# Resomnium > Resomnium designs organizations where humans and AI agents work together, then builds what they run on. The method is open (CellOS). The agents are live. Every chain of accountability ends with a human. Resomnium d.o.o., Zagreb, Croatia. Founded 2024. Works with a small number of long-term partners. ## What we do Most companies bolt AI onto an org chart built for humans working alone. The tools work. The structure doesn't. We redesign the structure: small accountable units (cells) where humans own judgment and agents handle execution. Then we build the agent systems those cells run on: scoped authority, signed receipts, escrowed work, a named human at the end of every chain. ## Work with us Three ways in. All long-term, all built so the partner owns the outcome. 1. **Embedded design partnership.** We work inside your organization until the new structure holds without us. Decision mapping, cell design, the agent layer. 2. **Venture and ecosystem co-build.** When the upside is real, we co-build agent-native products and infrastructure instead of billing hours. 3. **Research and open standards.** Agent governance and human-AI coordination, in consortium or commissioned. Output ships in public wherever possible. Start a conversation: https://resomnium.com/contact or filip@resomnium.com ## Try the method - **Organization Gap Analysis agent**: a paid agent that diagnoses an organization against the Cell framework and returns a written report. https://apify.com/resomnium/organization-gap-analysis ## Key concepts - **The Cell**: the smallest viable accountable unit. Five steward roles (Clarity, Execution, Narrative, Access, Integrity), each owned by a human. Flat: no boss node. Each steward runs AI agents beneath them, scoped to sub-functions. An agent team can mirror the same five roles one level down, so the structure is fractal. - **CellOS**: open-source framework (Apache 2.0) that encodes cells, roles, scope boundaries and escalation chains for human-AI teams. Early: v0.2.0. https://github.com/Resomnium/cellos - **Jobs, Skills, and Company of the Future**: three essay series by Filip Blagojević that lay out the thinking. https://resomnium.com/writing ## Writing - Index: https://resomnium.com/writing - Full text of every essay: https://resomnium.com/llms-full.txt - Original publication: https://filipblagojevic.substack.com ## For agents - **Zach, the Resomnium inbox (A2A)**: send a message and it is delivered to Filip Blagojević, who reads every inquiry and replies by email. Include an email address if you want a reply. No API key needed for message/send or tasks/get. Agent card: https://resomnium.com/.well-known/agent-card.json · Endpoint: https://api.resomnium.com/a2a - **MCP server**: CellOS spec, ways to work, inquiry delivery, and the paid Organization Gap Analysis agent. Card: https://resomnium.com/.well-known/mcp.json · Endpoint: https://mcp.resomnium.com/mcp - Zach does not generate answers or make commitments on Resomnium's behalf. A human does. ## Partners - **Apex Fusion**: multi-chain ecosystem. Its agent-native chain, Vector, has been live on mainnet since August 2025. - **HAL8**: the company building Apex Fusion. - **Thespian**: engineering partner. https://thespian.eu/ ## Founder Filip Blagojević. Leads Vector. Director of Delivery at HAL8. Head of Ecosystem at the Apex Fusion Foundation. Background: industrial automation (PLC software, factory commissioning), then protocol-level blockchain delivery, then production AI agents. ## Contact - Website: https://resomnium.com - Email: filip@resomnium.com - LinkedIn: https://www.linkedin.com/company/resomnium/ - X: https://x.com/ResomniumX - GitHub: https://github.com/Resomnium # Full text of all essays --- # Jobs of the Future, part 1: What Survives When AI Does Everything Else By Filip Blagojević, 2026-02-11. Canonical: https://filipblagojevic.substack.com/p/jobs-of-the-future-what-survives On-site: https://resomnium.com/writing/jobs-what-survives-when-ai-does-everything-else We are living through the fastest capability shift in human history. In 2020, the median expert prediction for artificial general intelligence was 50 years away. Today, that number has collapsed to somewhere between 5 and 10 years. Whether or not you believe those timelines, the practical reality is already here: AI can now perform knowledge work that, five years ago, required expensive professionals and years of training. This changes everything about how we think about careers, organizations, and value. I’ve spent years working at the intersection of technology, organizations, and incentive design—including time in industrial automation, where I watched firsthand what happens when machines become capable of doing what humans used to do. The pattern is always the same: first denial, then disruption, then reorganization around whatever humans still do better. The difference now is speed. What took manufacturing decades is happening to knowledge work in years. What I’ve come to believe—and what I’ll argue in this series—is that the future belongs to a specific set of human roles that AI cannot replace. Not because AI lacks capability, but because these roles exist precisely at the boundary where capability meets judgment, where execution meets meaning, and where systems meet accountability. This is not a comforting message about how “humans will always be needed.” That framing is lazy and probably wrong. Many humans, doing many jobs, will not be needed—at least not in the way they’re currently organized. The Bureau of Labor Statistics, McKinsey, and the World Economic Forum all converge on a stark projection: somewhere between 30% and 50% of current work hours are automatable by 2030. The WEF estimates 92 million jobs will be displaced in the next five years, even as 170 million new ones emerge. --- ## The Execution Trap Most professionals today are paid for execution. They take inputs, apply learned processes, and produce outputs. Analysts build models. Developers write code. Marketers create campaigns. Lawyers draft contracts. Consultants produce slide decks. Every single one of these tasks is now within reach of AI systems—and improving at a rate that makes yesterday’s benchmarks irrelevant. The current generation of frontier models passes professional licensing exams, writes production-quality code, and generates first drafts of almost any knowledge work product in minutes rather than hours. Here’s what this means: if your job can be described as a process—if someone could write a manual for what you do—you are in the execution trap. Your timeline isn’t a decade. It’s probably three to five years before the economics of your role fundamentally shift. I’ve seen this before. In industrial automation, we didn’t replace workers all at once. We replaced tasks, then processes, then entire functions. The workers who survived were the ones who moved up the abstraction ladder—from doing the work to designing the systems that did the work, and eventually to deciding what work was worth doing at all. The same pattern is now playing out in every knowledge profession. The abstraction ladder is the only way up. --- ## What AI Cannot Do AI systems, for all their power, share a common limitation: they optimize for objectives they’re given, but they cannot set those objectives. They can execute with superhuman speed and scale, but they cannot decide what’s worth executing. They can generate options, but they cannot exercise judgment about which options serve human purposes. David Autor at MIT calls this the “decision-making function”—the capacity to apply context, exercise judgment, and take responsibility for outcomes in situations where the right answer isn’t derivable from data alone. McKinsey’s research identifies three irreplaceable capabilities: aspiration (setting audacious goals), judgment (making calls under uncertainty), and creativity (generating solutions beyond existing patterns). The World Economic Forum’s analysis of 2,800+ skills found something remarkable: zero skills showed “very high capacity” for AI replacement when those skills were rooted in human interaction, contextual judgment, or accountability. This is where durable value lives. --- ## The Shape of What Remains Based on my research and experience designing organizations that function in high-complexity, high-stakes environments, I’ve identified five roles that I believe will remain fundamentally human through at least 2040—and possibly beyond, depending on how AI development unfolds. This list is not exhaustive. There are likely other durable roles I haven’t yet identified, and I expect to expand this framework as the landscape evolves. But these five represent a starting point—a core set of functions that appear robustly resistant to automation based on everything we currently know. These are job titles, not abstract competencies. Each one represents a distinct function that requires clear ownership and undivided accountability. They cannot be distributed across committees or shared among teams. When everyone is responsible, no one is responsible—and in a world moving this fast, diffused accountability is organizational suicide. One person holds each role. One person makes the call. One person owns the outcome. I’m not going to name them all here. Over the coming weeks, I’ll publish a deep dive on each one: what it actually involves, how it’s evolving, what skills it requires, and how to position yourself to hold it. What I will say is this: the five roles cluster around a common theme. They are all about things AI cannot do because they require human judgment, human relationships, or human accountability. They sit at the top of the abstraction ladder—above execution, above process, above optimization. They are about deciding what matters. --- ## The Qualities That Cut Across All Five Before I reveal the specific roles, I want to address something more fundamental: what kind of person can hold any of these positions? Because the roles themselves matter less than whether you have the underlying qualities to perform them. **Agency.** This is the foundation. The people who thrive in the AI age will be those who act without being told, who take ownership without being assigned it, who see what needs to happen and make it happen. AI is a tool for people with agency. For people without it, AI is just a more efficient way to wait for instructions. **AI fluency.** You cannot lead in a world you don’t understand. The humans who remain valuable will be those who use AI daily, who understand its capabilities and limitations from direct experience, who can orchestrate AI systems as naturally as they once managed spreadsheets. This isn’t about technical expertise. It’s about functional fluency—knowing what AI can do well, what it does poorly, and when to trust or override its outputs. **Learning velocity.** The half-life of technical skills has collapsed to under 2.5 years. The half-life of durable human skills—judgment, communication, relationship-building—is over 7.5 years. But even durable skills require constant updating as the context shifts. The people who hold these roles will be relentless learners, comfortable with perpetual incompleteness, treating every week as an opportunity to upgrade their mental models. **Comfort with ambiguity.** AI excels in well-defined problem spaces. Humans remain essential in ill-defined ones. If you need certainty before you act, you will be paralyzed. The valuable humans will be those who can make decisions with incomplete information, adjust as reality clarifies, and take responsibility for outcomes they couldn’t fully predict. **Taste.** This is the hardest to define and the most important. Taste is the ability to discern quality, to know what good looks like, to make judgment calls that can’t be reduced to metrics. AI can generate a thousand options. Taste is knowing which one is right. It comes from experience, from exposure, from caring deeply about craft. It cannot be taught directly, but it can be developed—and it will be the ultimate differentiator. These qualities aren’t nice-to-haves. They’re prerequisites. Without them, the specific roles don’t matter—you won’t be able to hold them anyway. --- ## The Timeline Problem I want to be honest about uncertainty. The durability of these roles depends on assumptions about AI capability growth that may prove wrong. If AGI arrives in 2030—which some credible researchers now assign 50% probability—then “durable through 2040” may be optimistic. If AI development plateaus or faces unexpected obstacles, these roles may remain valuable much longer. My working assumption is that these five functions represent a core of human competitive advantage in knowledge work—though certainly not the entirety of it. They may compress, evolve, or require continuous redefinition. But they will not disappear, because they address problems that are fundamentally about human purpose, human relationships, and human accountability. Organizations that recognize this early will restructure around these roles. Individuals who develop mastery in these functions will remain valuable regardless of what AI can do. --- ## What Comes Next In the articles that follow, I’ll go deep on each of the five roles I’ve identified so far. I’ll share the research, the frameworks, and the practical implications for anyone trying to build a career or an organization that lasts. And as the landscape evolves, I’ll add new roles to this series as they emerge. This is not career advice for a stable world. The world isn’t stable. This is a framework for navigating a transition that will be measured in years, not decades. The future of work isn’t about competing with AI. It’s about becoming the kind of human that AI makes more valuable, not less. --- *This is Part 1 of the Jobs of the Future series. Next week: the first of five roles—and why it might be the highest-leverage skill in the AI age.* --- # Jobs of the Future, part 2: The Clarity Steward By Filip Blagojević, 2026-02-11. Canonical: https://filipblagojevic.substack.com/p/jobs-of-the-future-the-clarity-steward On-site: https://resomnium.com/writing/jobs-the-clarity-steward Every organization has a strategy. Almost none of them can articulate it clearly. I’ve watched teams of smart, capable people work for months on projects that should never have been started. I’ve seen companies burn through funding building features nobody asked for, solving problems that weren’t the real problem. And in almost every case, the failure wasn’t intelligence or effort. It was clarity. Someone forgot to ask: *What are we actually doing here? Why does this matter? How does this connect to everything else?* In the age of AI, this problem gets worse, not better. When execution becomes cheap and fast, the cost of building the wrong thing drops to near zero—which means organizations will build more wrong things than ever before. The bottleneck is no longer “can we make this?” It’s “should we make this?” and “does everyone understand why?” This is the domain of the Clarity Steward. --- ## What This Role Actually Is The Clarity Steward is the guardian of strategic coherence. Their job is to ensure that everyone on the team—every human participant—understands the vision, believes in it, and can connect their daily work to it without ambiguity. This is not the same as being the CEO or the founder. The Clarity Steward may or may not set the original vision. But they are responsible for keeping it alive, keeping it sharp, and keeping it from getting fuzzy or watered down as it passes through the organization. Think of them as a translator operating on multiple frequencies simultaneously. At the highest level, they hold the overarching vision of the project—the “why” that justifies everything else. They can articulate it in a sentence or expand it into an hour-long conversation without losing coherence. At the individual level, they translate that vision into the language of each team member. They understand what each person cares about, what motivates them, what they’re optimizing for. And they connect the abstract strategy to those personal drivers in a way that makes the work feel meaningful, not mandatory. This is harder than it sounds. Most strategic communication fails because it’s broadcast, not translated. The Clarity Steward doesn’t send the same memo to everyone. They have different conversations with different people, finding the angle that resonates, the framing that clicks. --- ## The Incubation Function Before any strategic message goes public—to users, investors, partners, or the market—it must first achieve internal clarity. This is what I call the incubation phase. The Clarity Steward owns this phase. They are the quality gate. Nothing leaves the building until it passes through their filter for coherence and alignment. This means working closely with whoever handles external communication, marketing, and brand. But the division is clear: others shape the message for the world. The Clarity Steward shapes the thinking before the message exists. They are upstream of public narrative, ensuring that what gets communicated externally is rooted in something real and understood internally. Once incubation is complete and clarity is set, the Clarity Steward steps back. Others have autonomy to execute. But the Clarity Steward never stops monitoring. They watch for drift, for dilution, for the slow erosion that happens when strategy meets reality. --- ## Institutional Memory as Infrastructure Organizations forget why they made decisions. This is one of the most underrated sources of dysfunction in fast-moving teams. Someone proposes an idea. Others object: “We tried that before.” But no one remembers exactly what was tried, why it failed, or whether the conditions have changed. So the same debates happen repeatedly. The same mistakes get made. Institutional knowledge lives in people’s heads, and when those people leave or forget, the knowledge disappears. The Clarity Steward is responsible for preventing this. They own the institutional memory—not by personally remembering everything, but by designing and maintaining the knowledge architecture that captures decisions, reasoning, and context. This means saving files, conversations, and meeting minutes in structured ways. It means building systems—heavily reliant on AI—that make it easy to answer questions like: *Why did we decide X? What were the alternatives? What did we know at the time?* AI is exceptionally good at this kind of retrieval and synthesis, but only if someone has organized the raw material properly. The Clarity Steward is that someone. They don’t need to be a technologist, but they need to understand how to structure information so that both humans and AI can access it. This function will become more important, not less, as teams become smaller and more leveraged. When three people do the work of thirty, they cannot afford to lose context. The Clarity Steward ensures they don’t. --- ## The Game Theory Underneath Great Clarity Stewards understand something that average ones miss: strategy is not just about goals. It’s about incentives. They think in terms of game theory—not necessarily formally, but intuitively. They understand that people respond to incentives, that coordination problems are real, and that misaligned motivations will undermine even the clearest vision. This means understanding internal dynamics: how to structure work so that individual incentives align with collective goals, how to prevent defection, how to maintain cooperation when it would be easier to optimize locally. It means understanding external dynamics: anticipating how competitors will respond, how partners will behave, how the market will react to different strategic choices. And it means understanding the psychology of the team: what makes each person feel valued, what triggers resistance, what creates buy-in versus compliance. The Clarity Steward doesn’t need to be a game theorist. But they need to intuitively grasp that clarity without alignment is useless. You can explain the strategy perfectly and still fail if people’s incentives push them in different directions. --- ## The Low-Ego Requirement This role demands a specific personality trait that cannot be faked: low ego. Low ego means accepting when you’re wrong without it bothering you emotionally. It means changing your mind publicly when the evidence warrants it, without feeling diminished. It means holding strong opinions loosely—being willing to defend a position vigorously and then abandon it completely when someone makes a better argument. Low ego means leaving space for others to shine. The Clarity Steward often does invisible work—the alignment conversation that prevents a costly mistake, the reframing that unlocks someone’s motivation, the institutional memory that saves hours of rediscovery. None of this is glamorous. Much of it goes unnoticed. Low ego means celebrating others’ successes without jealousy. When teammates deliver exceptional results, the Clarity Steward takes quiet satisfaction in knowing the foundation was solid. This is not weakness. It’s a specific kind of strength—the strength to subordinate personal recognition to organizational coherence. People with high ego can be brilliant strategists, but they make poor Clarity Stewards. They hold onto positions too long. They make conflicts personal. They can’t create the psychological safety that allows others to challenge ideas. --- ## What Failure Looks Like A bad Clarity Steward is unsure. They are easily swayed in different directions by the people around them. They adjust the strategy based on whoever spoke last, whoever spoke loudest, whoever applied the most pressure. The team senses this and loses faith—not in the strategy, but in the strategist. A bad Clarity Steward lets the project lose focus. What started as a sharp, coherent vision becomes muddled. Scope creeps. Priorities multiply. Everyone is working on something, but no one can explain how it connects to the whole. A bad Clarity Steward does not track decision-making and reasoning properly. They avoid documentation because they’re afraid of what future revisions might reveal. They prefer ambiguity because it protects them from being proven wrong. But this is exactly backwards—the Clarity Steward’s job is to create accountability through transparency, including accountability for themselves. The core failure mode is this: confusing flexibility with weakness. A good Clarity Steward changes their mind when warranted, but from a position of strength—because the evidence changed, not because someone complained. A bad one changes their mind because it’s easier than holding the line. --- ## Who Should Consider This Role If you’re reading this and recognizing yourself, here are the signs that Clarity Steward might be your path: You naturally think in terms of coherence. When you hear a strategy or plan, you immediately notice the gaps, the contradictions, the unstated assumptions. You’re a translator. You can take a complex idea and explain it differently to different people, finding the framing that works for each. You have low ego but strong convictions. You don’t need credit, but you do need to be right—or at least to be honestly wrong, learning from mistakes rather than hiding them. You’re comfortable with ambiguity but driven to resolve it. You don’t freeze when things are unclear; you move toward clarity. You understand people. Not in a manipulative way, but in a genuine way—you’re curious about what makes others tick, what they care about, what they’re afraid of. If this sounds like you, the age of AI is your opportunity. As execution becomes automated, the human who can answer “why are we doing this?” and make everyone believe it becomes the most valuable person in the room. --- ## The Leverage Is Enormous Here’s what makes this role so powerful: clarity is a multiplier. When the vision is clear and everyone is aligned, execution accelerates. Decisions happen faster because the criteria are known. Conflicts resolve more easily because there’s a shared reference point. Mistakes get caught earlier because people understand what success looks like. When clarity is missing, everything slows down. Every decision becomes a negotiation. Every priority becomes contested. Every misalignment compounds into dysfunction. In a world where AI handles execution at unprecedented speed, the limiting factor is no longer how fast you can build. It’s how clearly you can think. The Clarity Steward is the role that makes everything else work. It’s invisible when done well and catastrophic when done poorly. And as organizations get smaller and more leveraged, there is no room to distribute this function across committees or share it among teams. One person holds the clarity. One person makes the call. One person owns the coherence. That’s the Clarity Steward. --- *This is Part 2 of the Jobs of the Future series. Next week: the role that turns strategy into reality by orchestrating the humans and machines that build.