Why brilliant teams sometimes make stupid decisions
The intelligence is already inside your company. Here is why it’s stuck.
Every CEO has had this moment: you’re in a room with people you hired because you thought they could do the job. You have good data. Adequate resources. But the team just made the same mistake it made six months ago … Or failed to act on something everybody knew … Or the decision coming out of the meeting was dumber than what any individual in the room would have decided alone.
So you’re forced to do heroics to salvage the situation.
You have probably blamed the usual suspects: bad communication, wrong incentives, the wrong person in the wrong seat. Maybe you restructured, maybe you brought in a consultant. The problem came back wearing different clothes.
It came back because the problem is not your people, your data, or your incentives. The problem is structural. Your organization has the intelligence it needs — the seeing, the understanding, the ability to act — and something in its architecture is preventing that intelligence from being used.
I call this Trapped Intelligence, and after over twenty years of running companies, I believe it’s the single most common reason good organizations underperform.
Where intelligence gets trapped
There are three places intelligence gets stuck in any organization, and they’re different problems with different fixes.
| What your org can SEE | What your org can UNDERSTAND | What your org can REACH |
|---|---|---|
| The information reaching the people who need it. Sales reps hear what customers want. Does that signal make it to the product team with enough fidelity to act on? | The work your people do with what they receive. Your analyst reads the competitive report. Can she interpret what the pricing data means, or does it sit unprocessed? | The actions your people can actually take. Your middle manager sees the problem and understands it. Does he have the authority, the budget, or the mechanism to do anything about it? |
Most organizational failures are not failures of effort or intent. They are failures at one of these three points.
- The signal did not arrive.
- The signal arrived but nobody could make sense of it.
- The signal was understood but nobody had the lever to act.
Same symptom — the company is not doing what it knows — three completely different causes, three completely different interventions.
A restructuring that fixes a “reach” problem when the real problem is “see” makes things worse: you rearranged authority when the issue was information flow. A training program that fixes an “understand” problem when the real problem is “reach” wastes everyone’s time: your people already get it, they just can’t act.
The hidden fourth: what you think you have versus what you actually have
There is a fourth trapping mechanism that is often the most consequential and the easiest to fix. It is the gap between what your organization thinks its capabilities are and what they actually are.
A senior engineer on your team has deep experience in deployment automation from his previous company. He could cut your pipeline time in half. His manager doesn’t know this — he was hired for API design. The engineer doesn’t think it’s relevant to his current role. The capability is real, but nobody’s self-model includes it. The intelligence isn’t trapped because it’s missing, but because it’s invisible.
This works in the other direction too. An executive who believes he understands a domain he doesn’t — regulatory risk, technical architecture, a market he has never operated in — makes decisions whose consequences he can’t model. His self-model exceeds his actual capability. The intelligence he does have gets misdirected because the map is wrong.
The gap between self-model and reality is the single highest-leverage intervention point, because it requires changing the least. You don’t need to hire someone new, build something, or need to restructure. You need to make what is already there visible — or make what isn’t there visible too, before it causes damage.
This is not a theory. This is an operating model.
I ran six companies over twenty years. I wrote a book about how organizations process information and make decisions. I arrived at this way of seeing organizational failure through building and running companies, not through reading academic papers.
Years later, I discovered that researchers in biology, cognitive science, and AI were converging on the same structural patterns from completely independent directions. That convergence is what convinced me the pattern is real and worth sharing.
What you have just read is the surface of a framework I call the Cognitive Boundaries Framework. It gives you a methodology and a small vocabulary for recognizing the same structural problem in places you would never have thought to look.
What to do next
If you want to diagnose your own organization, start here:
- The Trapped Intelligence Diagnostic — The top-level tool. A structured walkthrough to find exactly where intelligence is stuck in your organization and what kind of intervention would release it.
Then go deeper with these:
- The Team Trade-Off Matrix — See where combining people made your organization smarter and where it made it dumber.
- The Franz Ferdinand Check — Find the positions in your org where someone’s authority exceeds their understanding. The blast-radius diagnostic.
- The Capability Reality Check — Discover what your team thinks its capabilities are versus what they actually are. The invisible-expertise and overconfidence diagnostic.
If you want to understand the vocabulary, the Framework Vocabulary defines every term with precision and gives examples for both business and research contexts.
If you are a researcher looking at this from an academic perspective, the framework connects to work by Michael Levin, Karl Friston, and others. That is a different entry point — reach out and I will point you in the right direction.