Appearance
What Is an AI Meeting Assistant?
An AI meeting assistant is software that joins your video calls (Zoom, Google Meet, Teams) and helps with meetings. Most focus on documentation — transcription, note-taking, summarizing. A few try to do actual work during the meeting.
Two Categories
Documentation tools (most of them)
These capture what happened and process it after the meeting ends:
- Transcription — a text record of what was said
- Summaries — key points and decisions extracted
- Action items — "I'll send that by Friday" pulled into a list
- Search — query across your meeting history
Value arrives 10 minutes after the meeting ends, in your inbox.
Execution tools (newer)
These try to do work while the meeting is happening:
- Research — "How does Figma handle this?" answered in 30 seconds
- Artifacts — documents, specs, mockups created live
- Shared workspace — everyone sees what the AI is working on
Value arrives during the call, before you move on.
Why This Distinction Matters
Every meeting creates work:
- "Can someone look into how competitors price this?"
- "We should create a ticket for that"
- "Let's follow up on X"
With documentation tools, that work still happens after the meeting. Someone has to remember, find time, and do it — often with half the context lost.
With execution tools, some of that work happens immediately. The research appears in 30 seconds. The document is created while everyone's still on the call. The gap between "we need X" and "we have X" collapses.
Common Tools
Otter.ai — The original. Excellent transcription accuracy, good search, solid free tier. Primarily focused on creating a searchable archive.
Fireflies.ai — Good all-rounder with lots of integrations (CRM, Slack, Zapier). Works for a variety of team types.
Fathom — Built for sales teams. Fast summaries, CRM integration (Salesforce, HubSpot), coaching features.
Grain — Focused on video clips. Great for sharing key moments with stakeholders who weren't in the meeting.
Potato — Does work during meetings: research in 30 seconds, artifacts created live, shared workspace everyone sees. Experimental — sometimes impressive, sometimes misses things.
Do You Need One?
Probably useful if:
- You have a lot of meetings and struggle to track what was decided
- You find yourself taking notes instead of participating
- Meetings generate follow-up work that takes days
- You want questions answered immediately, not researched later
Probably not necessary if:
- You have few meetings
- Your meetings are casual/social
- You already have a good system for capturing and executing on decisions
What Potato Does Differently
Most meeting assistants give you output after the meeting — a transcript, a summary, action items in your inbox.
Potato does work during the meeting:
Research in 30 seconds — "What's the market size for X?" — answered with sources before you change topics. That's research that would take 2 hours, done while everyone's still on the call.
Shared workspace — Everyone sees what the AI is working on. One view. Anyone can say "that's wrong" or "add context." Not personal notes that differ between people.
Artifacts created live — Need a spec, a summary, or a mockup? It appears while context is fresh — not reconstructed from notes later.
It's experimental. Sometimes it's shockingly good. Sometimes it gets confused. We're improving it every day.
No signup required to try it — just paste a meeting link at meetpotato.com.