* --- # Jobs of the Future, part 3: The Execution Orchestrator By Filip Blagojević, 2026-02-11. Canonical: https://filipblagojevic.substack.com/p/jobs-of-the-future-the-execution On-site: https://resomnium.com/writing/jobs-the-execution-orchestrator Ideas are cheap. Execution is everything. This has always been true, but it meant something different when execution required large teams, long timelines, and massive budgets. Today, a single person with the right skills can orchestrate AI agents, automated systems, and selective human support to build what once required dozens of people and months of work. The bottleneck has shifted. It’s no longer “can we find enough people to build this?” It’s “can we find someone who understands the entire stack well enough to orchestrate its construction?” This is the domain of the Execution Orchestrator. --- ## What This Role Actually Is The Execution Orchestrator is the person who turns vision into reality. They take the strategy, the requirements, the desired outcome—and they figure out how to actually build it. Not by doing everything themselves, but by orchestrating the systems and resources that do. This requires a specific kind of knowledge: not deep expertise in one area, but informed competence across every level of execution. They understand design, engineering, infrastructure, quality assurance, deployment, and operations—not necessarily well enough to be the best practitioner in any single domain, but well enough to direct, evaluate, and course-correct the work being done. Think of them as a deeply informed generalist. Having been a former practitioner helps—someone who has written code, shipped products, or built systems from scratch brings invaluable intuition. But what matters most is understanding how all the pieces fit together and how to control the machinery that assembles them. The machinery, increasingly, is AI. --- ## The New Execution Stack In the emerging model, the Execution Orchestrator’s primary collaborators are not humans. They are AI agents. The ideal execution stack is 100% automated. AI agents handle code generation, testing, documentation, deployment. Robotic and automated hardware systems handle physical processes where relevant. Human sub-contractors are the last resort—brought in only when automation cannot yet handle the task or when specialized human judgment is temporarily required. This is a fundamental shift. Traditional project managers coordinated human teams. The Execution Orchestrator coordinates intelligent systems. The skills are related but not the same. An elite Execution Orchestrator comes prepared. They have agentic workflow templates refined through iteration. They have volumes of tested prompts for different purposes—prompts that reliably produce quality outputs for specific task types. They have deep familiarity with the tools: which AI models excel at what, which frameworks handle which workflows, where the failure modes hide. This toolkit is never finished. A great Execution Orchestrator is constantly building, testing, and refining their systems. Not because someone told them to, but because mastery of the craft demands it. The toolkit is how they multiply their leverage. --- ## The Abstraction Ladder The Execution Orchestrator’s goal is to operate at the highest level of abstraction possible while maintaining control over quality. At the lowest level, you’re doing the work yourself—writing code, designing interfaces, fixing bugs. This is where you start when learning a new domain or when automation fails. One level up, you’re directing AI agents to do the work while reviewing their outputs closely. You’re in the details, catching errors, re-prompting, iterating. Higher still, you’re designing workflows that AI agents execute semi-autonomously, with automated quality checks that flag issues for your attention. At the highest level, you’re orchestrating entire systems of agents and automated processes, intervening only on exceptions and strategic decisions. The elite Execution Orchestrator always strives to climb this ladder—but not too quickly. Move up before you’ve established reliable quality control at the current level, and you lose visibility into problems until they compound into disasters. The art is knowing when you’ve earned the right to abstract away. When things break—and they will—the Execution Orchestrator must be willing to descend the ladder and get their hands dirty. They are the ultimate owner of quality. Automated checks are the first line of defense, but when those fail, the Execution Orchestrator personally diagnoses and fixes the issue. Whatever it takes. However long it takes. This is not optional. If you’re not willing to go into the details when necessary, you cannot hold this role. --- ## The Practical Mindset The Execution Orchestrator is relentlessly practical. They think in terms of time, money, and feasibility. When someone proposes a feature, the Execution Orchestrator immediately starts calculating: How long will this take? What dependencies does it create? What could go wrong? What’s the minimum viable version? They have an intuitive sense for complexity—they can smell when something that sounds simple will actually be a nightmare to implement. This practicality must be paired with awareness of context. The Execution Orchestrator needs to understand what stage the project is in and calibrate accordingly. In early stages, speed matters more than polish. Ship fast, learn fast, accumulate technical debt deliberately with eyes open. In later stages, quality and scalability become paramount. The same shortcut that was smart in month one becomes dangerous in month twelve. This is a balancing act with no universal formula. Speed, quality, cost—pick two, sacrifice one. The Execution Orchestrator navigates this tradeoff daily, making judgment calls that depend on context, priorities, and risk tolerance. They don’t make these calls alone. The decision about what stage the project is in, and therefore what tradeoffs are appropriate, is a mutual decision across roles. But the Execution Orchestrator’s voice in that discussion is essential. They’re the one who knows what’s actually possible and at what cost. --- ## The Obligation to Push Back Here’s where many Execution Orchestrators fail: they become order-takers. Someone with strategic authority says “build this.” The Execution Orchestrator says “okay” and starts building. They focus so intensely on delivery that they never stop to think: Should we actually build this? Is this the right scope? Are the assumptions behind this request sound? This is a failure of the role, not a feature of it. The Execution Orchestrator has not just the authority but the *obligation* to push back when something doesn’t make sense. They are often the first person to see that a plan is impractical, that a timeline is unrealistic, that a feature set is bloated, that a technical approach is flawed. Imagine this scenario: someone wants to jam too many features into version one of a product. They’re optimistic, excited, eager to impress users or investors. The Execution Orchestrator sees the reality—that this scope will drastically prolong development, make testing exponentially harder, and turn debugging into a nightmare. A weak Execution Orchestrator stays quiet and tries to deliver the impossible. A strong one speaks up: “Here’s what this scope actually means in time and risk. Here’s what I recommend instead—a sensible version one that we can ship in half the time and iterate from. Let’s discuss.” This isn’t insubordination. It’s the job. The Execution Orchestrator is the reality check on ambition. They ground the vision in what’s actually achievable. Without this voice, organizations build castles in the sky and wonder why they never ship. --- ## What Failure Looks Like The most common failure mode is subordination—putting yourself in a lower position of influence, preferring to stick to execution rather than engaging with strategy. This happens for understandable reasons. Execution is concrete. You can point to what you built. Strategic discussions feel abstract, political, exhausting. It’s tempting to retreat into the comfort of building and leave the “big picture” to others. But an Execution Orchestrator who doesn’t engage strategically becomes a bottleneck at best and a liability at worst. They build the wrong things efficiently. They optimize for output rather than outcomes. The opposite failure is also real: pushing back so much that you become a blocker. Saying “that’s impossible” to everything. Letting perfect be the enemy of good. Hiding behind complexity to avoid difficult work. The Execution Orchestrator must challenge assumptions without becoming the person who kills every initiative. Another failure mode is getting lost in details. The Execution Orchestrator who spends all their time debugging individual issues has no time or energy to see the project whole. They lose the forest for the trees. They miss systemic problems because they’re too deep in symptomatic ones. And there’s a psychological trap specific to this role: convincing yourself that you’re the only one doing real work. Everyone else is “just talking” while you’re actually building. This breeds resentment, damages collaboration, and is usually wrong. The antidote is genuine generalism. Immerse yourself in other functions. Understand what the strategic roles actually do—the complexity of their decisions, the difficulty of their tradeoffs. We all tend to find our own work hardest and most valuable. Technical people, in my experience, are particularly prone to this bias, including myself. The Execution Orchestrator must actively resist it. --- ## Good Versus Great A good Execution Orchestrator gets things done. They listen carefully, execute what’s planned, and push back when absolutely necessary. They’re reliable. They ship. A great Execution Orchestrator does all that and more. They proactively challenge assumptions, not waiting for obvious problems but anticipating them. They don’t just execute the plan—they improve it through the act of building. They build quick and dirty prototypes to show their ideas, not just describe them. When they see a better approach, they demonstrate it. They make the alternative tangible. They stay at a high level most of the time, which gives them the time and energy to see the project whole. They know when to zoom in and when to zoom out. They ask for help when they need it rather than drowning quietly. And perhaps most importantly: they love the game. Execution is not a burden to them—it’s a craft. They find genuine satisfaction in making things work, in solving the puzzle of delivery, in the moment when the system finally does what it’s supposed to do. This is not something you can fake. Either you find joy in the act of building, or you don’t. If you don’t, this role will burn you out. --- ## Who Should Consider This Role If you’re reading this and recognizing yourself, here are the signs that Execution Orchestrator might be your path: You’ve built things before. Maybe not at expert level in any one domain, but you’ve shipped. You know what it feels like to take something from idea to reality. You’re a systems thinker. You naturally see how pieces connect, where dependencies hide, what order things need to happen in. You’re practical to your core. Abstract discussions frustrate you if they don’t lead to concrete action. You want to know: what are we actually doing? You’re learning AI tools obsessively. Not because someone told you to, but because you see the leverage. You’re already experimenting with agents, prompts, automated workflows. You have strong opinions but update them fast. You’ll argue for an approach, but if someone shows you a better one, you switch without ego. You enjoy the grind. Not masochistically, but genuinely—there’s something satisfying to you about working through hard problems until they’re solved. If this sounds like you, the age of AI is your moment. The Execution Orchestrator who can reliably turn vision into reality—fast, at high quality, with minimal human overhead—will be among the most valuable people in any organization. --- ## The Leverage Is Enormous Here’s what makes this role so powerful: execution has become infinitely scalable for those who know how to orchestrate it. A decade ago, building a product meant hiring a team. Today, one skilled Execution Orchestrator with the right toolkit can match or exceed that team’s output. The constraint isn’t labor—it’s the ability to direct intelligent systems effectively. This means the gap between good and great Execution Orchestrators is enormous. A good one delivers. A great one delivers at 10x the speed, or with 10x the scope, or at a fraction of the cost. The leverage compounds. Organizations are just beginning to understand this. Most are still organized around human teams, with managers who coordinate people rather than orchestrators who coordinate systems. That model is already obsolete for many types of work. The organizations that recognize this first will outcompete those that don’t. One person who can orchestrate AI agents, automated systems, and selective human support to deliver a complete product—reliable, scalable, on time—is worth more than a traditional team of ten. That’s the Execution Orchestrator. --- *This is Part 3 of the Jobs of the Future series. Next time: the role that shapes how the world sees what you’ve built.* --- # Jobs of the Future, part 4: The Narrative & Distribution Steward By Filip Blagojević, 2026-02-11. Canonical: https://filipblagojevic.substack.com/p/jobs-of-the-future-the-narrative On-site: https://resomnium.com/writing/jobs-the-narrative-and-distribution-steward People don’t buy products. They buy stories. This has always been true, but it mattered less when distribution was expensive and attention was captive. You ran ads, people saw them, some percentage bought. The funnel was inefficient but functional. Today, attention is fragmented across infinite channels. AI can generate unlimited content. Every company can produce professional-looking marketing materials in minutes. The noise is deafening. In this environment, the ability to cut through—to make people feel something, to build genuine emotional connection, to turn a product into a story people want to be part of—becomes the scarcest and most valuable skill. This is the domain of the Narrative & Distribution Steward. --- ## What This Role Actually Is The Narrative & Distribution Steward is the person who shapes how the world perceives what you’ve built. They own the story, the brand, the feel, the tone—and the channels through which all of it reaches people. This is not traditional marketing. Traditional marketing teams optimize for views, clicks, and interactions—intermediate metrics that have become disconnected from actual business outcomes. Entire industries exist to pump these numbers through schemes that generate impressive dashboards but don’t move revenue or build lasting brand equity. The Narrative & Distribution Steward rejects this. They own the *outcome*: how the market actually perceives the product, whether that perception translates into meaningful business results, and whether the brand accumulates value over time. This is hard to measure perfectly, and that’s part of the job. The Narrative & Distribution Steward must work on defining what success looks like—revenue, conversion, brand sentiment, customer loyalty—and build systems to measure what can be measured while maintaining judgment about what can’t. If you need a neat KPI dashboard to feel confident, this role isn’t for you. --- ## The Two Stories Every product has two stories that matter. The first is the *origin story*: Why does this exist? Who built it and why? What problem did they see that others missed? What beliefs and values drove them to create this? Humans are wired for origin stories. We want to know where things come from. A product with a compelling origin story feels different from a product that just appeared. It has meaning, intention, soul. The second is the *user story*: How does this product fit into my life? What does it make possible for me? Who do I become when I use it? What tribe am I joining? This story is about the customer, not the company. It answers the question every potential user is subconsciously asking: “What’s in it for me, and what does choosing this say about who I am?” The Narrative & Distribution Steward must craft both stories and weave them together. The origin story creates authenticity and trust. The user story creates desire and belonging. Together, they create the emotional resonance that turns a product into a brand. --- ## Distribution Is Not an Afterthought The title includes “Distribution” for a reason. A brilliant narrative that reaches no one is worthless. The Narrative & Distribution Steward owns channel strategy completely. Which platforms? Which audiences? Paid or organic? Partnerships or direct? Broad reach or targeted resonance? These are not tactical questions to delegate. They are strategic questions that shape whether the narrative actually lands. The right answer depends on the product and the stage. Sometimes you want to reach a million loosely interested people to build awareness. Sometimes you want to reach ten thousand deeply aligned ones who will become evangelists. Sometimes you start narrow and expand. Sometimes you start broad and refine. The Narrative & Distribution Steward must be comfortable making these calls, adjusting as they learn, and defending their choices with reasoning—not just intuition. They must also be relentlessly curious about new channels. The landscape shifts constantly. Platforms rise and fall. Algorithms change. Attention migrates. The Narrative & Distribution Steward who relies only on established playbooks will eventually find those playbooks obsolete. --- ## The Tools of Storytelling Story can be told through any medium: text, audio, video, short films, infographics, interactive experiences, physical products, events—even scents if that’s what the brand calls for. The Narrative & Distribution Steward doesn’t need to be an expert creator in every medium. Like the Execution Orchestrator, they are a director who orchestrates specialists and AI systems to produce the work. But also like the Execution Orchestrator, they must be willing to get their hands dirty when necessary. This means understanding each medium well enough to evaluate quality, give meaningful feedback, and know when something isn’t working. You can’t direct what you don’t understand. AI has transformed what’s possible here. A single person can now produce content across multiple formats that would have required a full agency a few years ago. The Narrative & Distribution Steward who masters AI-assisted content creation gains enormous leverage. But AI also raises the floor. Everyone can produce competent content now. The Narrative & Distribution Steward’s job is to produce *exceptional* content—content that makes people feel something, that stands out from the algorithmically-generated noise, that carries genuine human insight and intention. This is where taste becomes essential. AI can generate options. The Narrative & Distribution Steward must know which options are right. --- ## The Domain Knowledge Problem Here’s where most marketing people fail: they don’t actually understand what they’re marketing. Traditional marketing teams like to do “marketing things”—campaigns, content calendars, brand guidelines, creative briefs. They expect someone else to prepare detailed narratives and explanations of the product. They’re comfortable with the surface. This doesn’t work anymore. Not in a world where audiences can smell inauthenticity instantly, where technical products require genuine understanding to position correctly, where the line between product and story has dissolved. The Narrative & Distribution Steward must be a generalist. They need to understand the domain deeply—not at the level of the builders, but far beyond the level of traditional marketers. They need to understand not just what the product does, but *why it matters*, *how it works*, and *what makes it different* at a level that allows them to tell authentic stories. This means learning constantly. It means sitting in technical discussions, asking dumb questions until they become smart ones, reading what the experts read, immersing themselves in the world the product inhabits. It also means actively participating in strategic discussions. The Narrative & Distribution Steward should be challenging other roles, pushing back when positioning doesn’t resonate, contributing to product decisions from a market perspective. They are not downstream of strategy. They are part of it. --- ## Collaboration and Pushback The Narrative & Distribution Steward works in constant collaboration with other roles—especially whoever owns strategic clarity. Internal strategic thinking must be incubated before it goes public. The Narrative & Distribution Steward takes that incubated thinking and translates it for the world. But the handoff is not clean or one-directional. It’s a continuous dialogue. Sometimes the Narrative & Distribution Steward must push back: “This positioning won’t resonate externally. Here’s what I’m hearing from the market. We need to rethink this.” This is not overstepping. It’s the job. Consensus is always the goal. All parties should debate, stress-test, and ultimately commit to a shared direction. But consensus doesn’t mean the Narrative & Distribution Steward defers to internal preferences. They bring the market’s voice into the room. When alignment is achieved, commitment must be total. A brand that sends mixed messages because internal stakeholders couldn’t agree—or couldn’t stick to what they agreed—will fail. The Narrative & Distribution Steward is a guardian of that consistency. --- ## What Failure Looks Like The most common failure is domain ignorance. A Narrative & Distribution Steward who doesn’t understand the product—who treats it as an interchangeable object to apply generic marketing frameworks to—will produce work that feels hollow. The audience senses it. The story doesn’t land. The brand never develops real equity. This happens because many people attracted to marketing prefer the creative and social aspects of the role and resist the hard work of understanding technical or complex products. They want to craft messages, not study domains. But you cannot tell an authentic story about something you don’t understand. And audiences can tell the difference. Another failure mode is metric fixation. The Narrative & Distribution Steward who optimizes for views and clicks—because those are easy to measure—will produce content that performs on dashboards but doesn’t move the business. They’ll celebrate vanity metrics while the brand stagnates. Another is channel stagnation. The Narrative & Distribution Steward who masters one platform and stops learning will eventually find their audience has migrated elsewhere. The landscape changes too fast for anyone to coast. And there’s a collaboration failure: becoming either too deferential or too dominant. Too deferential, and you become an order-taker who executes internal mandates without challenging them. Too dominant, and you override legitimate input from other roles who understand things you don’t. --- ## Good Versus Great A good Narrative & Distribution Steward knows the product, understands the market, and uses proven approaches effectively. They build solid brand foundations. They run campaigns that work. They’re reliable. A great Narrative & Distribution Steward does all that while constantly questioning the process. They study human psychology—not as an academic exercise, but as a practical obsession. They want to understand why stories work, how emotions drive decisions, what makes something memorable versus forgettable. They read widely. They notice what captures their own attention and reverse-engineer it. They experiment with new channels and new formats before they’re proven. They’re not reckless, but they’re not conservative either. They understand that the biggest opportunities exist where proven playbooks don’t yet exist. They learn from other roles, genuinely curious about how strategy forms, how products get built, how partnerships develop. This cross-functional fluency makes their narratives richer and more authentic. And they teach. They share their process with other roles, helping the entire organization develop better instincts about how to communicate, how to position, how to tell stories. A great Narrative & Distribution Steward elevates everyone’s ability to represent the brand. Most importantly, they love the craft. They have genuine agency and drive to excel—not because someone is measuring them, but because shaping perception and building emotional connection is endlessly fascinating to them. --- ## Who Should Consider This Role If you’re reading this and recognizing yourself, here are the signs that Narrative & Distribution Steward might be your path: You’re obsessed with why some things capture attention and others don’t. You find yourself analyzing ads, studying viral content, noticing what makes you feel something. You’re a storyteller. You naturally frame information as narrative. You understand that how you say something matters as much as what you say. You’re curious about everything. You don’t want to be confined to “marketing.” You want to understand the product, the market, the technology, the users—all of it. You’re comfortable with ambiguity in measurement. You can pursue outcomes that aren’t perfectly quantifiable without losing confidence or direction. You’re a student of human nature. You’re genuinely interested in why people do what they do, want what they want, believe what they believe. You’re willing to defend your perspective but also to change it. You push back when you’re right, but you listen when you’re wrong. If this sounds like you, the age of AI is your opportunity. When everyone can produce content, the human who can produce *meaning*—who can make people feel something real—becomes invaluable. --- ## The Leverage Is Enormous Here’s what makes this role so powerful: perception shapes reality. A product with a great story outsells a better product with no story. A brand with emotional resonance commands premium prices. A narrative that captures imagination attracts talent, partners, and customers who want to be part of something. In a world where AI commoditizes execution and content, the ability to shape how people *feel* about what you’ve built becomes the ultimate differentiator. The Narrative & Distribution Steward is not a support function. They don’t make the product look good after it’s built. They shape whether the product succeeds at all. One person who can craft authentic stories, distribute them through the right channels, and build genuine emotional connection with the market—that person is the difference between a product that exists and a brand that matters. That’s the Narrative & Distribution Steward. --- *This is Part 4 of the Jobs of the Future series. Next time: the role that opens doors no marketing can open.