Chapter 10 | Part 3: Operating

Meetings and Memory

Fathom, Otter, Fireflies. Setup in ten minutes. The meeting summary arrives before the call ends.

6 min read

Every meeting produces three things: decisions, action items, and context. Most organizations reliably capture zero of the three.

The Cost of Not Capturing Meetings

A one-hour executive meeting involves people whose time costs real money. That meeting produces decisions, commitments, and context that the organization needs to operate on for the next week, quarter, or year.

What typically gets captured: a vague set of notes, if anyone wrote them. What gets lost: the exact wording of the decision, the reasoning behind it, who committed to what, the nuance that would have prevented the misunderstanding three weeks later.

AI meeting tools solve this without changing how you meet.

They join the call silently, transcribe everything, and return a structured summary — decisions made, action items with owners, key points raised — before the meeting has fully wrapped. The record exists. It's searchable. It can be shared with the people who weren't in the room.

This is not a productivity trick. It's institutional memory.

The Tools

Fathom is free, installs as a browser extension, and works with Zoom, Google Meet, and Microsoft Teams. Setup takes ten minutes. From that point forward, any call you join can be recorded and summarized automatically.

What you get from each meeting:

  • A full transcript (searchable)
  • A structured summary with key points and action items
  • Timestamped moments you can click back to in the recording
  • A sharable link you can send to participants

The summary arrives before the call ends. By the time you're on the next task, the record already exists.

Fathom is the right starting point for individual executives. No team-wide deployment needed. No IT approval required for basic use. Start with your own recurring meetings.

Otter.ai

Otter covers both video calls and in-person meetings. The mobile app transcribes live in the room, which makes it the right choice if your team holds physical meetings that need to be captured.

Paid tiers add features like meeting highlight tagging, comment threads within transcripts, and integration with Salesforce for sales teams. The free tier is functional for most basic use.

Best for: executives who move between video calls and in-person sessions and want consistent capture across both.

Fireflies.ai

The enterprise option. Fireflies connects to your calendar, joins every meeting automatically, and builds a searchable database of all your organization's meeting history. CRM integration (Salesforce, HubSpot) means call summaries flow directly into deal or account records.

Best for: companies where meeting intelligence needs to be systematically distributed and integrated — sales teams tracking client calls, or organizations that want cross-functional visibility into meeting outcomes.

Start with Fathom individually. Evaluate Fireflies when you want team-wide deployment with system integration.

What the Summary Actually Gives You

A good AI meeting summary is not a transcript. It's a structured document covering:

Decisions made: The explicit choices the group reached. "We agreed to delay the Panama expansion until Q3." "The proposal from Carrier A was rejected; we'll proceed with Carrier B."

Action items with owners: What each person committed to. "Marcus to send the updated rate sheet by Friday." "Lucas to schedule the site visit in Q2."

Key discussion points: The main arguments or concerns raised, without the full back-and-forth.

Open questions: What was raised but not resolved. "We still need clarity on the regulatory requirements before committing to the international lane."

This is the document your EA would have spent 30 minutes writing. It arrives automatically, within minutes.

How to Use the Output

Share immediately

Send the summary to all participants before the day is out. This serves two purposes: it confirms what was decided (preventing the "I thought we agreed to X" conversation three weeks later) and surfaces corrections while people still remember the meeting.

A 10-second friction — hitting send on the Fathom link — prevents hours of re-work.

Store where people can find it

The summary is only useful if it's accessible. Options:

  • A shared meeting notes folder in Notion or Google Drive, organized by date and meeting type
  • Directly in the relevant project or client record
  • In your calendar as an attachment to the meeting event

The method matters less than the consistency. If summaries live in three different places, nobody knows where to look.

Use it to build your own context file

If you have recurring client relationships, keep a running document per client: key decisions, what you've discussed, what they care about, what's been promised. Paste the relevant action items and decisions from each meeting summary into that document.

After six months, you have something that no amount of memory can replicate: a reliable record of everything that was said and committed. Before every client meeting, you read that document in five minutes instead of spending 20 minutes trying to remember.

This is institutional memory that actually persists when people leave.

Setting It Up: Ten Minutes

Fathom on Zoom:

  1. Go to fathom.video, create a free account
  2. Connect your Zoom account (OAuth — a few clicks)
  3. In Fathom settings, set recording to "automatic for all meetings" or "ask each time"
  4. Join your next call — Fathom appears as a participant, introduces itself in chat
  5. After the call: the summary and transcript are in your Fathom dashboard within minutes

The only recurring step: if someone in the meeting asks about the recording, explain it's there for notes, not surveillance. Have a one-line answer ready. "We use Fathom to capture action items — the summary goes to everyone on the call."

Always tell meeting participants that a recording tool is running.

This is not optional. In many jurisdictions, recording calls without consent is illegal. Beyond legality: people speak differently when they know they're on the record. They should know.

Most tools (including Fathom) announce themselves when they join the call. That announcement counts. If you're using a tool that doesn't announce, announce it yourself.

"I use a meeting tool that captures notes — you'll get the summary afterward." That's enough.

For sensitive conversations — negotiations, performance discussions, anything where candor depends on informality — consider turning the tool off. The presence of a record changes the conversation. Use judgment about when the record is worth more than the openness.

Beyond Calls: Personal Meeting Prep

The capture tools handle the output of meetings. There's also an AI application on the input side: preparation.

Pre-meeting brief from AI: Before a significant client meeting or pitch, paste what you know about the person or company into Claude and ask: "I'm meeting with person from company in an hour. Here's what I know. Give me: three things I should ask, two things I should listen for, and the one thing I need to make sure I communicate."

This takes five minutes. It produces better preparation than 30 minutes of unfocused review.

Combine it with the running client context document you've been building from past meeting summaries. You walk in prepared in a way that demonstrates you remember every conversation — because you have a reliable record of every conversation.

The Bigger Picture

The constraint on most executive decision-making is not information — it's reliable memory of past information. What did we decide three months ago? What did the client say they needed? What did the analyst tell us in the briefing?

AI meeting tools don't make you smarter. They remove the friction between information and memory. Decisions you made, commitments you took, context you gathered — it's there, searchable, whenever you need it.

A year of consistent meeting capture produces an organizational intelligence layer that would have taken a team of analysts to build manually.

Set it up today. It runs in the background from then on.

Next: The Practice

Stay in the loop

Occasional updates on AI systems, tools, and new writing.

Ormus — Diego Bodart