Skip to content
Inside the AI Orchestrator OS

AI meeting notes that file themselves into a wiki.

Most AI meeting notes end up in a folder nobody opens. In the AI Orchestrator OS, every recorded call is written up and filed into a wiki, usually within minutes of the transcript landing. Each one is linked to the client, the people and the decisions. Your agents read it before they do the work.

Included in the AI Orchestrator OS.

  • Every recorded call filed within minutes of its transcript, with no prompt from you.
  • Plain text on your own Mac. Any model can read it.
  • Your agents read the wiki before they write a word.
The problem

Meeting notes AI apps give you storage, not memory.

Everyone has an AI note-taker now. It records the call and writes a summary. Then the summary sits there.

  • A summary sees one call. A note-taker is one guard watching one monitor. He writes up that call and never sees the other twenty. So he can't tell you the client who went quiet in March is the one at risk in June.
  • Finding anything is a hunt. You want what a client said three weeks ago. You're back in the tool, scrolling.
  • Your teaching is gone when the call ends. The stories you use to explain the hard stuff land once, then you pay to explain them again.
  • Your AI starts from zero. The model does not get smarter the more you use it. What grows is the context you feed it. A folder of summaries feeds it nothing.
What you get

An AI meeting summary covers one call. The wiki joins them all.

A call is not dumped in raw. Claude reads the transcript and writes it up into linked pages. The thinking is done when the call goes in, so every later read is fast.

People and clients

Who was on the call, what matters to them, and what you have agreed over time.

Decisions and the why

What got decided and the reason behind it, so nobody argues it out again.

Concepts and methods

The ideas you teach, built up across every session you explain them in.

The teaching itself

The stories and metaphors you use to make the hard stuff land, in your own words.

Answers when you ask

Show me the last three meetings with this client. What did we agree? What got missed? You ask, and the wiki takes the shortest route to the answer.

Agents that read first

When a draft is needed, the agent pulls your real stories and phrasing from the wiki, then writes. That is why the work sounds like you.

Before and after

Ask about a client. Get the whole story.

Same question to both. One has seen one call. The other has every call, filed and linked.

  1. You

    Show me the last three meetings with this client. What did we agree? What got missed?

An AI note-taker

  1. Note-taker

    Here's the summary of your last call. I can only see that one call, so I can't say what came before it.

One guard watching one monitor. It never sees the other calls.

Your meetings wiki

  1. Meetings wiki

    Here are the last three meetings with this client, linked together. What you agreed over time, and why (from decisions and the why). What got missed. This is the client who went quiet in March (from across every call), and they're the one at risk in June.

  • Decisions and the why
  • Across every call

Every call is linked to the client, the people and the decisions.

An illustration, not a client result.
How it works

How a call gets filed.

Nobody does this by hand. It runs whether you show up or not.

Your old calls go in first

If you have a backlog of transcripts, we run it through on day one. The wiki starts full.

You run a call

Your recorder captures it. That is your whole job.

A routine picks it up

It runs on a schedule, with no prompt from you, and pulls the transcript.

The call is written up and linked

It becomes wiki pages, joined to the people, clients and topics it touches. The raw transcript stays as it is. That is the source of truth.

The rest of the OS uses it

The wiki is indexed again so it can be searched fast. The other parts of your OS read from it, and every call makes it bigger.

Who it's for

This is for you if

  • You run a proven, paying program, and your calls are where the teaching happens.
  • You have months of call summaries nobody reads.
  • Your team needs to know what was agreed with a client without asking you.
  • You want AI drafts built from your real stories, not generic copy.
Who it's not for

This is not for you if

  • You only want a note-taker app. Buy one. They are cheap and good at that job.
  • You want the meetings wiki on its own. It is not sold alone. It is one part of the OS.
  • Your main problem is getting more clients. This fixes delivery, not sales.
FAQ

Common questions.

A note-taker records one call and writes one summary. Then it stops. The meetings wiki picks up from there. It takes the transcript and files it into linked pages about your clients, your people, your decisions and your teaching. One summary stores one call. The wiki joins all of them, and your agents work from it.

The wiki does not record anything. It works from the transcript your recorder makes. In our public guide that recorder is Fathom. Which recorder you use, and whether it shows up in the call, is a choice we make with you during setup.

On your own Mac. The wiki is plain text files that sit on your machine. That is on purpose. Plain text survives any tool change, and any model can read it. If you stop working with us, the wiki stays with you.

Nothing on its own. The meetings wiki is one part of the AI Orchestrator OS. It is not sold alone. Ask about the price on a call.

Yes. The OS is shared across your coaches and consultants, and training them on it is part of the setup. They can ask the wiki what was agreed with a client and why, instead of asking you.

Yes. Our Meeting Wiki Brain guide gives you the full build prompt for free. The OS is for founders who want it built with them, joined to the rest of the business and shared with the team.

The meetings wiki is one part of the OS.

See what else is in the AI Orchestrator OS, and how the six setup weeks run.