* --- # Jobs of the Future, part 5: The Power & Access Steward By Filip Blagojević, 2026-02-11. Canonical: https://filipblagojevic.substack.com/p/jobs-of-the-future-the-power-and On-site: https://resomnium.com/writing/jobs-the-power-and-access-steward Some doors only open from the inside. You can have the best product in the world, the clearest strategy, the most compelling narrative—and still fail because you couldn’t reach the people who matter. The investor who could fund your next phase. The partner who could open an entire market. The enterprise client who could transform your revenue overnight. These people don’t respond to cold emails. They don’t click on ads. They don’t fill out contact forms. They operate in a different economy—one based on trust, reputation, and human relationships that no algorithm can replicate. This is the domain of the Power & Access Steward. --- ## What This Role Actually Is The Power & Access Steward is the person who opens doors that would otherwise stay closed. They handle the highest-stakes external relationships: top-tier customers, strategic partners, and capital sources of all kinds. This is similar to what business development professionals do today, but with a critical difference: the Power & Access Steward has real authority. They’re not a messenger carrying proposals back and forth. They have decision-making power, accountability, and the organizational standing that high-value stakeholders demand. This matters because the people on the other side of these conversations—the decision-makers at potential partners, the managing directors at investment firms, the executives at enterprise clients—want to speak with someone who matters. They can tell immediately when they’re talking to someone without authority. And they disengage just as quickly. The Power & Access Steward must genuinely be important, not just appear important. This means real equity, real decision-making power over certain domains, and real accountability for outcomes. Organizations must be careful to calibrate this correctly—too little power and the role becomes ineffective; too much and it creates imbalance. --- ## The Three Domains The Power & Access Steward operates across three interconnected domains, with priority shifting based on what the project needs most at any given stage. **High-value customers.** These are the top tier of whatever customer segment matters—the accounts that pay the most, influence the most, or open the most doors. In B2B, this might be enterprise clients. In other contexts, it might be key accounts or strategic early adopters. The threshold varies, but the principle is constant: this person handles the relationships that are too important to automate or delegate. **Strategic partnerships.** Not every partnership is meaningful. The Power & Access Steward focuses on the ones that are—partnerships based on genuine strategic fit, complementary capabilities, or shared values. What makes a partnership meaningful depends entirely on context. One project might need a partner with distribution reach. Another might need technical capabilities. Another might need credibility by association. The Power & Access Steward identifies which partnerships actually matter and pursues them relentlessly. **Capital.** This includes all forms: venture capital, angel investment, strategic investments from partners, grants, debt—whatever the project needs at its current stage. The Power & Access Steward is the face of fundraising. They have the accountability that comes with it and the authority to make commitments within defined parameters. The trap is spreading too thin. Every high-value stakeholder wants attention. The Power & Access Steward must prioritize ruthlessly, understanding that saying yes to one conversation means saying no to another. Unlike other functions where you might hire additional people to scale, this role doesn’t work that way. The whole point is that stakeholders want to talk to *this* person—the one with authority and standing. Adding more bodies doesn’t solve the constraint. --- ## The Network Advantage Let’s be direct: having an existing network helps enormously. The Power & Access Steward who enters a domain with established relationships, a track record, and a reputation has a massive head start. Doors that take years to open for an unknown are already ajar for someone with social proof. Does this mean someone without an existing network can’t succeed? Not necessarily. Passion, energy, and a proven track record of building networks in other domains can compensate—especially in emerging fields where few people have established relationships. But it’s harder. The ramp is steeper. The role demands that you build trust quickly with people who have many demands on their attention and little patience for unproven actors. The ideal Power & Access Steward brings at least some experience and connections in the relevant domain. They don’t start from zero. They start from a foundation of relationships they can leverage and expand. --- ## The Human Element AI is transforming this role, but it cannot replace its core. The Power & Access Steward uses AI heavily for prospecting, lead generation, research, and preparation. Before any high-stakes conversation, they should know everything knowable about the person across the table—their background, interests, recent activities, mutual connections, potential alignment. AI makes this preparation faster and more comprehensive than ever before. But the conversation itself? That must be human. Decision-makers can sense when they’re being handled by a script or a process. They respond to authenticity, to genuine understanding, to the feeling that the person across from them actually cares about the relationship and not just the transaction. This is true for investors evaluating whether to trust you with their capital. It’s true for partners evaluating whether to tie their reputation to yours. It’s true for enterprise clients evaluating whether you’ll be there when things go wrong. The human touch is not a nice-to-have. It’s the entire point. This means the Power & Access Steward must be genuinely relational. Not performatively friendly, but actually interested in people. The classic extroverted salesperson archetype has some truth to it—this role requires comfort with strangers, energy in social situations, and the ability to build rapport quickly. But more important than extroversion is authenticity and domain knowledge. The Power & Access Steward who doesn’t fully understand the product, who can’t speak intelligently about the domain, who relies on other roles to prepare all their materials—that person will fail in rooms where sophistication is expected. High-value stakeholders ask hard questions. They probe for depth. They notice when someone is faking it. --- ## The Relationship to Other Functions The Power & Access Steward works in constant dialogue with other roles, particularly whoever handles narrative and brand. There’s often natural tension here—similar to the current tension between marketing and business development. The narrative function casts a wide net, building brand recognition and driving volume. The Power & Access Steward hunts specific targets with personalized approaches. In a retail-focused product, the narrative function might drive masses of consumers to the point of sale while the Power & Access Steward works on expanding partnerships and strategic outreach. In a B2B context, the narrative function might focus on brand recognition and thought leadership while the Power & Access Steward manages key accounts and closes enterprise deals. The ideal is synergy: brand creates air cover and credibility; targeted relationship-building converts the highest-value opportunities. But achieving this synergy requires ongoing communication, mutual respect, and clear understanding of who owns what. The Power & Access Steward must also stay deeply connected to whoever owns strategic clarity. High-stakes conversations require deep understanding of what the project is, where it’s going, and what tradeoffs are acceptable. The Power & Access Steward who operates in isolation—who makes commitments that conflict with strategy or misrepresents the product’s direction—creates chaos. --- ## What Failure Looks Like The most common failure is self-dealing. The Power & Access Steward has access to relationships and opportunities that could benefit them personally, sometimes at the project’s expense. They could prioritize growing their own network over pursuing the partnerships the project actually needs. They could angle for roles at other companies through conversations that are supposed to be about partnerships. They could spend time on high-status but low-value activities because those activities serve their personal brand. This is fundamentally an integrity problem, and there’s no structural solution that fully addresses it. You can create accountability mechanisms, but ultimately this role requires someone whose character you trust. This is typically assessed through intuition, track record, and references from mutual connections. It’s imperfect. But hiring someone untrustworthy into this role is catastrophic. Another failure mode is laziness disguised as delegation. The Power & Access Steward who expects other roles to prepare all materials, craft all messaging, and set up all meetings—who sees their job as showing up to conversations others have arranged—is not doing the job. Yes, there’s collaboration. Yes, others contribute. But the Power & Access Steward owns the outcomes and must own the work that drives them. A third failure mode is mechanical execution without strategic thinking. Setting up meetings, following up on leads, checking boxes—this is necessary but not sufficient. The Power & Access Steward who operates mechanically, without proactively seeking novel opportunities or challenging assumptions about where value might hide, leaves enormous potential on the table. --- ## Good Versus Great A good Power & Access Steward is diligent. They set up the meetings, do the follow-ups, maintain the relationships, execute the playbook. They’re reliable. Deals close. Partnerships form. Capital comes in. A great Power & Access Steward does all that while operating on a different level. They proactively identify opportunities others miss. They see connections between their network and the project’s needs that aren’t obvious. They bring ideas to other roles: “I was talking to someone who made me think about our positioning differently” or “There’s an adjacent market we haven’t considered.” They use their network creatively. Not just for direct business development, but to find novel ways to solve problems—introductions that unlock stuck situations, insights from people who’ve faced similar challenges, access to resources through unexpected channels. They stay deeply engaged with the substance of the project. They don’t outsource understanding to other roles. They can speak about the product, the strategy, the vision with genuine depth because they’ve invested in learning it. This makes their external conversations richer and more credible. And they maintain genuine relationships even when there’s no immediate transactional value. They understand that networks compound over time, that today’s loose connection might be tomorrow’s crucial partner, that reputation is built through consistent behavior over years. Most importantly, they put the project first. Their personal network growth happens as a byproduct of doing the job well, not as the primary objective. When there’s tension between what’s good for them and what’s good for the project, they choose the project. This is rarer than it should be. --- ## Who Should Consider This Role If you’re reading this and recognizing yourself, here are the signs that Power & Access Steward might be your path: You’re naturally relational. You enjoy meeting people, building connections, maintaining relationships over time. It energizes you rather than draining you. You have some network already. Maybe not in the exact domain you want to work in, but somewhere. You’ve demonstrated you can build trust and maintain relationships. You’re comfortable with ambiguity and rejection. High-stakes business development involves many nos for every yes. You don’t take rejection personally, and you can operate without constant validation. You understand the substance. You’re not content to be a “people person” who relies on others for expertise. You want to genuinely understand what you’re representing. You have integrity. When you examine your motivations honestly, you find that you care about outcomes beyond your personal advancement. You can be trusted with access to opportunities. You’re energized by the hunt. Finding opportunities, pursuing them, closing them—this isn’t exhausting to you. It’s the game you want to play. If this sounds like you, the age of AI is your opportunity. AI can do research and preparation better than any human. But AI cannot sit across from a skeptical investor and make them believe. AI cannot build the trust that makes a partner take a risk on you. AI cannot maintain the human relationships that open doors. The Power & Access Steward who combines AI-augmented preparation with genuine human connection becomes extraordinarily effective—more effective than any previous generation of business development professionals could be. --- ## The Leverage Is Enormous Here’s what makes this role so powerful: access is a multiplier that nothing else can replace. The best product without access to customers is a failed product. The best strategy without access to capital can’t execute. The best partnership opportunity without access to decision-makers never materializes. The Power & Access Steward is often the difference between potential and reality. They convert what could be into what is. And because this role is fundamentally human—because trust cannot be automated and relationships cannot be scaled through technology—it becomes more valuable, not less, as AI transforms everything around it. One person who can open the right doors, build the right relationships, and bring the right resources to the table—that person makes things possible that would otherwise be impossible. That’s the Power & Access Steward. --- *This is Part 5 of the Jobs of the Future series. Next time: the role that keeps everything from falling apart.* --- # Jobs of the Future, part 6: The Integrity Steward By Filip Blagojević, 2026-02-11. Canonical: https://filipblagojevic.substack.com/p/jobs-of-the-future-the-integrity On-site: https://resomnium.com/writing/jobs-the-integrity-steward Every system decays without maintenance. Teams that start aligned drift apart. Incentives that once motivated become stale or perverse. Agreements that were clear become ambiguous. Small resentments accumulate into major conflicts. People leave. People underperform. People act in their own interest at the expense of the whole. This is not a failure of character. It’s physics. Entropy applies to organizations just as it applies to everything else. Without active effort to maintain coherence, any human system will gradually fall apart. Most organizations pretend this isn’t true. They build structures, write contracts, set compensation—and then assume the machine will run itself. It won’t. Someone has to watch for dysfunction, address it early, and continuously tune the system to keep it healthy. This is the domain of the Integrity Steward. --- ## What This Role Actually Is The Integrity Steward is the immune system of the organization. They protect against threats—bad actors, misaligned incentives, legal exposure, cultural decay, unresolved conflicts—and they maintain the conditions that keep people motivated, aligned, and productive. This role combines two functions that are usually separated: the stick and the carrot. The stick is governance and accountability. Legal structure. Contracts—both internal and external. Consequences for bad behavior. Clear boundaries. The Integrity Steward ensures that agreements are documented, enforceable, and enforced. The carrot is incentive design. Compensation. Equity. Recognition. The Integrity Steward ensures that everyone is rewarded in a way that motivates them to be productive and effective, even when things get hard. Most organizations split these functions. Lawyers handle contracts. HR handles compensation. Managers handle disputes. The Integrity Steward integrates all of it because these things cannot be separated. How you structure agreements shapes behavior. How you design incentives determines whether agreements hold. How you handle conflicts reveals whether your stated values are real. Someone must see the whole picture. That’s the Integrity Steward. --- ## The Carrot: Designing Incentives That Work Compensation is not just about money. It’s about alignment. The Integrity Steward owns how everyone is paid, how much equity they receive, and how rewards are distributed. But more importantly, they own the *logic* behind these decisions—the system that determines who gets what and why. This requires deep understanding of game theory, even if not formally trained in it. The Integrity Steward must anticipate how people will actually behave given the rules they’re operating under. They must design systems that make good behavior easy and bad behavior costly—not through surveillance and punishment, but through intelligent structure. Consider: If you reward individual performance but need collaborative behavior, you’ll get competition when you need cooperation. If you distribute equity equally regardless of contribution, you’ll demotivate your highest performers. If you pay below market, you’ll lose people at the worst moments. If you pay above market without accountability, you’ll attract people who want comfort over challenge. Every incentive structure has consequences, intended and unintended. The Integrity Steward thinks through these consequences before they manifest. But incentive design isn’t purely analytical. It’s also deeply human. People care about fairness, not just absolute outcomes. They care about recognition, not just money. They care about social standing within the team, not just their title. The Integrity Steward must understand these softer dimensions and factor them into decisions. This means knowing the people—their private circumstances, their motivations, their concerns. What keeps the Execution Orchestrator motivated to deliver might be different from what keeps the Narrative Steward engaged. Compensation isn’t one-size-fits-all. It’s personalized within a coherent framework. --- ## The Stick: Governance and Accountability The Integrity Steward owns the legal and contractual foundation of the organization. This includes the legal structure of the entity itself, employment agreements, equity agreements, partnership contracts, vendor agreements, customer contracts—anything that creates binding commitments. They don’t need to be a lawyer, but they need to understand legal frameworks deeply enough to make strategic decisions. They can work with external counsel for technical matters, using AI to optimize costs and minimize billable hours, but the judgment calls are theirs. More importantly, they own the *social* contracts—the unwritten agreements about how people work together, what behavior is acceptable, and what happens when someone crosses a line. Every organization has these social contracts, whether explicit or not. The Integrity Steward makes them explicit. They document expectations. They establish processes for handling violations. They ensure that when consequences are necessary, they’re applied fairly and consistently. This is uncomfortable work. Most people avoid it. They let small violations slide, hoping they won’t repeat. They defer hard conversations until conflicts become crises. They tolerate ambiguity because clarity requires difficult decisions. The Integrity Steward cannot afford this avoidance. Part of their job is to walk toward the tension that others walk away from. --- ## Arbitration and Dispute Resolution When people can’t resolve disagreements themselves, the Integrity Steward arbitrates. This should be seen as an internal failure—a sign that boundaries weren’t clear enough or that something in the system allowed the conflict to escalate. The Integrity Steward resolves the immediate dispute, but more importantly, they learn from it. What ambiguity allowed this to happen? What clearer boundary needs to be set? How do we prevent this class of conflict in the future? Good arbitration isn’t about making everyone happy. It’s about reaching a fair resolution that people can accept, even if they don’t love it. This requires trust—the parties must believe that the Integrity Steward is impartial, informed, and fair. This trust is built over time through consistent behavior. The Integrity Steward who plays favorites, avoids hard calls, or changes their standards based on who’s involved will lose the moral authority the role requires. A critical distinction: a good Integrity Steward makes peace quickly without taking sides. A great Integrity Steward isn’t afraid to take sides when the truth demands it. This might seem contradictory, but it isn’t. Many disputes involve shared blame, and the great Integrity Steward will call out both parties because they value truth, integrity, and honesty over comfortable resolution. They don’t pretend both sides are equally right when they aren’t. They name reality clearly—and then help everyone move forward. People trust this more than they trust false neutrality. They can accept losing a dispute to someone who calls it straight. They cannot accept a system where truth is sacrificed for peace. --- ## The Immune System Framing I described this role as “the immune system of the organism” because that’s exactly what it is. An immune system does three things: it prevents threats from taking hold, it detects problems early, and it corrects dysfunction when it occurs. The Integrity Steward does all three. Prevention means designing systems that make dysfunction unlikely in the first place. Clear contracts. Aligned incentives. Explicit expectations. Good structure prevents most problems before they start. Detection means watching for early signs of trouble. Resentments building. Commitments slipping. Incentives becoming misaligned as circumstances change. The Integrity Steward maintains awareness of organizational health and notices when something is off. Correction means acting when problems emerge. Having hard conversations. Adjusting agreements. Resolving conflicts. Removing people who cannot or will not align. The Integrity Steward doesn’t wait for dysfunction to become crisis. The threats this immune system defends against are varied: bad actors who prioritize themselves over the project, incentive structures that have become counterproductive, legal exposure from unclear agreements, cultural decay as standards slip, personal disputes that poison collaboration, unforeseen consequences of earlier decisions, team absences, people leaving, and the general entropy that affects all human organizations. No system designed at the beginning will remain optimal forever. The Integrity Steward continuously tunes and adjusts, keeping the organization healthy as circumstances evolve. --- ## Culture and Morale This might seem like an odd responsibility for someone who handles contracts and compensation, but it belongs here. The Integrity Steward is responsible for keeping people motivated and feeling good about working on the project. Not through cheerleading or superficial perks, but through the deeper work of ensuring that people feel fairly treated, clearly valued, and genuinely aligned with the mission. This is a softer, more human dimension—not easy to quantify, but essential. When incentives are right, agreements are clear, disputes are handled fairly, and people feel respected, morale takes care of itself. When any of these fail, no amount of team-building exercises will fix it. The Integrity Steward must be a leader in this sense. They explain their reasoning. They help people understand why decisions were made. They convince people that working for the greater good of the project benefits them personally—not through manipulation, but through transparent alignment of interests. This requires partnership with whoever owns strategic clarity. The vision gives people meaning. The Integrity Steward ensures that the structure supports that meaning rather than undermining it. --- ## The Personality That Succeeds This role requires a rare combination: