--- title: "Conway's Law Has a Corollary - In Parallel" description: "Your operating model shapes your AI. But a context layer that only records will mirror your dysfunction faithfully — unless it also coordinates. We measured the cost: 16.5 hours a week, per manager." url: https://www.in-parallel.com/insight/conways-law-has-a-corollary --- AI in Business # Conway's Law Has a Corollary Pen-and-watercolor illustration of interconnected organizational nodes and communication flows, representing how systems mirror the organizations that build them. Published June 13, 2026 Markku Mäkeläinen Co-founder, CEO [LinkedIn](https://www.linkedin.com/in/markkumakelainen/) Contents In 1968, Melvin Conway observed that organizations design systems that copy their own communication structures. Build software with four teams who don’t talk, and you ship four modules that don’t talk either. The system mirrors the org. It’s having a moment again, and for good reason. A recent Forrester piece argued that in the age of agentic AI, your operating model matters more than your AI model. That’s correct. If the operating model is fragmented, siloed, and built for an earlier era of work, AI doesn’t transcend those flaws — it inherits them and reproduces them at machine speed. Agentic systems don’t escape your org chart. They amplify it. But there’s a corollary that doesn’t get said out loud, and it’s the one that actually tells you what to fix first. ## The wound, measured — not asserted Most of the conversation about the AI productivity paradox is assertion: _gains dissipate inside operating models built for human-only work._ True. We went and measured it. In June 2026 we asked 247 managers across five countries to itemize where their week actually goes — not how they feel about meetings, but the hours. The average manager loses **16.5 hours a week** to coordination overhead: status meetings that exist to transfer state, re-explaining context to people and tools that should already have it, and hunting for decisions that were definitely made somewhere. That’s **2.1 working days, every week.** At a loaded cost, it’s roughly **$64,600 per manager, per year.** We call it the coordination tax. Most teams can’t name it, because it arrives in fifteen-minute slices that never hit a budget line. Ask managers directly and they’ll say coordination eats 21% of their week. Ask them to count the hours and it’s 41%. The largest cost in knowledge work is invisible to the people paying it. ## The AI paradox, proven Here is the part that should stop every leader betting on a model upgrade. **86%** of the managers we surveyed already use AI at work. The daily users carry **more than twice the coordination load** of non-users — 20.3 hours a week versus 9.1. Not because AI made them slower, but because a stateless assistant is one more thing to brief. **91%** load context manually before the tool is useful. Only 9% say their assistant usually knows enough to help. The bottleneck has moved. It is no longer model capability. It is context access. Which is exactly why your operating model beats your model: a smarter model pointed at a fragmented org just produces fragmentation, faster. ## The corollary Here it is. **If systems mirror organizations, then a context layer that only records will mirror a broken organization faithfully.** It will capture the silos, the orphaned decisions, the dependencies nobody flagged — and serve them back at machine speed. Recording is not intelligence. A perfect transcript of a dysfunctional meeting is still dysfunction, now searchable. The difference between a recording layer and organizational intelligence is **coordination**: context tied to goals, decisions, owners, and dependencies, so it composes into something usable instead of just accumulating. Capture without coordination digitizes the mess. That’s the trap most “AI memory” products walk straight into. ## How you escape it Escaping the corollary takes four things, and they aren’t four initiatives. They’re one substrate — what we call an **Shared Context** **Aligned on outcomes, not use cases.** A use case describes a problem to automate. An outcome describes what the organization is trying to become. Design around isolated use cases and you get brittle point solutions; design around outcomes and the work composes. In Parallel pushes thinking up the ladder — goals first, then the means to reach them. **Automated at the skill level.** This is where capability becomes reusable. In Parallel’s own execution model is composed of more than 30 discrete skills that compose against a shared context — not agents bolted onto yesterday’s workflows, but a system built skill-first from the ground up. It’s the architecture the analysts are now telling enterprises to adopt; it’s how we already work. **A shared context layer that is organizational intelligence.** Every decision, every meeting outcome, every action item — captured, structured, and made machine-readable for your people and, through MCP, for any AI tool they use. Claude, GPT, Gemini: all drawing on the same always-on org memory instead of starting from zero every session. **Accountable by design.** Every commitment captured, drift surfaced as it happens rather than at the next steering committee, every decision owned. Coordination without accountability is just a nicer-looking silo. ## From the tax to the dividend The Coordination Tax Index measured the downside. The more interesting number is the upside, and managers have already priced it: an AI that could reach the team’s shared context would return **4.4 hours a week — about $17,000 per manager per year, before any process change.