- The company brain isn’t on or off: it has five levels, from “lives in heads” (L0) to “agent-operated” (L4).
- Almost every company sits at L0 or L1, and often doesn’t know it: that’s where AI stays a toy.
- The jump that matters is from L1 (written but dead) to L3 (self-updating and queryable): where knowledge stops rotting.
- In two minutes the self-check tells you your level and your next step, without switching tools.
The five levels of the company brain
It’s not a yes-or-no question. A company’s brain lives at a precise level, and the level tells you how much its knowledge is yours, alive, and useful, for people and for AI. Here are the five.
- L0In heads and chats
Knowledge lives in people and disconnected chats. Every AI chat starts from zero.
- L1Written but dead
A wiki exists, but it’s stale and nobody updates it. The AI ignores it: a false sense of safety.
- L2Owned and structured
Decisions, state and rules in files you own and govern, updated by habit. People and agents read them.
- L3Self-updating and queryable
It absorbs from email, chat and calls, answers with the source, and captures tacit knowledge through interviews.
- L4Agent-operated
Agents run the work on the brain and keep it fresh. The company operates from it.
What level are you at? Find out in 2 minutes
Seven questions about your company, not your software. At the end you get your level and a single next step, the right one for you.
What level is your company brain at?
0 of 7Answer 7 questions about your company. At the end you’ll know your level, from L0 to L4, and what the next step is.
01Where does your project state live today?
02If a key person leaves tomorrow, their know-how…
03When you open a new AI chat to work, how much context does it have?
04Your knowledge base or wiki…
05Important decisions made in calls or chats…
06Can you “ask your company” and get an answer with its source?
07Recurring routines (reports, reviews, follow-ups)…
The problem isn’t a lack of AI. It’s the level underneath.
95% of enterprise AI projects leave no measurable mark[1]. Almost never the model’s fault: what’s missing is the level underneath, the company brain that provides the context. An advanced agent on an L0 company doesn’t work, because you’re asking it to reason over knowledge that lives nowhere it can read.
There’s a reason this need stays invisible: almost nobody knows they’re at L0 or L1. You only find out when you measure it. That’s exactly the point of the ladder: to give a name to a problem you already have.
Companies don’t have a memory problem. They have a level problem: the knowledge is there, but it lives where nobody, and no AI, can use it.
From “written but dead” to “self-updating and queryable”
Most companies that “gave it a try” are at L1: a wiki or a folder of documents nobody updates and the AI never reads. Moving to L2 means owning and structuring that context in files you govern, which is exactly what our open-source kit cowork-os does[2]. Moving to L3 means you stop updating it by hand: the brain absorbs from the sources, answers with the source, and the know-how of whoever leaves comes in through an interview instead of disappearing.
You don’t need to switch tools to start, you need to know where you stand. If you want the practical method, we wrote it: how to build a company brain on files.
The maturity model, in practice
What is the company brain maturity model?
It’s a five-level scale (L0-L4) that measures how a company’s operating knowledge lives: from “in people’s heads” to “agent-operated”. It helps you see where you are and what your next step is, without slogans.
What level are most companies at?
L0 or L1: knowledge lives in heads and chats, or in a wiki nobody updates. It’s not a judgment, it’s the normal starting point. The value is knowing your next step.
Do I need to buy something to level up?
No. The first jump, to L2, is a method: put decisions, state, and rules in files you own. The open-source cowork-os kit gives you the structure for free. Higher levels add automations and agents, but the starting point is free and yours.
Is the self-check reliable?
It’s an indication, not an audit. Seven questions give you your level honestly; for a serious assessment of your case, we start from a call and your real process.
Sources
- [1]Fortune, “MIT report: 95% of generative AI pilots at companies are failing”, August 18, 2025. fortune.com
- [2]cowork-os, open-source repository. github.com
The company brain maturity model is a Yempik editorial model. The self-check is an indicative tool, not an audit. The kit cited is our open-source cowork-os (MIT license). Where we cite a number, you’ll find the source below.
Want to know your real level, on your actual process?
We start from your process, not the tool. We help you level up, from putting context on files you own to agents that work on it. Fixed price and timeline, the code is yours.