warmth and backbone. The Integrity Steward must be comfortable creating tension—naming problems, setting boundaries, pushing decisions. But they must also be skilled at maintaining connection—preserving respect, trust, and psychological safety even when delivering hard messages. The personality pattern that succeeds usually includes high assertiveness (willingness to speak up, not intimidated by status), high conscientiousness (standards, follow-through, intolerance for vague commitments), high empathy without high people-pleasing (they care, but they don’t melt), and low neuroticism (stays calm when things get heated, doesn’t become reactive or avoidant under stress). The vibe is: calm, direct, fair, and surprisingly warm. Attachment style matters more than most people realize. Those with secure attachment—or who’ve trained themselves into secure behaviors—handle conflict without making it personal and handle closeness without losing boundaries. That’s the “hard conversation plus good feelings” sweet spot. Certain backgrounds tend to produce this skillset: cross-functional leadership roles where ambiguity is high, people management in organizations where feedback is frequent and standards are real, mediation or facilitation work, military or emergency response leadership, community governance, or open-source project maintenance where you must create alignment without formal authority. The real separator isn’t conflict tolerance—it’s conflict skill. Many people tolerate conflict, or even enjoy it, but still damage teams. The ones you want do three things consistently: they separate person from behavior (”You’re valuable; this behavior isn’t working”), they speak in observable facts and impact rather than vibes and blame, and they repair quickly after tension without pretending nothing happened. --- ## What Failure Looks Like The most common failure is inaction. The bad Integrity Steward ignores issues. They aren’t proactive. They actively avoid making course corrections because corrections are uncomfortable. They let small problems become big problems. They defer hard conversations indefinitely. This usually comes from conflict avoidance, but it manifests as neglect. The organization slowly decays while the Integrity Steward hopes things will work themselves out. They never do. Another failure mode is false neutrality. The Integrity Steward who refuses to make judgments, who treats every dispute as equally valid on both sides, who prioritizes peace over truth—this person enables dysfunction rather than correcting it. People learn that bad behavior has no consequences as long as they’re willing to argue about it. Another is rigidity. The Integrity Steward who applies rules mechanically, without understanding context or exercising judgment, will create resentment and drive away good people. Fairness isn’t identical treatment—it’s appropriate treatment given circumstances. And there’s the capture failure: the Integrity Steward who becomes aligned with one faction, who loses the impartiality the role requires, who uses their authority to advance particular interests rather than organizational health. Once trust is lost, the role becomes impossible to perform. --- ## Good Versus Great A good Integrity Steward reacts swiftly to disputes, makes peace quickly, ensures the team is fairly compensated, and maintains clear agreements. They’re reliable. The system works. A great Integrity Steward operates proactively. They don’t wait for conflicts to emerge—they anticipate them. They watch for early signs of misalignment and address them before they become disputes. They regularly open difficult discussions, not because there’s a crisis, but because they know that unspoken tensions accumulate. They aren’t afraid to drill down into any issue, no matter how uncomfortable. They ask the questions others avoid. They surface the problems others hope will disappear. They design systems that prevent dysfunction rather than just correcting it. They think several moves ahead, anticipating how today’s agreements will play out as circumstances change. And they maintain genuine warmth throughout. They’re not cold enforcers—they’re trusted leaders who people seek out when things are hard. People feel better after talking to them, even when the conversation was difficult, because they feel heard, respected, and fairly treated. The great Integrity Steward makes the organization feel safe—not safe from challenge or accountability, but safe to be honest, safe to raise concerns, safe to trust that the system will treat them fairly. --- ## Who Should Consider This Role If you’re reading this and recognizing yourself, here are the signs that Integrity Steward might be your path: You’re drawn to fairness. Inequity bothers you. You notice when systems are imbalanced or when someone is being treated poorly. You’re comfortable with conflict. Not aggressive or combative, but genuinely unbothered by tension. You can have hard conversations without your heart racing. You’re trusted. People tell you things. They seek your counsel in difficult situations. They believe you’ll be fair. You think in systems. You see how incentives shape behavior, how agreements create expectations, how small decisions compound into culture. You’re warm but not soft. You care about people genuinely, but you don’t let that caring prevent you from holding standards. You have backbone. You can say no to powerful people. You can deliver unwelcome messages. You can hold a position when pressured. If this sounds like you, this role needs you. Most organizations suffer from too little integrity enforcement, not too much. The person who can maintain alignment, design fair incentives, and handle conflict with skill and warmth is rare and invaluable. --- ## The Leverage Is Enormous Here’s what makes this role so powerful: everything else depends on it. The clearest strategy fails if incentives push people in different directions. The best product fails if key people leave because they feel unfairly treated. The strongest partnerships fail if agreements are unclear or unenforced. The Integrity Steward is the foundation that makes everything else possible. When they do their job well, the organization can focus on its actual work instead of internal dysfunction. When they do it poorly—or when the role doesn’t exist—entropy wins. In the age of AI, where teams are smaller and more leveraged, this role becomes more critical, not less. There’s no room for unresolved conflicts or misaligned incentives when three people are doing the work of thirty. Every friction point is amplified. Every departure is catastrophic. Every failure of trust is existential. One person who can design fair systems, maintain alignment, enforce accountability, and keep people motivated through difficult times—that person is the difference between an organization that endures and one that falls apart. That’s the Integrity Steward. --- # Skills of the Future, part 1: The Human Toolkit By Filip Blagojević, 2026-02-13. Canonical: https://filipblagojevic.substack.com/p/skills-of-the-future-the-human-toolkit On-site: https://resomnium.com/writing/skills-the-human-toolkit Roles are what you hold. Skills are what you bring. In my Jobs of the Future series, I mapped five roles that will remain fundamentally human as AI transforms organizations—the Clarity Steward, the Execution Orchestrator, the Narrative & Distribution Steward, the Power & Access Steward, and the Integrity Steward. Each represents a function that requires clear ownership and undivided accountability. Each sits above execution on the abstraction ladder, in the territory where judgment meets meaning and systems meet responsibility. But here’s what I didn’t fully address: what makes someone capable of holding those roles in the first place? The roles describe the shape of the work. The skills describe the shape of the person. And as AI continues to absorb execution-level capability at an accelerating rate, the gap between people who can fill these roles and people who cannot is widening into a chasm. This series is about what sits on the human side of that chasm. --- ## The Skill Inversion For most of the twentieth century, the most valuable professional skills were technical. You learned a body of knowledge—law, medicine, engineering, accounting—and that knowledge retained its value for decades. A doctor who graduated in 1970 could practice effectively in 1990 without fundamentally relearning their profession. The half-life of a professional skill was measured in decades. That world is gone. The half-life of technical skills has collapsed to under 2.5 years and continues to shrink. LinkedIn data shows that job skill sets have changed by 25% since 2015 and are projected to be 50% different by 2027. The World Economic Forum estimates 39% of core job skills will change by 2030. Technical knowledge that took years to acquire can now be rendered obsolete by a single model release. Meanwhile, something strange is happening on the other side of the ledger. Durable human skills—judgment, communication, relationship-building, ethical reasoning—have a half-life exceeding 7.5 years and show no signs of compression. Workday’s global research across 22 countries found that the skills deemed least likely to be replaced by AI are also considered the most valuable at work: ethical decision-making, connection and relationship building, emotional intelligence, and conflict resolution. This is the skill inversion. The things we spent centuries treating as “soft”—secondary to real expertise, nice-to-have supplements to technical competence—are becoming the primary source of professional value. And the things we treated as “hard”—the technical knowledge that justified credentials and commanded premiums—are becoming commodities that AI provides on demand. The inversion is not coming. It has already arrived. Most people just haven’t updated their mental model. --- ## Why Skills Matter More Than Roles I could give you the five roles and you could nod along, understanding them intellectually. But understanding a role and being able to perform it are entirely different things. The Clarity Steward needs the judgment to know when to hold a strategic position and when to change it. They need the empathy to translate vision into language that resonates with different people. They need the integrity to create accountability through transparency, including accountability for themselves. The Execution Orchestrator needs the intelligence to understand systems at every level of abstraction. They need the spine to push back when something doesn’t make sense. They need the taste to know when output is good enough and when it’s not. The Integrity Steward needs the emotional depth to understand what actually motivates people. They need the discernment to arbitrate disputes fairly. They need the character that isn’t for sale when pressure mounts. Every role in the framework depends on the same underlying set of human capabilities. The roles are the positions on the field. The skills are whether you can actually play. --- ## What AI Cannot Provide AI systems have reached a remarkable inflection point. They pass professional licensing exams. They write production-quality code. They generate strategic recommendations that sometimes outperform human consultants. A Harvard Business School study found that AI could provide specific, effective business advice—in one case, about chicken farming operations—that was genuinely useful. But the same study revealed something crucial: the humans who benefited most from AI advice were the ones who already had the underlying capability to know which advice to follow. Low-performing entrepreneurs took generic AI suggestions—lower prices, increase advertising—and those suggestions backfired. The AI wasn’t wrong, exactly. It was giving reasonable-sounding advice that didn’t fit the specific context. The humans who lacked the necessary skill couldn’t tell the difference. This pattern appears everywhere AI meets real-world application. A Nature study on healthcare AI found that human-AI teams rarely achieve full complementarity—not because the AI is bad, but because humans struggle to calibrate when to trust and when to override. Research on cognitive offloading shows that people who defer to AI without critical evaluation see their own decision-making abilities decline over time. The risk isn’t that AI replaces human skills. It’s that AI atrophies them. Researchers at the Centre for International Governance Innovation call this “agency decay”—a subtle erosion of human autonomy resulting from overreliance on AI. People shift from experimentation to integration to reliance and, potentially, to addiction. Each step feels like efficiency. Cumulatively, it’s a loss of capability. This means the skills I’m going to explore in this series aren’t just valuable. They’re existentially necessary. Not because AI can’t do them yet—some of them, it’s getting remarkably close to simulating—but because without them, you lose the ability to use AI effectively. You become a passenger in a system you can’t steer. --- ## The Shape of What Makes Someone Irreplaceable I’ve identified a specific set of skills that I believe represent the core of human competitive advantage in the AI era. I’m not going to name them all here. Over the coming weeks, I’ll publish a deep dive on each one: what it actually is, why AI can’t replace it, how to develop it, and how to recognize when it’s present or absent in yourself and others. What I will say is this: these skills share several properties that make them fundamentally different from the technical capabilities AI is absorbing. They are developable but not teachable through instruction alone. You can’t read a book and acquire them. They require practice, exposure, and often failure. This means they resist the kind of rapid scaling that makes technical skills vulnerable—you can’t train a million people in these capabilities the way you can train them in Python. They compound across domains. The person who develops one of these skills in one field carries it into the next. Unlike technical knowledge, which often stays locked to its original context, these skills transfer. They make you more valuable everywhere, not just in the job you currently hold. They interact as a system. None of them works in isolation. One without the others can actually be dangerous—you can probably imagine the damage done by someone who has drive without discernment, or brilliance without the ability to understand other people. The people who hold the durable roles of the future won’t excel at one of these skills. They’ll have developed the full set to a functional level, with particular depth in the ones their specific role demands most. And they become more valuable under pressure. Technical skills degrade when the environment shifts. These skills are what you rely on precisely when everything else is uncertain—when the data doesn’t give you a clear answer, when the team is fractured, when the stakes are real and the path is unclear. --- ## The Developmental Challenge Here’s what makes this both urgent and hopeful: every one of these skills can be developed. They are not personality traits you’re born with. They are capacities built through progressive challenge, deliberate practice, and honest self-assessment. The research is clear on this. Durable human skills respond to development even more reliably than technical skills—and once developed, they transfer across domains, industries, and career transitions. This is the most important career insight of the AI era: invest in the skills that compound across your entire life, not the ones that depreciate with the next software update. But development requires knowing what you’re developing. Most people have a vague sense that “soft skills matter” without any precision about which ones, in what combination, or to what standard. That vagueness is a problem. You can’t deliberately develop what you can’t name. This series will name them. Precisely. One at a time. --- ## The Timeline I want to be honest about uncertainty, as I was in the Jobs of the Future series. These skills appear durable through at least 2040—but that durability depends on assumptions about AI capability growth that may prove wrong. If artificial general intelligence arrives in the early 2030s, even these human skills may face competitive pressure from systems that can simulate them at superhuman levels. But here’s what I believe: even in that scenario, the humans who have developed these skills will be the last to be displaced and the first to find new sources of value. Because these skills don’t just make you useful to an organization. They make you capable of navigating uncertainty itself. And uncertainty is the one thing that isn’t going away. --- ## What Comes Next In the articles that follow, I’ll go deep on each skill I’ve identified. I’ll share the research, the frameworks, the failure modes, and the practical implications for anyone trying to remain valuable—and valuable in a way that matters—as the world reorganizes around artificial intelligence. This isn’t a self-help framework. It’s a map of what makes a human irreplaceable. The future of work isn’t about competing with AI. It’s about becoming the kind of human that AI makes more powerful, not more obsolete. That starts with the toolkit. --- *This is Part 1 of the Skills of the Future series, a companion to the Jobs of the Future series. Next: the one skill that separates people who use AI from people AI uses.* --- # Skills of the Future, part 2: Agency By Filip Blagojević, 2026-02-17. Canonical: https://filipblagojevic.substack.com/p/skills-of-the-future-agency On-site: https://resomnium.com/writing/skills-agency There are two kinds of people in the AI age. The first kind sees a problem and fixes it. They don’t wait for permission. They don’t wait for a meeting. They see the gap, figure out a solution, and move. AI makes them terrifyingly effective because they already knew what they wanted to do. AI just made it faster. The second kind opens ChatGPT and types “what should I do today?” That’s the divide. Not intelligence. Not technical skill. Not credentials. Agency. The ability to act without being told. To generate your own direction. To take ownership of outcomes nobody assigned to you. Everything else I’ll write about in this series builds on this. You can’t develop judgment if you never make decisions. You can’t build integrity if you never take responsibility. Agency is the foundation. Without it, the other skills have nowhere to live. --- ## What agency actually is I want to be precise here because this word gets thrown around loosely. Agency is not ambition. Ambitious people want things. People with agency do things. There’s a massive gap between wanting a promotion and restructuring your workflow so you become indispensable. The first is a desire. The second is agency. Agency is not confidence either. Plenty of confident people wait for direction. They believe they’ll crush it once someone tells them what to crush. People with agency don’t wait for that signal. Here’s my working definition: **agency is the capacity to identify what needs to happen and make it happen, without external prompting.** You see the problem before someone points it out. You start building the solution before anyone asks. Psychologists call this “personal initiative” - going beyond assigned tasks, developing your own goals, solving problems that haven’t occurred yet. Researchers at the University of Sheffield found that proactive, self-directed workers don’t just perform better. They experience their work as more meaningful. They connect with the future through their actions. That last part matters more than you think. --- ## Why AI makes this the number one skill AI is the most powerful tool ever created for people who know what they want to do. And it’s completely useless for people who don’t. AI can write your code, draft your strategy, generate your marketing copy, design your product mockups. It can do all of this in minutes. But it cannot decide what’s worth building. It cannot look at your organization and say “this is broken and here’s why.” That requires a human with direction. In my Jobs of the Future series, I described five roles that remain fundamentally human - roles built around judgment, meaning-making, and accountability. Every single one demands someone who acts first and asks permission later. The Clarity Steward doesn’t wait to be told the strategy is drifting. The Execution Orchestrator doesn’t wait to be told the architecture is wrong. They see it and they move. AI is also flattening organizations. Teams are shrinking. Hierarchies are compressing. Companies are hitting $10M revenue with teams under 10 people. At that scale, everyone is self-directed or the company doesn’t exist. There’s no manager to give you quarterly objectives. You set your own direction or you produce nothing. --- ## The agency decay problem Here’s what scares me. There’s a concept researchers call “agency decay.” It works like muscle atrophy. The less you use your capacity for self-direction, the weaker it gets. And AI makes it very easy to stop using it. The pattern: you start by experimenting with AI. Then you integrate it into your workflow. Then you rely on it. Then you can’t function without it. Each step feels like efficiency. Cumulatively, it’s a loss of capability. A study from SBS Swiss Business School found a strong negative correlation between frequent AI use and critical thinking ability. The neuroscience is straightforward - neural pathways you don’t use get pruned. When AI handles your analysis, your planning, your problem-solving, the brain regions responsible for those functions weaken. This is the paradox. AI amplifies agency. But if you don’t have agency to begin with, AI becomes the thing that destroys what’s left of it. You outsource your thinking until you can’t think anymore. I’ve seen this in organizations already. People who can’t write a paragraph without running it through AI first. People who can’t make a decision without asking a chatbot. People who have basically become prompt-typists - sitting between a manager giving orders and an AI executing them, adding no value in between. --- ## What it looks like in practice A person with agency in a product team doesn’t wait for the product manager to identify bugs. They use the product themselves, notice the rough edges, and either fix them or raise them with specific recommendations. Not “this feels off.” Instead: “the onboarding flow loses 40% of users at step three. Here’s why, and here’s a fix.” A person with agency in their career doesn’t wait for their industry to collapse and then scramble. They learn AI tools now. They build new skills now. They position themselves for what’s coming, not what’s already here. The common thread: they generate action from within. The prompt comes from them, not from someone else. The negative definition is useful too. Lack of agency is the person who says “nobody told me to do that.” The person who waits for the meeting to raise what they’ve known for weeks. The person who complains about the problem without ever attempting a solution. I’m not judging. I’ve been that person at different points in my life. But I am saying clearly: in the AI age, that posture is a death sentence for your career. Not in twenty years. In three to five. --- ## Can you build it? Yes. Full stop. Agency is not a personality trait you’re born with. It’s a capability you develop through practice. It develops through taking progressively larger risks and owning the outcomes. Start small. Identify one problem in your work that nobody asked you to solve. Solve it. Don’t ask permission first. Just do it and present the result. Notice the resistance you feel before doing it - the voice that says “that’s not my job” or “what if I’m wrong?” That resistance is the thing you’re training against. Then go bigger. Propose a project you believe in. Build a prototype of something you think should exist. Push back on a decision you disagree with - not by complaining in private, but by making your case to the person who made it. Each time you do this, you’re strengthening the muscle. And like any muscle, it gets stronger with consistent use. There’s a dark side to acknowledge. Agency without other skills is dangerous. The person who acts decisively but in the wrong direction causes more damage than the person who does nothing. Agency without judgment is recklessness. Agency without empathy is bulldozing. That’s why this is the first skill in the series, not the only one. Agency gives you the capacity to act. The skills that follow determine whether your actions are worth a damn. --- ## How to know if this is you You have a default bias toward action. When you see a problem, your first instinct is to do something about it, not analyze it endlessly. You generate your own projects. You see gaps and fill them without being asked. You’re comfortable being wrong. People without agency often avoid action because they’re terrified of failure. People with agency fail constantly - they just don’t let it stop them. You take responsibility easily. When something goes wrong and you were involved, you own it, learn from it, and move on. If this sounds like you, the AI age is your moment. The leverage available to one person with real agency and AI fluency is unlike anything in human history. And if this doesn’t sound like you yet - that’s fine. It’s developable. Start small. Start now. Don’t wait for someone to assign you the project of developing your own agency. That would sort of defeat the purpose. --- *This is Part 2 of the Skills of the Future series. Next: the skill that keeps everything from falling apart when pressure mounts.