** That’s 27% of the tax, recovered by fixing context access alone. That recovered capacity is the next thing worth measuring, and we’re building the instrument to do it. Call it the coordination dividend: what an organization gains when the coordination infrastructure is actually in place. ## The point “Operating models deliver outcomes” is the right axiom. But an operating model only delivers outcomes if it can remember what it decided, coordinate across the teams that have to deliver it, and account for who owns what. Conway’s Law tells you your AI will end up looking like your organization. The corollary tells you what to fix before you scale it: not the model — the layer underneath it. **See the full picture.** The Coordination Tax Index 2026 — 247 managers, five countries, nine industries — breaks the tax down by role, country, tool count, and AI usage. Download the free report at [in-parallel.com/coordination-tax](https://www.in-parallel.com/in-parallel.com/coordination-tax) Already running Copilot, Gemini, or Claude and want them drawing on your org’s shared context? Book a demo at [in-parallel.com/demo](https://www.in-parallel.com/in-parallel.com/demo) ### Get insights like this in your inbox One email per week on execution intelligence, team coordination, and enterprise AI. No fluff. By subscribing you agree to our [Privacy Policy](https://www.in-parallel.com/privacy-policy). Unsubscribe anytime. [Back to all insights](https://www.in-parallel.com/insight) [Book a demo](https://www.in-parallel.com/demo) ## Related articles The Coordination Tax — Richard H. Carson on how change actually works Leadership Insights · Oct 20, 2026 ### You Either Manage Change or It Manages You: Richard H. Carson on how change actually works Richard H. Carson, PhD — 40 years inside county governments, regional planning agencies, nonprofits, and the state of Oregon — on why every organization has a risk manager but almost none has a change manager, and what that asymmetry costs. A rope suspension bridge over a dark river towards the fells, drawn in pen and watercolour AI in Business · Oct 9, 2026 ### AI-Assisted or AI-Native: The Bridge Is Shared Context At a CIO data event in Espoo, HumanX in Amsterdam and a post-merger integration forum, large companies explained how they brought AI in by mapping processes and fitting AI into the gaps. That keeps a structure built around human limits. Getting to AI-native means rebuilding information flows into shared context first. The Coordination Tax — Adam Christing on humor as leadership Leadership Insights · Oct 7, 2026 ### You Can't Fix Culture With a Mission Statement: Adam Christing on humor as leadership Adam Christing — professional master of ceremonies, founder of Clean Comedians, and author of The Laughter Factor — on why culture is a feeling rather than a statement, and why humor is the fastest way for a leader to shift it. The Coordination Tax — Jennifer Recla on the urgency mindset breaking your team Leadership Insights · Sep 22, 2026 ### Urgency Isn't Leadership. It's Addiction: Jennifer Recla on the mindset breaking your team Jennifer Recla — leadership coach who has worked with hundreds of leaders across nonprofits, mission-driven organizations, and large financial institutions — on the urgency mindset behind most organizational chaos, and why most leaders have no idea how much of it they're creating themselves. The Coordination Tax — Barry Wolfe on value-centric leadership Leadership Insights · Sep 15, 2026 ### HR Is the Last Line of Defense Against Chaos: Barry Wolfe on value-centric leadership Barry Wolfe — HR consultant, founder of Argos HR, and author of It's All in Your Head — on why HR is the last line of defense against organizational chaos, why every role needs 3–5 clear value outcomes, and why the foundation of a well-run company is one document under 300 words. Two colleagues sharing one open binder — context handed over at the resolution the other person needs Leadership Insights · Sep 9, 2026 ### Your Job Is Already Context Management On The Coordination Tax, Stanford's Robert E. Siegel put it as an equation: truth equals facts plus context. Facts are cheap now. Context is not. Read the manager's week against that equation and the 16.5 hours a week lost to coordination stop looking like overhead — they are context management, performed manually, by the most expensive person in the room. The Coordination Tax — Alberto Torres on surviving disruption Leadership Insights · Sep 8, 2026 ### You Can't Sword-Fight a Bullet: Alberto Torres on surviving disruption Alberto Torres — former Nokia executive, ex-McKinsey partner, and author of The Z Curve — on what actually happened at Nokia, why great peace-time leaders quietly fail in war-time, and what every company facing AI right now should learn from the companies that got disruption right. The Coordination Tax — Olivia Montgomery on why decisions belong lower in the org Leadership Insights · Sep 1, 2026 ### Your Leadership Doesn't Know What's Going On: Olivia Montgomery on why decisions belong lower in the org Olivia Montgomery — principal analyst at Data Frame and former Gartner researcher — on why every layer up the org chart knows less about what's really going on, and why the fix isn't more visibility for leadership but pushing decisions down two or three levels. The Coordination Tax — Craig Denison on the founder's hardest transition Leadership Insights · Aug 25, 2026 ### You Don't Get to Be the Rock Star Anymore: Craig Denison on the founder's hardest transition Craig Denison — head of growth at The Table Group (Patrick Lencioni's firm) and founder of Green Thumb Consulting — on the transition every founder hits when execution stops running through their brain, and why 99% of the time the founder is creating the dysfunction they are trying to fix. A manager mid-explanation, watch visible — the hour spent re-explaining context AI in Business · Aug 24, 2026 ### AI Saves a Day a Week. The Coordination Tax Collects It. BCG's fourth AI at Work survey found 42% of regular frontline AI users save at least a full workday every week, and that two-thirds get little or no guidance on what to do with it. The usual reading is that leaders forgot to give direction. The better reading is that the hours were already spoken for. AI compresses the work; it doesn't touch the work about the work.