What team memory is (and why your wiki isn’t it)

Every team writes things down. Meeting notes, wikis, specs, Slack threads, comments on tickets. And somehow, when you actually need to know something, you can’t find it. You ask in a channel. Someone who has been around longer answers from their head.

This isn’t a discipline problem. Teams that document everything often have it worse, not better. The more you write, the more places an answer can hide.

The real problem is that a document is frozen. It captures what was true on the day someone wrote it. Then the plan changes, the customer upgrades, the deadline moves — and the document stays exactly as it was. Nobody goes back to fix old prose. There is too much of it.

Documents store sentences. Memory stores facts.

A memory is a different kind of thing. Not pages of prose, but small facts that connect to each other. “Acme is on the Team plan.” “Sarah owns the billing service.” “We dropped the Redis idea in March.” Each fact lives in exactly one place. When it changes, you change it once, and everyone sees the new version.

In Ontonym, those facts are objects with properties and relations between them. A customer links to its contract. The contract links to the decision that shaped it. To answer a question, you follow the connections. You don’t skim twelve documents hoping the sentence you need is in this one.

The part that changes everything: your AI reads it too

Here is where this stops being a nicer wiki. Your AI tools can read the same memory you do. Ask Claude “which customers are still on the old plan” and it answers from current facts — not from a doc that was right in April. It sees exactly what you have permission to see, and nothing more.

Wikis still have a place. Long-form explanations, guides, essays — prose is good at those. But the stuff your team asks each other all day is facts. Facts belong in memory.

If you want to try it, personal memories are free and unlimited. Connect your AI in a few minutes with the setup guide.