* --- # Skills of the Future, part 3: Integrity By Filip Blagojević, 2026-02-19. Canonical: https://filipblagojevic.substack.com/p/skills-of-the-future-integrity On-site: https://resomnium.com/writing/skills-integrity Trust is about to become the scarcest resource in the world. Not talent. Not capital. Not AI capability. Trust. The ability to look at another person and believe that what they say matches what they do. That they won’t change the story when it becomes convenient. That they’ll hold their commitments when holding them gets expensive. We are entering a period where nothing looks real. Deepfakes are indistinguishable from reality. AI-generated content floods every platform. The World Economic Forum ranks misinformation and disinformation as the number one global short-term risk - ahead of armed conflict, ahead of climate. Researchers at UC Berkeley found that people are now just as likely to call something real “fake” as they are to call something fake “real.” In this environment, plenty of people will do what people always do in chaos. They’ll exploit it. They’ll cut corners. They’ll say whatever gets them through the next quarter and worry about consequences later. Some will do it deliberately. Most will drift into it because the pressure is real and the guardrails are gone. Integrity is the skill that keeps you from becoming one of those people. And paradoxically, it’s the skill that makes you most valuable precisely because so many others will abandon it. --- ## What integrity actually is Integrity is not honesty. Honesty is a component, but it’s not the whole thing. Here’s my working definition: **integrity is the consistency between what you believe, what you say, and what you do - maintained over time and under pressure.** That last part is the whole game. Everyone has integrity when things are easy. The test is what happens when integrity becomes costly. When telling the truth means losing a deal. When keeping a commitment means working through the weekend. When admitting a mistake means looking incompetent in front of people whose respect you need. Integrity means your word is your word. Period. Not “my word is my word unless circumstances change.” Not “my word is my word unless I find a better option.” You said you would do it. You do it. If you can’t do it, you say so early, clearly, and you own the consequences. This sounds basic. It is not basic. It is extremely hard to maintain consistently, and most people don’t. --- ## Why the AI era makes this existential Three things are happening simultaneously that make integrity more valuable than it’s been in decades. First, trust in everything is collapsing. Trust in AI dropped from 63% to 56% globally since 2022. Trust in digital content is effectively gone. A Berkeley professor says detecting fakes will become “essentially impossible.” That’s not a future problem. That’s now. Second, chaos creates opportunity for bad actors. Every period of rapid change produces people who exploit confusion for short-term gain. The crypto boom had its scammers. The AI boom has its grifters. Ruthless people move fast when guardrails are down. Third, AI is quietly eroding our capacity for ethical judgment. Researchers call this “moral deskilling.” Same pattern as agency decay. When systems make decisions for us, the neural pathways responsible for ethical reasoning weaken from disuse. A manager who lets algorithms handle hiring stops developing the discernment to assess people. The muscle atrophies. Put these three together and you get a world where trust is scarce, exploitation is common, and the average person’s ability to navigate ethical complexity is declining. Now look at the other side. If you are a person others can rely on - genuinely, consistently, under pressure - you become extraordinarily rare. And rare is valuable. Game theory confirms this. In repeated interactions, reputation compounds. The person known for keeping their word gets opportunities the person known for cutting corners never sees. Short-term cheaters win individual rounds. Long-term players with integrity win the game. --- ## Integrity is not passive Most people think of integrity as a constraint. A list of things you don’t do. Don’t lie. Don’t cheat. Don’t break promises. That’s the passive version, and it’s not enough. Active integrity means proactively creating transparency. It means telling people bad news early instead of hoping the problem resolves itself. It means raising concerns before they become crises. It means documenting your reasoning so others can challenge it. It means admitting what you don’t know instead of performing confidence. In my Jobs of the Future series, I described the Integrity Steward as the immune system of the organization - the person who prevents dysfunction, detects problems early, and corrects them before they compound. That role demands active integrity. But so does every other role in the framework. The Clarity Steward who hides strategic uncertainty loses the team’s faith. The Execution Orchestrator who overpromises timelines destroys trust with every missed deadline. The Power & Access Steward who shades the truth in a pitch meeting might close the deal but poisons the relationship. Integrity isn’t one role’s responsibility. It’s the foundation all of them stand on. --- ## Can you build it? Yes. And I want to be direct about this because there’s a common belief that integrity is something you either have or you don’t. That it’s character, not skill. That you’re born honest or you’re not. That’s wrong. Research on moral development - from Kohlberg’s stages to Lennick and Kiel’s work on moral intelligence - consistently shows that ethical reasoning is ability-based. It develops through progressive challenge and deliberate practice, the same way any other skill does. Here’s how it works in practice. **Start with small commitments and keep them.** This sounds trivial. It’s not. Most people’s integrity erodes not through dramatic betrayals but through a slow accumulation of broken small promises. You said you’d send that email by Friday. You didn’t. You said you’d review that document. You forgot. Each one is tiny. Together, they build a pattern that others notice before you do. Pick commitments deliberately. Make fewer promises. Keep every single one. This is the foundational exercise. **Practice delivering bad news early.** The next time something goes wrong - a deadline slipping, a budget overrun, a mistake you made - tell the relevant people immediately. Not after you’ve tried to fix it quietly. Not after you’ve softened the message. Just tell them. “This happened. Here’s what I know. Here’s what I’m doing about it.” This will feel terrible the first few times. Do it anyway. You’re building the muscle that matters most. **Create accountability for yourself before others create it for you.** Document your decisions and reasoning. Share your work publicly. Ask for feedback on your process, not just your results. Make it easy for others to see where you were wrong. This is the hardest part, and it separates people who have integrity from people who perform it. **Notice when you’re rationalizing.** We all do it. “This situation is different.” “Nobody will notice.” “I’ll make up for it later.” “Everyone else is doing it.” These are the sounds integrity makes when it’s cracking. Learn to hear them. The dark side applies here too, same as with agency. Integrity without judgment is rigidity. The person who keeps a commitment that has become destructive - who won’t adapt because they said they would - isn’t displaying integrity. They’re displaying stubbornness. Integrity means being honest about changed circumstances too. --- ## What failure looks like Integrity failure is almost never dramatic. It’s not the big scandal. It’s the slow drift. You shade the truth slightly in a meeting because the full truth is complicated. You let a small commitment slide because you’re overwhelmed. You avoid a hard conversation because it’s uncomfortable. You take credit for something that wasn’t entirely yours. Each one feels minor. Together, they create a reputation. And reputations, once lost, are nearly impossible to rebuild. Especially in an era where information travels at the speed of a screenshot. The most dangerous failure mode is self-deception. The person who genuinely believes they have integrity while consistently cutting corners. They’ve rationalized so effectively that they can’t see the gap between their self-image and their behavior. If this sounds uncomfortable, good. It should. We’re all closer to this failure mode than we like to admit. --- ## How to know if this is your skill You feel genuine discomfort when you can’t keep a commitment. Not embarrassment at being caught - actual discomfort at the broken promise itself. You default to transparency. When something goes wrong, your instinct is to say so, not to manage the narrative. You hold the same standards for yourself that you hold for others. This one’s rare. Most people have asymmetric integrity - high expectations of others, generous exceptions for themselves. You’ve been tested and held. You’ve had moments where breaking your word would have been easier, more profitable, or more comfortable - and you didn’t. If this sounds like you, the coming years are your competitive advantage. In a world drowning in noise, manipulation, and synthetic everything, the person who is simply, verifiably, consistently real becomes the most valuable person in the room. And if this doesn’t sound like you yet - start building. Small promises kept. Hard truths delivered early. Accountability created voluntarily. The muscle grows the same way every muscle grows. Through use. --- *This is Part 3 of the Skills of the Future series. Next: the skill that separates the people who understand from the people who merely know.* --- # Skills of the Future, part 4: Intelligence By Filip Blagojević, 2026-02-24. Canonical: https://filipblagojevic.substack.com/p/skills-of-the-future-intelligence On-site: https://resomnium.com/writing/skills-intelligence Intelligence is not what you think it is. Most people hear “intelligence” and picture analytical horsepower. Logic. Pattern recognition. The ability to solve hard problems quickly. That’s part of it. But it’s the part AI already does better than you. The intelligence that matters now - the kind that keeps you valuable - is something different. It’s the ability to think clearly across domains. Analytical, emotional, social, practical. Not mastering one. Moving fluidly between all of them and integrating what you find into something coherent. AI can analyze data faster than any human alive. It can detect sentiment in text. It can generate strategies that sound plausible. What it cannot do is sit in a room where the data says one thing, the team feels another, the client wants a third, and the market is shifting beneath all of it - and make sense of the whole picture. That integration is the skill. And almost nobody is training for it. --- ## The Simulation Problem Here’s what makes this moment dangerous. AI now produces outputs that look like every type of intelligence. It writes analytically. It responds with apparent empathy. It generates strategic recommendations that read like wisdom. Judge by output alone and AI looks intelligent across every dimension. But simulation is not understanding. Researchers at Emory coined a term for this: epistemia - the illusion of understanding produced when AI’s probabilistic output mimics deeper reflection. The AI generates plausible patterns. The human generates meaning. One optimizes for coherence. The other searches for truth. When the simulation is good enough, people stop noticing the difference. And when they stop noticing, they stop developing the real thing. A 2025 study found a significant negative correlation between frequent AI usage and critical thinking scores. The mechanism is cognitive offloading - delegating hard thinking to the tool until the muscle atrophies. Same pattern as agency decay, same pattern as moral deskilling. What gets short-circuited is not intelligence. It’s the process of becoming intelligent about something. The struggle of working through confusion, sitting with ambiguity, constructing understanding from scratch - that’s where intelligence develops. Skip it enough times and you lose the ability to do it at all. --- ## Intelligence Is Not One Thing We’ve been trained to think of intelligence as a single score. IQ. SAT. GPA. One number that captures how smart you are. This is wrong, and it’s been wrong for decades. Intelligence operates across multiple dimensions. Some of them overlap. Some of them conflict. The ones that matter most in the AI era are the ones that require you to operate across several dimensions simultaneously. Analytical intelligence is the classic: logical reasoning, pattern recognition, mathematical thinking. AI dominates here. If this is all you bring, you’re already competing with something faster and cheaper. Emotional intelligence is the ability to understand your own emotional states and those of others - and to use that understanding to navigate social situations. AI can simulate empathetic responses. It cannot feel the tension in a room, read the thing someone isn’t saying, or sense that a team member is about to quit before any data suggests it. Social intelligence extends beyond emotions into understanding power dynamics, cultural context, group behavior, and the unspoken rules that govern how humans actually cooperate. AI doesn’t operate in social environments. It processes text about them. Practical intelligence is what some call street smarts - the ability to solve messy, real-world problems that don’t have clean answers. The kind of thinking that helps you navigate a negotiation where every party has hidden motivations and the formal terms don’t capture what’s actually being agreed to. Each of these is valuable. But here’s my actual point: none of them is the skill. The skill is integration. --- ## The Integration Edge AI can do each type of intelligence in isolation. Give it an analytical problem, it performs well. Ask it to generate emotionally aware language, it produces something passable. Request a strategic recommendation, it delivers one. What AI cannot do is weave them together in real time, grounded in lived context, while adjusting for things it cannot observe. Think about what a Clarity Steward actually does - the role I described in the Jobs of the Future series as the guardian of strategic coherence. They need analytical intelligence to evaluate whether a strategy makes sense. They need emotional intelligence to understand what each team member actually cares about. They need social intelligence to navigate the politics that determine whether a strategy survives contact with the organization. And they need practical intelligence to know whether the plan is actually executable. Not one of these. All of them. At the same time. Synthesized into a single coherent read of the situation. This is what I mean by intelligence as a skill. It’s not about being the smartest person in the room on any single axis. It’s about being the person who can hold multiple types of knowing simultaneously and act on the integrated picture. AI processes inputs sequentially or in parallel, but within a defined frame. It cannot step outside its own frame. Humans can. We notice that the analysis is correct but the timing is wrong because we feel the team’s exhaustion. We recognize that the strategy is sound but the relationship with the partner won’t survive the execution. We understand that the numbers work but the culture won’t support it. This cross-domain synthesis - the constant movement between analytical and emotional, between social and practical - is where human intelligence becomes irreplaceable. Not because AI can’t touch each domain, but because it can’t hold them all at once and navigate the contradictions between them. --- ## Why It’s Getting Harder The cognitive offloading problem hits intelligence harder than any other skill. When you let AI do your analysis, you lose analytical sharpness. When you let AI draft emotionally sensitive communications, you stop developing emotional fluency. When AI summarizes meeting dynamics, you stop reading rooms yourself. Each delegation seems harmless. But intelligence isn’t a static asset. It’s a capacity that strengthens or weakens based on use. Every time you skip the hard thinking - the confusion, the ambiguity, the slow wrestling with problems that don’t have obvious answers - you lose a little of what makes you valuable. The insidious part is that it feels like you’re getting smarter. Better outputs. Higher volume. You look more intelligent. But the underlying capability is eroding. You’re borrowing intelligence instead of building it. The 82% of workers who told Workday researchers they’ll crave more human connection as AI integrates deeper - they’re sensing something real. Something is being lost even as productivity rises. The something is depth. The capacity to truly understand rather than merely know. --- ## How to Build It Intelligence across domains is trainable. It requires deliberate, uncomfortable practice. **Think in a different domain every week.** If you’re technical, spend time understanding what motivates the people around you. If you’re naturally empathetic, force yourself through a rigorous analytical exercise. The point isn’t mastery. It’s building the pathways that let you move between modes of thinking. Cross-domain learning activates distant neural circuits. That activation is where original insight comes from. **Stop outsourcing your confusion.** When you hit a problem you don’t understand, resist the urge to immediately ask AI. Sit with it. Work through it yourself first. The struggle is where intelligence forms. Use AI afterward to check your thinking, not to replace it. **Practice real-time integration.** In every meeting, every decision - ask yourself: what does the analysis say? What do the people feel? What are the power dynamics? What’s practically possible? Hold all four simultaneously. Most people never even try. **Develop your weakest type.** Whatever dimension you naturally avoid - that’s your growth edge. The integrated thinker doesn’t need to be exceptional at every type. They need to be functional across all of them. Your weakest link limits your synthesis capacity. --- ## Failure Modes Intelligence fails in predictable ways. The most common is lopsidedness. The analytically brilliant person who is emotionally oblivious makes decisions that are technically correct and humanly catastrophic. The emotionally attuned person who is analytically weak makes decisions that feel good and don’t work. Another is intellectual arrogance. The conviction that your type of intelligence is the only type that matters. I see this constantly in technical people - the belief that analysis is real thinking and everything else is soft nonsense. I’ve been guilty of it myself. It’s a trap that gets you outmaneuvered by people who understand dimensions you dismiss. And there’s borrowed intelligence. The person who appears smart because their AI-augmented outputs are excellent, but whose underlying capacity is hollowing out. They perform well until they face a novel situation where AI has no good answers - and discover they can’t think their way through it alone. --- ## The Self-Assessment You have real intelligence - the durable kind - if you can navigate situations where different types of intelligence give you conflicting signals and still act coherently. If you can hold analytical, emotional, social, and practical considerations simultaneously without collapsing into one. You’re at risk if you rely on one type of intelligence and dismiss the others. If you consistently outsource your hard thinking to AI. If you perform well with AI support but feel lost without it. The test is simple: when was the last time you sat with a genuinely confusing problem and thought your way through it from scratch? Not prompted an AI. Not Googled the answer. Actually thought. If you can’t remember, that’s your signal. --- *This is Part 4 of the Skills of the Future series, a companion to the Jobs of the Future series. Next: the skill that tells you which of AI’s thousand options is the right one.* --- # Skills of the Future, part 5: Taste By Filip Blagojević, 2026-02-26. Canonical: https://filipblagojevic.substack.com/p/skills-of-the-future-taste On-site: https://resomnium.com/writing/skills-taste Taste is not about aesthetics. It’s not about knowing which wine pairs with which dish or having opinions about furniture. Those are symptoms. The actual skill is deeper. Taste is the ability to discern what matters from what doesn’t. In a world of infinite options, it’s knowing which one to choose - and more importantly, which thousand to ignore. This has always been useful. It’s about to become essential. When AI can generate anything - code, content, strategies, designs, entire products - the bottleneck is no longer creation. It’s selection. The person who can look at a thousand AI-generated options and immediately feel which one is right has a skill that no algorithm can replicate. Because the algorithm generated all thousand. It has no preference. It has no sense of what matters. That’s your job. And here’s the part most people miss: taste isn’t just professional. It’s personal. The choices you make about how to live - what to pay attention to, what to ignore, how to spend your time, what to believe, who to trust - these are all expressions of taste. In the age of AI, those choices define you more than your skills ever will. --- ## The Overabundance Problem We used to live in a world of scarcity. Information was hard to find. Products were expensive to build. Content took effort to create. The challenge was access - getting enough of the right inputs to make good decisions. That world is gone. Now the problem is the opposite. There is too much of everything. Too much information. Too much content. Too many tools, frameworks, opinions, takes, strategies, and options. Every platform, every scroll, every notification is competing for the same finite resource: your attention. AI made this exponentially worse. When the cost of creating something approaches zero, the volume of creation explodes. Most of it is noise. Competent, polished noise - but noise. The signal-to-noise ratio is collapsing across every domain. In this environment, the old metrics of intelligence don’t help you. Knowing more doesn’t make you better. Processing faster doesn’t give you an edge. The edge comes from knowing what to ignore. From being able to walk through a flood of information and pick up the one thing that actually matters. That’s taste. --- ## Taste Is Identity Here’s where this gets philosophical, and I make no apology for it. Your taste is the sum of your choices. Not just the big ones - career, relationships, where to live. The small ones too. What you read. What you skip. What catches your attention. What bores you. What you think is beautiful and what you think is garbage. These choices, accumulated over a lifetime, form a pattern. That pattern is you. Philosopher Friedrich Schiller argued that aesthetic taste forms the core of identity - it shapes how you see yourself in the world. Recent research confirms this intuitively: people perceive changes in someone’s taste as genuinely identity-altering. If you switch from loving classical music to loving pop, people don’t just think your preferences changed. They think you changed. This matters because real taste - the kind that’s actually valuable - is purely individual. It cannot be crowdsourced. It cannot be optimized by committee. The moment you start choosing things because everyone else chose them, you’ve lost it. You’re running on consensus, not taste. True taste will get you weird looks. It has to. If every choice you make is legible and approved by your peer group, you’re not exercising taste. You’re performing conformity. The person with genuine taste makes choices others don’t understand - until, sometimes, they do. And sometimes they never do. That has to be okay. Bourdieu argued that taste is social - shaped by class and culture, not individual at all. He’s partially right. Your starting point is shaped by where you come from. But where you go from there is yours. The skill isn’t having taste handed to you. It’s developing it through deliberate, conscious choices that are yours alone. --- ## Taste in the AI Era In the age of AI, taste expresses itself through a new set of choices that most people aren’t thinking about carefully enough. How will you use AI? Will you use it at all? For what? To what extent? Will you let it draft your thinking or only check it? Will you use it for creative work or keep that human? Which model will you use? Does it matter to you if it’s open-source? Do you care where your data lives? These aren’t technical questions. They’re taste questions. They reveal what you value, what you’re willing to trade, and how you think about the relationship between human effort and machine output. The person who lets AI do everything without thinking about it has no taste. They’ve abdicated the most important choices of their professional life to convenience. The person who refuses AI entirely out of principle also has no taste - they’ve substituted ideology