- A knowledge base, a wiki, or an intranet is a passive archive: someone writes it, the rest of the team looks things up, and from there it ages.
- A company brain is alive: it updates by habit, answers with the source, and is read by both the people and the agents who do the work.
- A wiki is enough while the context changes little, the team is small, and no one asks the AI to work on it. Above that threshold it becomes a false sense of safety.
- You don’t throw the wiki away: you own it, structure it, and keep it alive. It’s the jump from "written" to "operating".
You consult an archive, you work with a brain
A knowledge base, a wiki, a Notion, or an intranet all do the same thing: they’re a place where someone writes down what they know so others can find it when they look. The value is real, but the model is passive. A person produces the content at a specific moment, then it sits still while the work moves on. No one updates it until it’s needed, and by the time it’s needed it’s already stale. The AI, usually, doesn’t even look at it.
A company brain starts from the same raw material, decisions, rules, project state, but the verb changes. You don’t consult it: you work with it. It updates by habit, not in fits and starts; it answers a question with the source, not with a link to a page you have to read; and it’s read by both the people and the agents that carry out the work. It’s the difference between an archive that tells you how it was and context that knows how it is, right now.
A wiki is a place where you put what you know and hope someone finds it. A company brain is the context the team, and the AI, work with every day.
The same information, a different behavior
The difference isn’t in the technology or how nice the interface is: a wiki can be perfectly tidy and still be a dead archive. The difference is in behavior, and it’s clearest row by row. Where the context lives, who owns it, whether it’s readable and auditable, what happens when you switch tools, and whether the team, and the AI, really share the same state.
This comparison puts context on files you own on the right, the idea behind our open-source kits cowork-os and code-os, and a tool’s automatic memory on the left. It applies to a wiki the same way: one row leans toward the archive on purpose, because starting with a shared page costs nothing. It’s not advertising, it’s a criterion for choosing.
You don’t always need a brain: when a wiki is the right answer
A wiki is enough, and perfectly fine, when the context changes little, the team is small, and the knowledge sits in a few heads that talk every day. If you need to document stable policies, onboarding, credentials, or procedures that change once a year, a tidy archive is the correct tool: you don’t need a brain to remember things that don’t move.
It becomes a problem when three things happen at once: the context changes fast, the people who know it leave or multiply, and you start asking the AI to work on it. That’s when the wiki stops being useful and turns into a false sense of safety: something is written down, but it’s stale, no one updates it, and the agent reasons over it as if it were true. That’s exactly the "written but dead" level on the maturity ladder.
From archive to brain: you don’t throw it away, you make it live
Moving from a wiki to a company brain doesn’t mean throwing away what you’ve written and starting over. It means three things: owning it (moving the context onto files you govern, not inside a tool someone else controls), structuring it (decisions, rules, state, and open questions, not free-form pages), and keeping it alive (updating it by habit, so the AI and the team always read the latest version). It’s the jump from "written" to "operating".
Before you move, it helps to know where you start. The company brain maturity ladder gives a name to the exact point you’re at, and a wiki nobody updates is the level we call "written but dead". Find out your level, then decide the next step.
The hard part isn’t moving the files, it’s transferring the knowledge that lives in heads into readable context. The practical method, with interviews, transcription, and validation, we wrote here: how to build a company brain on files.
Company brain and wiki, in practice
What’s the difference between a company brain and a knowledge base?
A knowledge base or a wiki is an archive to consult: someone writes it, the rest of the team looks it up when needed, and from there it ages. The company brain is the live operating context you work with: it updates by habit, answers with the source, and is read by both people and agents. The knowledge base tells you "how it’s built", the company brain tells you "how we work, right now".
Can a wiki or a Notion be enough?
Yes, as long as the context changes little, the team is small, and no one asks the AI to work on it. For stable policies, onboarding, or procedures that change once a year, a tidy archive is the right tool. It becomes a problem when the context moves fast and you start using it as the basis for an agent: that’s when a stale page turns into a false sense of safety.
Do I have to throw away my wiki to have a company brain?
No. The content you’ve written is raw material. The jump is owning it (moving it onto files you govern), structuring it (decisions, rules, state, open questions), and keeping it alive with a routine. You don’t start from scratch: you move from "written" to "operating".
Why doesn’t the AI use my wiki?
Because a wiki is built to be read by a person who searches it, not to give context to an agent. Often the content is scattered, unstructured, and stale, and the AI has no way to know what’s still valid. A company brain is written to be readable by people and agents together, with the source next to every answer.
Sources
- [1]cowork-os, open-source repository. github.com
This page is written by Raffaele Zarrelli, founder of Yempik, with editing done with Claude. The company brain and its maturity model are Yempik editorial models. The kits cited are our open-source cowork-os and code-os (MIT license).
Has your wiki been frozen for months? Turn it into live context.
We start from your process, not the tool. We help you move decisions, rules, and state onto files you own, governed, that the AI and the team can re-read and update. Fixed price and timeline, the code is yours.