for discernment. Taste is in the middle. It’s the conscious, thoughtful, sometimes uncomfortable process of deciding: here, AI helps me. Here, it doesn’t. Here, the human effort matters. Here, it’s wasted. These boundaries are different for everyone. They should be. That’s the point. The Narrative & Distribution Steward from the Jobs of the Future series is a vivid example - but taste is essential for every role in that framework. The Clarity Steward needs taste to know which strategic direction is worth pursuing. The Execution Orchestrator needs it to know when output is good enough. The Integrity Steward needs it to sense when a system is subtly wrong before the data confirms it. The Narrative Steward just makes the case most visible. AI can generate unlimited content. Every company can produce professional marketing materials in minutes. The noise is deafening. The Steward’s entire job is cutting through it - knowing which story is real, which angle resonates, which channel matters. That’s taste applied to narrative. Without it, you’re just adding to the flood. --- ## How It Develops Taste isn’t innate. But it can’t be taught through instruction either. It develops through conscious choice-making. Every time you encounter options and deliberately choose - not defaulting, not following the crowd, not optimizing for what’s easy - you strengthen the muscle. Every time you think about why you chose what you chose, you refine it. It develops through exposure. The person with broad, deep exposure to different domains, cultures, ideas, and perspectives develops richer taste than the person who stays in one lane. This is why the intelligence article before this one matters - cross-domain thinking feeds taste. The wider your inputs, the more refined your filter becomes. It develops through living in the moment. This sounds soft, but it isn’t. Taste requires presence. You can’t discern what matters if you’re not paying attention. The person scrolling mindlessly is not developing taste. The person who pauses, notices, and asks themselves “do I actually care about this?” is. It develops through taking life seriously enough to contemplate it. Not grimly serious. But seriously enough to believe that your choices matter. That how you spend your attention is a real decision with real consequences. That the difference between a life lived with taste and a life lived on autopilot is the difference between meaning and noise. And it develops through getting it wrong. Taste without failure is just luck. The choices that turned out poorly, the instincts that were off, the times you committed to something that didn’t work - those are the experiences that sharpen discernment. You learn what matters partly by discovering what doesn’t. --- ## What Failure Looks Like The most common failure is defaulting. Choosing whatever is popular, whatever is recommended, whatever everyone else is doing. This isn’t taste. It’s the absence of it. And in a world where AI can optimize for popularity, defaulting makes you indistinguishable from an algorithm. Another failure is rigidity. The person who found their taste at 25 and never updated it. Taste that doesn’t evolve becomes dogma. The world changes. What mattered yesterday might not matter tomorrow. Taste must be alive - constantly tested, occasionally revised, never frozen. Another is performative taste. Choosing things not because you actually value them but because they signal sophistication or belonging. This is Bourdieu’s trap - using taste as social currency rather than genuine discernment. People can feel the difference. Performative taste is brittle. Genuine taste is magnetic. And there’s overthinking - analyzing every choice until you can’t make any. Taste requires some courage. Sometimes “because it feels right” is a valid reason, even in a data-driven world. The person who needs a spreadsheet to justify every decision has outsourced their taste to metrics. --- ## The Self-Assessment You have taste if your choices form a coherent pattern that’s recognizably yours - not a copy of someone else’s, not an optimization for approval, not a default. You have taste if you can walk into a room full of options and feel drawn to the right one before you can fully explain why - and then explain it when asked. You have taste if some of your choices confuse other people, and you’re comfortable with that. You’re at risk if you can’t remember the last time you chose something that surprised even you. If every decision feels safe and validated. If your life looks optimized but not yours. The test: take away the algorithms, the recommendations, the social proof. What would you choose? If you don’t know, that’s where the work begins. --- *This is Part 5 of the Skills of the Future series, a companion to the Jobs of the Future series. With this article, the series is complete - for now. As the landscape shifts and new patterns emerge, I’ll add new skills when they earn their place here.* --- # Company of the Future, part 1: How Organized Effort Looks Like Post AI By Filip Blagojević, 2026-03-03. Canonical: https://filipblagojevic.substack.com/p/company-of-the-future-how-organized On-site: https://resomnium.com/writing/company-how-organized-effort-looks-like-post-ai Everyone is talking about what AI builds. Almost no one is talking about how we organize to build it. This is the bigger problem. And almost nobody sees it yet. I’ve spent time in industrial automation. I watched factories go from hundreds of workers on a line to a handful of operators managing robotic systems. The pattern was always the same. First, the machines replaced individual tasks. Then they replaced entire processes. Then someone looked around and realized the org chart made no sense anymore. All those layers of supervision, all those coordination roles, all those managers managing managers - they existed to solve a problem that the machines had already solved. The factories that adapted fastest didn’t just buy better robots. They redesigned how humans were organized around those robots. The ones that didn’t adapt kept their old structures, added machines on top, and created expensive chaos. That exact pattern is now playing out in knowledge work. Except it’s happening in years, not decades. --- ## The execution problem is solved AI can write code. It can draft legal documents. It can build financial models, create marketing campaigns, generate research syntheses, and produce first drafts of almost anything a knowledge worker gets paid to do. GitHub reports developers merged 43 million pull requests per month in 2025, up 23% from the prior year. A quarter of current Y Combinator startups have 95% of their code written by AI. Companies are hitting $10M in revenue with teams under 10 people. The execution bottleneck is gone. Or at least, it’s collapsing so fast that planning around it is a mistake. Brynjolfsson’s research showed a 14% average productivity increase in call centers using AI, with 34% improvement for novice workers. Harvard found a 40% time reduction in writing tasks. These aren’t projections. These are measured results from the last two years. And the tools are getting better every quarter. So what’s the new bottleneck? Coordination. Judgment. Accountability. Deciding what to build, why to build it, whether it’s working, and who is responsible when it isn’t. These are organizational problems. And most organizations are still structured to solve an execution problem that no longer exists. --- ## The org chart is a fossil Think about what a traditional company structure actually is. It’s a system designed to coordinate human labor at scale. You have departments because specialization made individual humans more productive. You have middle management because information needed to flow up and decisions needed to flow down. You have hierarchies because someone had to ensure that hundreds or thousands of people were pulling in roughly the same direction. Every layer exists to solve a coordination challenge created by having lots of humans doing lots of tasks. Now remove the need for most of those humans to do most of those tasks. What happens to the layers? They become friction. They slow things down. They add cost without adding judgment. They create meetings about meetings. They diffuse responsibility until nobody owns anything. I’ve seen this in real organizations. A team of 30 people where maybe 5 are doing the actual work and 25 are coordinating, reviewing, approving, and reporting on that work. Remove the execution bottleneck with AI and what do you have? Five people orchestrating AI systems and 25 people with nothing to coordinate. That’s not a restructuring problem. That’s an existential one. Shopify’s CEO made headlines in 2025 by telling managers they had to prove AI couldn’t do a job before hiring a human. That got attention because it sounds bold. But it’s actually a half-measure. It optimizes headcount within the old structure. It doesn’t question whether the structure itself still makes sense. And it doesn’t. --- ## What actually needs to change Most people think the shift is from “big teams” to “small teams.” That misses the point entirely. The real shift is from organizations designed around executing tasks to organizations designed around exercising judgment. When AI handles execution, the humans who remain need to do the things AI cannot: set direction, make judgment calls under uncertainty, maintain accountability, build trust-based relationships, and keep the whole system honest. These functions don’t need departments. They don’t need middle management. They don’t need org charts with dotted lines and matrix reporting. They need clear ownership. One person, one function, full accountability. My strong opinion after years of thinking about this - studying the research, building organizations, watching what works and what doesn’t - is that the company of the future has a specific shape. It’s small. It’s flat. Every human in it holds real authority over a defined domain. And it scales not by adding people, but by adding AI leverage underneath those people. I call this model the Cell. I’m not going to fully explain it here. That’s coming in the next article. But I want to lay out the principles that any durable post-AI organizational model must satisfy, because these are non-negotiable regardless of what you call the structure. **Individual accountability, not shared responsibility.** Every critical function needs one person who owns it completely. When everyone is responsible, no one is responsible. This is true in a team of 5 and true in a company of 5,000. AI makes it more urgent because the speed of execution means mistakes compound faster than ever. **Human judgment at the top of every decision chain.** AI executes. Humans decide. Not because AI can’t make decisions - it increasingly can - but because someone has to be accountable for what those decisions produce. Legal liability, ethical responsibility, and stakeholder trust all require a human at the top. **Small enough for trust, large enough for challenge.** A single person can’t challenge their own thinking. A team of fifty can’t maintain the trust and speed required. The sweet spot is somewhere in between - small enough that everyone knows everyone, large enough that genuine disagreement is possible. **Generalists over specialists.** When AI can provide specialist knowledge on demand, the premium shifts to people who can synthesize across domains. The most valuable humans in the AI era are the ones who understand enough about everything to direct AI systems intelligently and catch their failures. **Designed for leverage, not headcount.** The measure of organizational capacity is no longer how many people you have. It’s how much each person can direct. One skilled human orchestrating AI agents and automated systems can produce what a traditional team of 10 or 20 produced. The organization should be designed around that reality. --- ## The timeline I want to be honest about what I know and what I don’t. I believe these principles are valid for the period between now and roughly 2040. That’s the window where AI capabilities are growing fast enough to transform knowledge work but haven’t yet reached the point where they can fully substitute for human judgment across all domains. What happens after 2040 depends on questions nobody can answer yet. If artificial general intelligence arrives on the timelines some researchers predict - and those timelines keep getting shorter - then even these organizational principles may need rethinking. The median expert prediction for AGI has compressed from 50 years in 2020 to roughly 5-7 years on current forecasting platforms. That should make everyone uncomfortable, including me. But here’s what I know from watching automation transform industries for years: the organizations that fail are never the ones that couldn’t adopt the new technology. They’re the ones that couldn’t reorganize around it. The technology is the easy part. The human structure is the hard part. Most companies right now are bolting AI onto their existing structures and hoping for the best. McKinsey found that 95% of generative AI pilots fail to deliver meaningful business impact. That number should terrify executives, but it shouldn’t surprise anyone. You can’t solve an organizational problem with a technology purchase. The AI works fine. The structure around it was built for a different world. --- ## What comes next This is the first article in the Company of the Future series. In my Jobs of the Future series, I mapped five roles that remain fundamentally human as AI transforms work. In my Skills of the Future series, I explored the capabilities those roles demand. This series completes the picture: how do you actually organize those people into a structure that works? Over the coming weeks, I’ll go deep on the organizational model I’ve been developing. I’ll start with the Cell - the core unit of the post-AI organization. Then I’ll show how it works at different scales: the single operator, the triad, the full five-person Cell. And I’ll address the hardest question of all - how existing organizations can evolve toward this model without burning everything down. This is not theory. This is a framework built from real experience designing organizations in high-complexity environments, validated against the best available research, and stress-tested against the reality of how AI is actually changing work right now. The product problem is solved. AI builds things. The organization problem is wide open. That’s where the real work begins. --- *This is Part 1 of the Company of the Future series. Next: the Cell - the smallest viable unit of human organization in the AI age.* --- # Company of the Future, part 2: The Cell By Filip Blagojević, 2026-03-05. Canonical: https://filipblagojevic.substack.com/p/company-of-the-future-the-cell On-site: https://resomnium.com/writing/company-the-cell Roles don’t work alone. In my Jobs of the Future series, I mapped five roles that remain fundamentally human as AI absorbs execution-level work. Each article went deep on a single function. The response was strong. People recognized themselves. They saw where they fit. But here’s what I didn’t address: none of those roles works in isolation. A brilliant strategist without someone to execute is a philosopher. A great executor without strategic clarity is building the wrong thing fast. A storyteller without substance to tell stories about is a bullshit artist. The roles are organs. Individually, they’re interesting. Together, they form an organism. I call that organism the Cell. --- ## What the Cell actually is The Cell is the smallest viable unit of human organization that can function with full accountability in the AI age. It’s not a team. It’s not a department. It’s a self-contained organism where every critical human function has a dedicated owner with real authority. Five core functions that must be covered. One or two flexible seats for mission-specific expertise when the Cell operates at full size. Five to seven people maximum in its ideal form - though it can run with fewer. Each person holds a defined domain with final authority over that domain. No committees. No shared ownership. No “we’ll figure out who decides later.” Every function has a name attached to it. The Cell scales not by adding humans, but by adding AI leverage underneath each function. One Execution Orchestrator with the right AI toolkit can deliver what a traditional team of 10 produced. The humans stay small. The output stays massive. This is the core idea. A fixed nucleus of high-level generalists, each owning a critical control plane, leveraging AI to multiply their capacity by orders of magnitude. --- ## The five core functions These are non-negotiable. Every Cell must have them covered. Not necessarily by five different people - but covered, with clear ownership, no matter how many humans are in the room. **The Clarity Steward** owns strategic coherence - the “why” behind everything the Cell does. They ensure everyone understands the vision, can connect their work to it, and that the strategy doesn’t drift or dilute as it meets reality. **The Execution Orchestrator** turns vision into reality by orchestrating AI agents, automated systems, and selective human support. They own delivery - timelines, quality, tooling, and the judgment calls about what gets automated versus what needs a human touch. **The Narrative & Distribution Steward** shapes how the world perceives what the Cell builds. They own the story, the brand, the channels, and the emotional connection between product and market. They don’t do “marketing things.” They make people feel something real. **The Power & Access Steward** opens doors that would otherwise stay closed. They handle the highest-stakes external relationships - top-tier customers, strategic partners, capital sources. The conversations that require a human with real authority sitting across the table. **The Integrity Steward** is the immune system of the organism. They own governance, incentive design, conflict resolution, and accountability enforcement. They keep the Cell honest, fairly compensated, legally sound, and culturally healthy. Without this function, every other function eventually breaks down. If you’ve read the Jobs of the Future series, these are familiar. If you haven’t, each one has a dedicated deep-dive article you can find on my Substack. The ideal Cell has one person per function. Five dedicated owners, each with undivided focus on their domain. But in practice, a Cell can operate with fewer people holding multiple functions. A solo founder holds all five. A triad of three covers them between its members. The functions don’t change. The distribution of who carries them does. What matters is understanding why these five functions, and not four, or six, or ten. --- ## Why these five Every durable organization needs to answer five questions. Who decides what we’re doing and why? Who actually builds it? Who tells the world about it? Who brings in the relationships and resources we can’t reach on our own? And who keeps the whole thing from falling apart? Strategy. Execution. Narrative. Access. Integrity. Remove any one and the organism fails. Merge two into one seat and that person becomes a bottleneck with divided attention on functions that deserve undivided focus. Add more for functions that AI can handle and you’ve added humans where you don’t need them. Five is the minimum for full coverage. Five is the maximum before you start duplicating or diluting authority. I didn’t arrive at this number by theory. I arrived at it by studying what actually breaks. When strategy and execution live in the same person, strategy always loses because execution feels more urgent. When integrity doesn’t have a dedicated owner, small problems fester until they become existential. When narrative is an afterthought, the market never understands what you’ve built. --- ## The flexible seats Beyond the five core functions, a full-size Cell has room for one or two flexible positions. These are permanent members - not contractors, not part-timers - but their anchor domain depends on the mission. A deep technical architect for a technology-heavy Cell. A financial engineer for a crypto project. A regulatory expert for a heavily regulated industry. Strict rules apply. Flexible seat holders must still meet the generalist bar. They participate fully in strategy, narrative, distribution, and governance discussions. Their specialization is an edge, not their identity. If someone can only contribute within their narrow domain, they belong on the contractor bench, not in the Cell. This preserves cognitive diversity without silos. Everyone sees the whole picture. Everyone can challenge everyone else. No one hides behind “that’s not my department.” --- ## The generalist requirement This is the part that makes people uncomfortable. Every person in the Cell must be a high-level generalist. Not a dabbler. Not someone who “knows a little about everything” in a superficial way. A generalist in the Cell means someone who has enough depth across multiple domains to direct AI systems intelligently, evaluate outputs critically, and engage meaningfully in every strategic conversation at the table. In practice, this means every member has a primary domain they own, but they can hold their own in any other domain’s discussions. The Execution Orchestrator can challenge the Narrative Steward’s positioning. The Integrity Steward can question strategic direction. The Clarity Steward can push back on execution timelines. When the whole Cell can pressure-test a decision from five different angles simultaneously, the decision quality is dramatically higher than anything a siloed organization can produce. Here’s the honest truth about how this works in practice: the higher the quality of the generalists, and the higher the standard they hold each other to, the better the Cell performs. A Cell of five exceptional generalists who challenge each other relentlessly will outperform any traditional organization many times its size. A Cell of five average performers who avoid hard conversations will produce average results with a fancy structure on top. The Cell is a multiplier. Put in great people who push each other, and it multiplies excellence. Put in mediocre people who play nice, and it multiplies mediocrity. The framework doesn’t do the work. The people do. --- ## How accountability works The Cell has a multi-layer accountability system. I’m going to outline the principles here and go deeper on the mechanics in a later article when I describe the full five-person Cell in operation. **Layer one: functional accountability.** Each steward has final authority in their domain and full accountability for outcomes. Narrative fails? The Narrative Steward owns that. Execution slips? The Execution Orchestrator owns that. No diffusion of responsibility. No “the team decided.” **Layer two: peer accountability.** Every member has both the right and the obligation to challenge any other member. This is institutionalized, not informal. Any member can trigger a full Cell review - a “red flag” pause - if they see integrity risk, strategy drift, or incentive corruption. No hierarchy can block this. **Layer three: integrity enforcement.** The Integrity Steward has special powers - incentive arbitration, behavioral enforcement, and the ability to initiate removal proceedings. Their decisions are binding unless overruled by supermajority. This prevents the slow cultural decay that kills organizations over time. And there’s a final layer most organizations are too afraid to define: removal. Grounds exist. The process is clear. Unanimous consent for admission, supermajority for removal. No ambiguity. No hostage situations. This system works because it assumes something uncomfortable: that good people can still act badly when incentives push them wrong, when accountability is unclear, or when nobody is watching. The Cell doesn’t rely on everyone being virtuous. It creates conditions where virtue is the easiest path. --- ## What makes this different Most organizational frameworks in the AI discourse focus on technology integration. Microsoft’s “Frontier Firm.” McKinsey’s “agentic organization.” They all start with the question: how do we plug AI into the existing model? The Cell starts somewhere else: what is the minimum viable human structure for judgment, accountability, and trust - and how do we build AI leverage underneath it? The technology is a given. AI handles execution. The unsolved part is the human structure. Who decides. Who owns. Who is accountable when things go wrong. Who keeps the system honest over time. --- ## The spectrum in practice Most people reading this are not going to assemble a full five-person Cell tomorrow. Some of you are solo operators right now, holding all five functions yourself. Some are in partnerships or small teams of two or three. Some run larger organizations wondering how any of this applies to them. All of those situations are valid starting points. The Cell framework scales across configurations, with each one having different strengths, limits, and failure modes. The single player has maximum speed but no one to challenge their judgment. The triad has the first real checks and balances but members carry stretched attention across multiple functions. The full Cell of five has complete coverage with one person per function - the ideal, but also the hardest to assemble. Over the next few articles, I’ll walk through each configuration. How it works. Where it breaks. Who it’s for. Starting with the loneliest and most common one: the single player. --- *This is Part 2 of the Company of the Future series. Next: the single player - what happens when one person tries to hold all five functions, and why AI doesn’t solve the problem you think it does.* --- # Company of the Future, part 3: Single Player By Filip Blagojević, 2026-03-10. Canonical: https://filipblagojevic.substack.com/p/company-of-the-future-single-player On-site: https://resomnium.com/writing/company-single-player The most common Cell configuration is also the most fragile. One person. All five functions. No one to challenge, no one to check, no one to share the load. Just a single operator leveraging AI to do work that used to require teams. This is where most people start. The freelancer, the solo founder, the independent consultant, the creator building their thing on nights and weekends. Tech CEOs love to predict that AI will create one-person billion-dollar companies. Maybe. But that’s not the interesting question. The interesting question is what happens to a human being when they try to hold strategy, execution, narrative, access, and integrity all at once - with no one watching. The answer is: some things work incredibly well. And some things break in ways you don’t notice until it’s too late. --- ## The single player’s superpower Speed. No alignment meetings. No consensus-building. No waiting for someone else to understand your vision. You think it, you decide it, you build it. The feedback loop between idea and execution is as tight as it can possibly be. With AI tools, this speed becomes extraordinary. A single player who has mastered AI-augmented execution can ship products, create content, manage operations, and handle customer relationships at a pace that would have required a team of ten just five years ago. The leverage is real. And for certain kinds of work - creative projects, early-stage products, consulting practices, content businesses - the single player model is not just viable. It’s optimal. The overhead of coordination with other humans can actually slow you down when the work is clear and the direction is set. This is worth acknowledging honestly. Not every project needs a team. Not every person needs collaborators. Some work is genuinely better done alone, at speed, with AI handling the parts that don’t require human judgment. The problems start when the work grows beyond what one brain can reliably oversee. --- ## Failure mode one: blind spots compound Every person has blind spots. Areas where their judgment is weak, where their experience is thin, where their biases go unchallenged. In a team, other people catch these. In the single player configuration, nobody does. The Execution Orchestrator who is also their own Clarity Steward will naturally gravitate toward building. Strategy becomes whatever justifies the thing they already want to make. The direction feels clear because no one is questioning it. The Narrative Steward who is also their own Power & Access Steward will tell stories that resonate with them personally and assume the market agrees. Without someone pushing back on positioning, they optimize for self-expression rather than market reality. AI doesn’t fix this. AI is excellent at executing within the frame you give it. It’s terrible at telling you the frame is wrong. You can ask an AI to critique your strategy, but the quality of that critique depends on how well you framed the question - which is exactly the skill your blind spot prevents you from exercising. The result is slow drift. Not dramatic failure. The single player doesn’t crash. They gradually wander off course in ways they can’t see, because seeing them requires a perspective they don’t have. --- ## Failure mode two: standards slip Accountability requires an audience. This is uncomfortable to admit, but it’s true. People maintain higher standards when someone else is watching. Not because they’re lazy or dishonest, but because human psychology is wired for social calibration. We sharpen our thinking when we know it will be challenged. We polish our work when we know it will be evaluated. We hold ourselves to commitments more firmly when someone else is tracking them. The single player has no audience. They are their own Integrity Steward - designing their own incentives, setting their own standards, evaluating their own performance. This is like being your own referee. Even with the best intentions, the calls start going your way. It shows up in small ways first. The deadline that slips because no one will notice. The quality standard that drops because “good enough” has no external benchmark. The strategic review that doesn’t happen because it feels unnecessary when you’re the only one who would attend. Over time, these small slips compound. The single player who started with high standards gradually settles into whatever level of performance they can sustain without external pressure. For some people, that level stays high. For most, it drifts downward. This isn’t a character flaw. It’s a structural problem. The Cell’s accountability system exists because even excellent people perform better with peer pressure. Remove the peers, and you remove the pressure. --- ## Failure mode three: attention fractures Five functions require five different modes of thinking. Strategy requires stepping back, seeing patterns, thinking long-term. Execution requires stepping in, managing details, thinking short-term. Narrative requires empathy for the audience, creative energy, emotional resonance. Access requires social energy, relationship maintenance, outward focus. Integrity requires vigilance, fairness, willingness to confront uncomfortable truths - including truths about yourself. No one can hold all five modes simultaneously. The single player must context-switch between them constantly. And every context switch has a cost. In practice, one or two functions always dominate. The technically skilled single player spends 80% of their time on execution and starves the other four. The naturally social one spends their energy on relationships and lets execution slip. The strategically minded one plans endlessly and ships rarely. The functions that get starved are usually the ones the single player is least comfortable with. And those are precisely the functions that most need attention - because they’re the ones where the single player’s skills are weakest. AI can absorb some of this load. It can handle execution tasks, draft content, do research, prepare for meetings. But it can’t make the judgment calls that each function requires. It can’t decide which strategic direction to pursue, which narrative angle will resonate, which relationship to prioritize, or when standards are slipping. Those decisions require human attention, and the single player has a fixed amount of it. --- ## The single player who works Despite these failure modes, some single players thrive. Not just survive - genuinely thrive, producing exceptional work at extraordinary speed. What separates them from the ones who drift? First, brutal self-awareness. They know their blind spots and build compensating systems. They seek external feedback obsessively - from advisors, peers, customers, mentors. They don’t wait for feedback to find them. They go looking for it. They treat every conversation as a potential challenge to their assumptions. Second, external accountability structures. They join communities, hire coaches, form advisory relationships, make public commitments. They deliberately create the social pressure that the Cell provides structurally. It’s not as robust as having a dedicated peer who knows your work deeply. But it’s vastly better than nothing. Third, ruthless prioritization. They accept that they can’t do all five functions well simultaneously. Instead of spreading thin, they identify which functions matter most at each stage and focus there. Early on, that might be execution and narrative. Later, it might shift to access and strategy. They don’t try to be excellent at everything. They try to be excellent at the right things at the right time. Fourth, they know when to stop being solo. The best single players have a clear mental model of when their configuration becomes a liability. They watch for the signs - repeated mistakes in the same domain, chronic neglect of a critical function, growing gap between their ambition and their capacity. When those signs appear, they start looking for their first Cell member. --- ## When to leave the single player configuration The single player model has a natural ceiling. It’s different for every person and every type of work, but it exists. The signals that you’ve hit it are consistent: the same types of mistakes keep recurring despite your awareness of them. One or two functions are permanently neglected no matter how hard you try. The quality of your judgment in at least one domain is visibly declining because you have no one to stress-test it against. You’re making decisions faster but with less confidence. Growth has stalled and you can’t see why clearly. These are not signs of personal failure. They’re signs that the work has outgrown the configuration. The single player who recognizes this early and evolves into a triad or a larger Cell isn’t admitting weakness. They’re exercising the judgment that defines the roles they’ll eventually share. The most dangerous single players are the ones who never hit this realization. Who attribute every problem to execution rather than structure. Who believe that more hours, more AI tools, or more discipline will solve what is fundamentally a design problem. You cannot be your own accountability check. You can build systems that approximate it. You can seek external input that supplements it. But the structural gap remains. And the longer you operate without acknowledging it, the wider it gets. --- ## The honest assessment The single player configuration is powerful, real, and legitimate. For many types of work, at many stages of development, it is the right choice. The AI revolution has made it more viable than it has ever been. But it is not the endpoint. It is the starting point. The Cell framework doesn’t exist to tell solo operators they’re doing it wrong. It exists to give them a map of what they’re missing and a model for what comes next when they’re ready. Five functions. All non-negotiable. The single player covers them all, imperfectly, with courage and speed. The question is not whether they can. It’s how long they can sustain it before the gaps become the thing that defines their ceiling. --- *This is Part 3 of the Company of the Future series. Next: what happens when you add two more people - and why three is where accountability is born.* --- # Company of the Future, part 4: The Cell of 3 By Filip Blagojević, 2026-03-12. Canonical: https://filipblagojevic.substack.com/p/company-of-the-future-the-cell-of On-site: https://resomnium.com/writing/company-the-cell-of-3 Two is a conversation. Three is a system. This is the most underappreciated structural truth in organizational design. When two people work together, disagreements are negotiations. Someone compromises or someone dominates. There’s no tiebreaker, no outside perspective, no third voice to say “you’re both wrong” or “you’re both right but missing something.” Add a third person and the entire dynamic changes. Suddenly there’s a majority. Suddenly there’s a witness. Suddenly the social physics shift from bilateral negotiation to something far more powerful: genuine peer accountability. Three is where the Cell comes alive. --- ## Why three is the threshold The single player holds all five functions alone. They’re fast, independent, and structurally blind. No one challenges their strategy. No one questions their standards. No one tells them the story they’re telling the market isn’t landing. The moment you add a second person, some of that blindness lifts. But dyads have their own problem: they’re fragile. When two people disagree, there’s nowhere to go. Either one person backs down, or the relationship absorbs the tension. Over time, dyads tend toward either dominance (one person consistently defers) or deadlock (neither will). Neither is healthy. Three breaks this pattern. Three creates the possibility of genuine resolution. When Person A and Person B disagree, Person C isn’t just an audience. They’re a structural tiebreaker who can evaluate both positions and side with the stronger argument. This isn’t voting for the sake of voting. It’s the birth of peer judgment - the moment when decisions get pressure-tested by someone who isn’t emotionally invested in either side. This is what I mean when I say three is where accountability is born. Not accountability to a boss. Not accountability to a metric. Accountability to peers who see your work, understand your domain well enough to challenge it, and have the structural standing to say “I disagree and here’s why.” --- ## How five functions fit three people Three people. Five functions. The math requires doubling up. Each person holds one primary function and shares responsibility for the remaining ones. There’s no universal formula for how to divide them. The right distribution depends on who your three people are - their strengths, their experience, their natural tendencies. But there are principles that matter. **Don’t combine functions that create conflicts of interest.** The person who owns access and relationships should not also own integrity and governance. The temptation to bend rules for a key partner or investor is too structurally predictable. Similarly, whoever owns execution should probably not own the integrity function - the pressure to ship creates natural tension with the patience that governance requires. **Do combine functions that share underlying skills.** Clarity and integrity both require systems thinking and the willingness to have hard conversations. Narrative and access both require understanding people and how to move them. Execution and narrative both require understanding the product deeply enough to represent it authentically. These overlaps aren’t accidental. They reflect real cognitive synergies. **Give each person at least one function that faces inward and one that faces outward.** This prevents the trap where two people are entirely focused on internal operations while the third handles everything external - or vice versa. Everyone should feel the tension between building and communicating, between making and selling. That tension keeps perspectives grounded. To make this concrete, here’s one hypothetical distribution for a tech startup triad. Person A holds Clarity and Integrity - they own the strategic vision and the governance that keeps the Cell honest. Person B holds Execution and Narrative - they build the product and tell its story, which work well together because deep product knowledge makes for authentic storytelling. Person C holds Access - the highest-stakes relationships, partnerships, and fundraising - while contributing to Clarity and Integrity discussions as a peer challenger. This isn’t the only way. Shift the people’s strengths and the configuration shifts with them. The principles matter more than any specific arrangement. The goal isn’t to find the perfect configuration. It’s to find one that works for your specific three people and then adjust as you learn where the stress points are. --- ## The triadic dynamics Three-person groups have structural properties that most people never think about. Understanding them is the difference between a triad that functions and one that tears itself apart. **The tiebreaker effect.** Every disagreement has a potential resolution path. Two people advocate, the third evaluates. This sounds simple but it fundamentally changes how people argue. In a dyad, disagreements feel personal because there’s nowhere else for the energy to go. In a triad, disagreements become presentations to a jury. The quality of arguments improves because there’s someone to persuade. Positions are stated more clearly, evidence is marshaled more carefully, reasoning becomes more rigorous - all because there’s a third mind that needs to be convinced. **The witness effect.** When someone drops their standards, two people notice. When someone avoids a hard conversation, two people see it. When someone makes a commitment and doesn’t follow through, two people know. This is social accountability in its most fundamental form. Humans behave differently when they’re observed. Not because they’re dishonest without witnesses, but because observation activates the part of the brain that cares about consistency and reputation. The witness effect is what keeps the single player’s failure modes in check. **The coalition risk.** This is the dark side of triadic dynamics. Two people can align against the third. This coalition can be conscious - a deliberate power play - or unconscious, where two people simply develop closer rapport and the third gradually becomes an outsider. Coalition dynamics are the primary threat to a Cell of three. They turn a peer accountability system into a two-against-one power structure. The person on the outside loses the ability to challenge effectively because every disagreement feels like - and often is - a losing vote. Healthy triads prevent this through two mechanisms. First, rotating alliances. If Person A sides with Person B on strategy but with Person C on execution, the alliances are fluid and issue-specific rather than fixed and personal. When alignments shift naturally based on the substance of the disagreement, no permanent coalition forms. Second, explicit awareness. The three members must openly acknowledge that coalition dynamics are a structural risk, not a personal failing. When someone feels consistently outvoted, that’s a signal to examine - not dismiss. The Integrity function, whoever holds it, should watch for this pattern actively. --- ## What a Cell of three can do A well-configured triad with AI leverage is a serious force. Three high-quality generalists covering all five functions can operate a meaningful business, launch a product, run a media operation, or manage a complex project - at a level that would have required fifteen or twenty people a decade ago. The AI multiplier works the same way it does for the single player, but now it compounds across three people who can divide and conquer. The triad is particularly strong in situations that require both speed and judgment. The single player has speed but no check on judgment. The full Cell of five has comprehensive judgment but more coordination overhead. The triad is the sweet spot for many projects - fast enough to be responsive, large enough to catch blind spots, small enough that coordination costs are minimal. It’s also the most natural starting configuration for founding teams. Two founders are common but structurally fragile. Four is awkward for decision-making. Three gives you resilience, tiebreaking, and enough coverage of the five functions to operate without critical gaps. --- ## Where the Cell of three breaks The limits are real. **Stretched attention.** Each person carries at least two functions. When workload spikes in both simultaneously - when narrative demands a product launch while execution demands a technical crisis - something gives. The single player faces this across all five functions. The triad faces it across two or three. It’s better, but it’s not solved. **Incomplete challenge.** Three generalists can challenge each other, but they may lack depth in specific domains. If none of the three has genuine experience in, say, legal governance, then the integrity function operates on instinct rather than expertise. The challenges they offer each other in that domain are well-intentioned but shallow. **Coalition decay over time.** Even triads that start healthy can develop fixed coalitions as familiarity breeds predictable alliances. The two people who naturally agree more begin coordinating informally. The third notices and starts disengaging. This is slow and subtle enough that by the time anyone names it, the damage is done. **Succession fragility.** If one person leaves a triad, you don’t lose one-third of your capacity. You lose the entire triadic dynamic. You’re back to a dyad with all its limitations, plus the additional burden of functions that were held by the departing member. Triads have no redundancy. Every member is load-bearing. --- ## The honest assessment The Cell of three is the minimum viable configuration for genuine peer accountability. It’s where you get the structural benefits that the single player lacks - tiebreaking, witnessing, social pressure to maintain standards - without the coordination overhead of a larger group. For many projects, it’s the right configuration permanently. Not every mission needs five dedicated function owners. Not every team can find five people who meet the generalist bar. Three exceptional people who trust each other, challenge each other, and distribute the five functions thoughtfully can build something remarkable. But the triad is also inherently tense. Three is a number that demands active management. Coalition risks are real. Stretched attention is real. The absence of deep specialization in every function is real. The people who make it work are the ones who understand that the tension is the feature, not the bug. The discomfort of being challenged by peers who know your work is exactly what keeps the Cell honest. The moment that discomfort disappears - because coalitions have formed, because standards have aligned downward, because challenge has been replaced by polite agreement - is the moment the triad stops functioning and starts coasting. Three is where accountability is born. Whether it survives depends entirely on whether the three people in the room want it to. --- *This is Part 4 of the Company of the Future series. Next: the full Cell of five - where every function gets a dedicated owner and the accountability system reaches its complete form.* --- # Company of the Future, part 5: The Cell of 5 By Filip Blagojević, 2026-03-17. Canonical: https://filipblagojevic.substack.com/p/company-of-the-future-the-cell-of-759 On-site: https://resomnium.com/writing/company-the-cell-of-5 This is where it all clicks. The single player holds all five functions and has no one to challenge their judgment. The triad covers the functions with stretched attention and coalition risk. Both configurations work. Both have structural limits that no amount of talent or discipline fully overcomes. The Cell of five is different. One person per function. No doubling up. No stretched attention. Every critical domain gets undivided focus from someone whose job is to own it completely. This is the full organism. --- ## What changes at five The shift from three to five isn’t just adding two more people. It’s a qualitative change in how the Cell operates. In the triad, everyone carries multiple functions. This means every person is constantly context-switching between domains that require different modes of thinking. The strategist who also handles integrity must shift between long-term vision and conflict resolution. The executor who also handles narrative must shift between building and storytelling. These are real cognitive costs that compound over time. At five, those costs disappear. Each person wakes up knowing exactly what they own. The Clarity Steward thinks about strategy. The Execution Orchestrator thinks about delivery. The Narrative Steward thinks about the story. The Power & Access Steward thinks about relationships. The Integrity Steward thinks about governance and fairness. This specialization of attention - not specialization of knowledge, every member is still a generalist - is what makes the Cell of five qualitatively stronger than smaller configurations. The strategist has time to think deeply about strategy because they’re not also worrying about contracts. The executor has space to optimize delivery because they’re not also managing investor relationships. And each person can challenge the others from genuine depth of focus. When the Narrative Steward pushes back on positioning, it’s not a passing thought between other responsibilities. It’s a considered opinion from someone whose full attention is on how the world perceives what the Cell builds. --- ## The problems it solves Every structural weakness from the single player and triad articles finds its resolution here. The single player’s blind spots? Five different perspectives, five different cognitive styles, five people whose job includes challenging every major decision. Blind spots don’t disappear, but they get caught. The single player’s slipping standards? Five witnesses. Every function has four peers who see the work, understand it well enough to evaluate it, and have the standing to say “this isn’t good enough.” The accountability that the single player must manufacture artificially is built into the structure. The single player’s fractured attention? Eliminated. One person, one primary domain, undivided focus. The triad’s coalition risk? Harder to sustain in a group of five. Coalitions of two don’t have majority power. Even a coalition of three can be challenged by the remaining two with enough force to trigger structural review. The dynamics are more complex, but also more stable. The triad’s incomplete challenge? Each function has a dedicated owner with the depth to defend their domain and the generalist breadth to challenge others. The challenges are sharper because they come from people with real focus, not stretched attention. --- ## The accountability system This is where the Cell of five earns its design. The accountability system has four layers, each one harder to implement and more important than the last. **Layer one: functional accountability.** Each steward has final authority in their domain and full accountability for outcomes. This is clean and simple. Narrative fails? The Narrative Steward owns that. Execution slips? The Execution Orchestrator owns that. No hiding behind “the team decided.” No diffusing responsibility across committees. This works because authority and accountability are fused. You can’t have one without the other. The person who makes the call is the person who answers for the result. **Layer two: structured dissent.** This is where most organizations fail because they treat disagreement as a problem to manage rather than a capability to institutionalize. In the Cell of five, dissent is formalized. At every major decision, one member is assigned the role of Devil’s Advocate. Their job is to attack assumptions, stress-test incentives, and surface blind spots. This role rotates. No one is permanently the contrarian. Everyone takes turns being the person who has to find the holes. This sounds like theater. It isn’t. When you know your job is to find weaknesses in a plan, you think differently. You stop looking for reasons to agree and start looking for reasons to worry. The quality of decisions improves not because the Devil’s Advocate is always right, but because the process forces the kind of adversarial thinking that groups naturally suppress. And there’s a harder mechanism underneath: the Red Flag Protocol. Any member, at any time, can trigger a full Cell review. This means an immediate pause, a structured examination of whatever triggered the flag - integrity risk, strategy drift, incentive corruption, power misalignment. No hierarchy can block it. No seniority can override it. The most junior member of the Cell can halt a decision they believe is dangerous. This is the structural equivalent of a fire alarm. You hope it never gets pulled. But knowing it exists changes how everyone behaves. **Layer three: integrity enforcement.** The Integrity Steward has powers that no other member holds. They can audit contribution claims and adjust compensation. They can resolve disputes with binding decisions. They can issue formal warnings, suspend voting privileges temporarily, and initiate removal proceedings. Their decisions are binding unless overruled by supermajority - four out of five, or five out of seven in a Cell with flexible seats. This is deliberate. The Integrity Steward needs enough authority to act without consensus, because the situations requiring their intervention are precisely the ones where consensus has broken down. This is the immune system I’ve referenced throughout the series. It doesn’t wait for infection to become sepsis. It acts at the first sign of dysfunction. **Layer four: removal.** This is the layer most organizations refuse to define until it’s too late. The Cell has explicit grounds for removal: integrity breach, persistent underperformance, loss of trust, strategic misalignment. The process is clear. The Integrity Steward initiates a review. Evidence is collected. The member hears the case against them. A supermajority vote decides. Exit terms are predefined before anyone joins. This isn’t punitive. It’s structural honesty. Every member of the Cell knows, from day one, what the boundaries are and what happens when they’re crossed. No ambiguity. No hostage situations. No drawn-out conflicts where someone should leave but nobody has the mechanism to make it happen. Admission works the same way. Unanimous consent is required to add a new member. There’s a trial period of three to six months with full access but limited voting. Then a formal confirmation vote. This keeps the bar high. One person who doesn’t meet the generalist standard, who can’t challenge and be challenged, who doesn’t fit the culture of radical honesty - that person will degrade the entire system. --- ## What the Cell of five demands The Cell of five is the ideal configuration. It is also the hardest to build. Finding five people who each meet the generalist bar is difficult. Finding five who also have the specific depth required for their primary function is harder. Finding five who can challenge each other honestly without damaging relationships is harder still. And finding five who all want to operate this way - with this level of transparency, this level of mutual accountability, this level of structural honesty about power and performance - is the hardest part of all. Most people don’t want this. Most people prefer organizations where accountability is vague, where authority is negotiable, where conflict is avoided rather than structured. The Cell of five is designed for people who find that model suffocating. Who would rather know exactly where they stand, exactly what they own, and exactly what happens if they fail. This is not for everyone. It shouldn’t be. --- ## Where this series lands I started this series with a simple observation: AI changes how we organize to build, not just what we build. The organizational structures we inherited - departments, hierarchies, management layers - were designed to coordinate human labor at scale. When AI handles execution, those structures become overhead. The Cell is what I believe replaces them. A minimum viable human structure for judgment, accountability, and trust, with AI leverage built underneath. We’ve walked through the spectrum. The single player, fast and blind. The triad, accountable but stretched. The full Cell of five, where every function gets dedicated focus and the accountability system reaches its complete form. But frameworks on a page are just frameworks on a page. The harder question is implementation. How does an existing company adopt this? How do you transition from traditional org charts to Cell-based structures? How do you find the right people, set the right incentives, build the right culture? That’s what comes next. --- *This is Part 5 of the Company of the Future series. Next: bringing the Cell into existing organizations - how to start the transition without burning everything down.* --- # Company of the Future, part 6: Updating the Old Company By Filip Blagojević, 2026-03-19. Canonical: https://filipblagojevic.substack.com/p/company-of-the-future-updating-the On-site: https://resomnium.com/writing/company-updating-the-old-company The old company is shrinking whether it wants to or not. In the first three months of 2026, over 55,000 tech workers lost their jobs. Not at struggling companies. At companies posting record revenue. Amazon cut 16,000 jobs while reporting $717 billion in annual sales. Block slashed 40% of its workforce - 4,000 people - and CEO Jack Dorsey called it “a new way of working.” Meta is planning up to 15,000 cuts while doubling its AI spending to $135 billion. Oracle is reportedly eliminating up to 30,000 positions to fund AI data centers. This follows 245,000 tech layoffs in 2025. Microsoft cut 15,000. Intel cut 15,000. Amazon cut 14,000. These are not struggling companies making desperate moves. These are the most profitable organizations on earth deliberately restructuring around smaller teams, flatter hierarchies, and AI-augmented workflows. They are being forced toward the logic of the Cell whether they use that language or not. --- ## The pattern Every one of these restructurings follows the same playbook. Cut management layers. Eliminate coordination roles. Consolidate overlapping functions. Invest in AI tools that let remaining employees do more with less. Shopify CEO Tobi Lutke made it explicit in an internal memo: before you can hire anyone, you must prove that AI can’t do the job. Performance reviews now include how effectively employees use AI tools. Shopify has cut its workforce by 34% since 2022 while growing revenue. Google generates over 25% of its new code through AI. Salesforce froze engineering hiring entirely. Meta’s Zuckerberg has floated the idea that AI could replace most mid-level developers. And Jack Dorsey, announcing Block’s 40% workforce reduction, said it plainly: “The intelligence tools we’re creating and using, paired with smaller and flatter teams, are enabling a new way of working which fundamentally changes what it means to build and run a company.” He’s right. He’s just not saying the uncomfortable part: the “new way of working” has no room for most of the people who built the old one. The pattern is clear. These companies are not just trimming fat. They are fundamentally rethinking how many humans they need and what those humans should be doing. The answer, consistently, is: fewer humans, doing higher-judgment work, augmented by AI. This is the Cell framework in action, even if nobody in these boardrooms is calling it that. Fewer people. Clearer ownership. More leverage per person. The five functions I’ve described throughout this series - strategy, execution, narrative, access, integrity - are exactly the functions these companies are trying to preserve while automating everything underneath. --- ## Why traditional structures resist If the logic is so clear, why is the transition so painful? Because traditional organizational structures are designed to resist exactly this kind of change. The traditional company is built on a simple premise: humans are the execution layer, and management exists to coordinate that execution. You need people to do the work, managers to coordinate the people, directors to coordinate the managers, and executives to coordinate the directors. This made sense for a century. It was an elegant solution to a real problem: how do you align hundreds or thousands of humans toward a common output? Remove the execution bottleneck - which AI is doing - and every coordination layer above it loses its justification. The director who coordinated ten teams doing work that AI now handles has nothing to coordinate. The project manager who tracked dependencies between human workstreams has no workstreams to track. The middle manager who translated executive strategy into team tasks has no team. But these people have titles, salaries, organizational authority, and political power. They will not quietly accept that their function has been automated. They will find new ways to justify their existence - creating meetings, requesting reports, building processes - all of which add coordination overhead that slows the organization down. This is the immune response of the traditional org chart. The structure protects itself even when it’s no longer serving the mission. And the bigger the organization, the stronger the immune response. A 50-person startup can restructure in a quarter. A 50,000-person enterprise can spend years rearranging chairs while the market moves without them. --- ## The mental model for transition Here’s how Cell principles map onto a traditional organization making this shift. **Identify the five functions, not the five people.** Don’t start by reorganizing the org chart. Start by asking: who actually owns strategic clarity right now? Who actually makes execution happen? Who actually shapes the narrative? Who opens the doors? Who maintains integrity and governance? In most organizations, these functions are distributed across dozens of people in fragmented, overlapping ways. Nobody fully owns any of them. Map the reality first. **Find the existing Cells.** Every large organization already has informal Cells operating inside it. Small teams of three to five people who actually get things done, who operate with trust and speed, who have figured out how to work around the bureaucracy. These are your proof of concept. Find them. Study how they work. Protect them from the organizational immune response that will try to absorb them back into the hierarchy. **Collapse coordination layers, not capability.** The goal is not to fire everyone and hope AI fills the gap. The goal is to remove the layers that exist purely to coordinate human labor - the management overhead that was necessary when execution required large teams. Push authority down to the people closest to the work. Give them AI tools and the autonomy to use them. Eliminate the reporting structures that existed to keep humans aligned when AI can do that faster. **Make ownership explicit.** The biggest disease of traditional organizations is diffused accountability. Twenty people touch a decision, nobody owns the outcome. The Cell principle that matters most for legacy companies is this: every function must have one name attached to it. Not a committee. Not a “cross-functional team.” One person who has the authority and the accountability. **Accept that the transition creates losers.** This is the part nobody wants to say out loud. When you collapse coordination layers, the people in those layers lose their jobs. When you consolidate functions, the people who held fragments of those functions become redundant. When you push authority down, the people who derived power from being in the chain of command lose that power. No framework can make this painless. The Cell doesn’t pretend otherwise. What it offers instead is a clear destination - a structure designed for the world we’re moving into, not the one we’re leaving behind. --- ## The three traps Organizations attempting this transition consistently fall into three traps. **The tool trap.** They buy AI tools and plug them into the existing structure, expecting transformation. It doesn’t work. Giving a seven-layer management hierarchy access to ChatGPT doesn’t make it a Cell. It makes it a seven-layer management hierarchy that uses ChatGPT. The structure must change, not just the tooling. **The halfway trap.** They cut headcount but preserve the organizational structure. Fewer people doing the same jobs in the same hierarchy with the same coordination overhead. This is the worst of both worlds - less capacity to execute but the same structural friction. The people who remain are overloaded and burned out, but the fundamental design hasn’t changed. **The avoidance trap.** They restructure everything except accountability. They flatten the hierarchy, they adopt AI tools, they reduce headcount - but they never answer the hard question: who owns what? Who has final authority? Who is accountable when things go wrong? Without this, you don’t have a Cell. You have a flat organization with no immune system. And flat organizations without clear authority structures are just chaos with good intentions. --- ## The honest assessment The companies making headlines with massive layoffs are not implementing the Cell framework. Most of them are doing something much cruder - cutting costs and hoping AI compensates for the lost capacity. Some will succeed despite the crudeness because they have enough momentum, enough capital, enough market power to absorb the turbulence. But crude restructuring without clear design creates its own problems. The remaining employees are overloaded, confused about ownership, and operating in structures that were designed for a workforce twice the current size. They have AI tools but no clarity about who owns what. They have fewer layers but the same diffused accountability. They have smaller teams but the same committee-driven decision-making. The organizations that will truly thrive are the ones that treat this moment not as a cost-cutting exercise but as a structural redesign. That ask the hard questions about who owns what. That build real accountability into their new, smaller teams. That recognize the five functions are non-negotiable even if the number of people holding them drops dramatically. The old company isn’t dying. It’s being compressed. The question is whether what emerges from that compression is designed or accidental. In the next articles, I’ll get specific. How a startup should build its first Cell. How a mid-size company restructures around Cell principles. How an enterprise creates Cells within existing structures without tearing everything down. The pressure is here. The direction is clear. What remains is the how. --- *This is Part 6 of the Company of the Future series. Next: how different company sizes adopt the Cell - starting with the organization that has no legacy to overcome.* --- # Company of the Future, part 7: The Cell in a Startup By Filip Blagojević, 2026-03-24. Canonical: https://filipblagojevic.substack.com/p/company-of-the-future-the-cell-in On-site: https://resomnium.com/writing/company-the-cell-in-a-startup Startups have one advantage that no legacy company can buy: they get to start clean. No management layers to collapse. No coordination roles to justify. No political power structures defending their existence. A startup building from zero can design around the Cell framework from day one, making the five functions the foundation of the company rather than a retrofit. But startups also have a constraint that legacy companies don’t: they can’t afford five exceptional generalists on day one. They start as a single player, or maybe a pair, holding all five functions with duct tape and determination. The full Cell is the destination. The question is how to build toward it without breaking along the way. Both of these realities matter equally. The advantage is real. The constraint is real. The startup that understands both will build something structurally sound. The one that ignores either will build something that looks like a company but operates like a mess. --- ## Design the Cell before you fill it Most startups don’t think about organizational structure until they have a structural problem. They hire based on what feels urgent, not based on what functions need to be covered. They end up with three engineers and no one who owns the narrative. Or two business development people and no one who owns execution quality. Or five people who all think they own strategy. The Cell framework gives you a different starting point. Before you hire anyone, map the five functions: clarity, execution, narrative, access, integrity. Ask yourself honestly: which ones do I cover well? Which ones am I weak at? Which ones am I neglecting entirely? This map becomes your hiring compass. Not “we need a developer” or “we need a marketer.” Instead: “I’m strong on execution and narrative but weak on integrity and access. My first partner needs to cover the functions I can’t.” This reframing matters because it prevents the most common startup hiring mistake: hiring people who are good at what you’re already good at. Technical founders hire more technical people. Sales-oriented founders hire more salespeople. The result is a team that’s deep in one or two functions and completely absent in the others. The Cell framework forces you to think in terms of coverage, not comfort. --- ## The founder is the first Cell Every founder starts as a single player. We covered the strengths and failure modes of that configuration earlier in this series. But in a startup context, there’s a specific dynamic worth highlighting: the founder’s relationship with the five functions sets the DNA of the company. Whatever the founder is good at becomes the company’s strength. Whatever they neglect becomes the company’s blind spot. And these patterns are remarkably persistent. A founder who neglects integrity from the start will build a company where governance is always an afterthought, where incentives are never quite right, where conflicts simmer because nobody owns their resolution. A founder who neglects narrative will build a product that nobody understands, no matter how good it is. This is why self-awareness matters so much at the founding stage. Not the vague kind where you say “I know I have weaknesses.” The specific kind where you can name exactly which of the five functions you’re worst at and commit to covering that gap as your first structural priority. The honest founders admit their gaps early and build around them. The delusional ones convince themselves they can do everything and wonder why the company keeps hitting the same walls. --- ## Principles for building toward the full Cell I’m not going to prescribe a sequence. Every startup’s path is different depending on the founder’s strengths, the nature of the product, the market, and the stage. But there are principles that hold regardless. **Cover the weakest function first.** Your first partner or hire should address your biggest gap, not reinforce your existing strength. If you’re a builder who can’t sell, your first addition covers access and narrative. If you’re a networker who can’t ship, your first addition covers execution. This feels counterintuitive because we naturally want to work with people who think like us. Resist that impulse. The Cell needs coverage, not redundancy. **Every person must cover at least one complete function.** Partial ownership is worse than no ownership. If you bring someone in to “help with marketing” but they don’t actually own the narrative function end-to-end, you haven’t solved anything. You’ve just added a person. Make it clear from day one: you own this function. You have authority over it. You are accountable for it. **Don’t add people to add capacity. Add people to add capability.** A startup that hires a second engineer because the first one is overwhelmed has added capacity. The work gets done faster but the structural gaps remain. A startup that hires someone who covers a missing function has added capability. The Cell gets closer to completeness. In the AI age, capacity problems are increasingly solvable with tools. Capability gaps are only solvable with the right humans. **Build the accountability habits before you need the formal system.** A two-person startup doesn’t need a Devil’s Advocate rotation or a Red Flag Protocol. But it does need the habit of genuine challenge. From the earliest stage, establish the norm that disagreement is expected, that pushback is welcomed, that “I think you’re wrong and here’s why” is the highest form of respect in your company. These habits are much harder to install later than they are to build from the start. **Define the integrity function early, even if no one owns it full-time.** Most startups delay governance, legal structure, and incentive design until they’re forced into it by a crisis. By then, the patterns are set and the damage is done. Even if you can’t dedicate a person to the integrity function yet, acknowledge it explicitly. Document your agreements. Be clear about equity and compensation logic. Set expectations for how conflicts get resolved. The Integrity Steward may not arrive until you’re a Cell of three or five. But the integrity function should be active from day one. --- ## What startups get wrong The most common structural mistake in startups is conflating roles with titles. Founders hand out titles - CTO, CMO, Head of Growth - without mapping those titles to the five functions. The result is titles that feel good but don’t create clarity. Three people all think they influence strategy. Nobody clearly owns integrity. The narrative function is split between marketing and the CEO with no single point of accountability. The Cell framework doesn’t care about titles. It cares about functions. You can call the Execution Orchestrator whatever you want. What matters is that one person owns execution, has authority over it, and is accountable for it. The title is cosmetic. The ownership is structural. The second most common mistake is hiring for the role you understand rather than the function you need. The technical founder who hires a “Head of Sales” because they’ve heard startups need sales is making a category error. They might need access and relationships. They might need narrative and distribution. They might need both. “Sales” is a traditional job category, not a Cell function. The founder who hires based on functions will build a better team than the one who hires based on convention. The third mistake is avoiding the integrity function because it feels premature. “We’re only three people, we don’t need governance.” Until equity disagreements surface. Until someone underperforms and nobody knows how to address it. Until a key hire leaves because incentives were never properly aligned. The integrity function is the one startups most often neglect and most painfully regret neglecting. --- ## The honest assessment Building a startup as a Cell from day one is harder than building one the traditional way. It requires more self-awareness from the founder. It demands uncomfortable conversations about who owns what before the company has even shipped anything. It means hiring for coverage rather than comfort, which is lonelier and slower. But it also means never having to do the painful restructuring that the previous article described. The legacy companies cutting thousands of jobs are trying to retrofit what you can build natively. Their immune response fights the change. You have no immune response to fight because you’re building the structure from scratch. This is the startup’s real advantage. Not speed, not scrappiness, not disruption. The advantage is design. You get to build the organism the way it should work, rather than trying to force an existing organism to evolve. Five functions. Clear ownership. Generalist standards. Accountability from day one. AI leverage underneath every seat. That’s the Cell-native startup. And it starts the moment the founder honestly maps their own strengths and gaps. --- *This is Part 7 of the Company of the Future series. Next: how mid-size companies restructure around the Cell - the most complex transition of all.*