The promise of AI meeting notes is simple: never scramble to write down what was decided again. The reality is often a second app you have to remember to invite, log into, and check afterward.

The notes are useful. The babysitting is not.

This guide explains how AI meeting notes actually work, what a good summary should capture, and how to get transcripts and action items without managing yet another tool.

Quick answer: An AI notetaker joins your Meet, Zoom, or Teams call, transcribes it, and writes a summary with decisions and action items. The friction is usually managing a separate app. An assistant like folk can auto-join calls on your calendar and deliver the notes straight to your texts, so there is nothing extra to open.

How do AI meeting notes work?

Under the hood, most AI notetakers do the same four things:

1. Join the call as a visible participant on Google Meet, Zoom, or Microsoft Teams.
2. Record and transcribe the audio into text, usually with speaker labels.
3. Summarize the transcript into key points, decisions, and action items.
4. Store it so you can search the transcript and revisit the summary later.

The model quality is broadly similar across tools now. What differs is the workflow around it: how the bot gets into the call, and where the notes end up.

What should a good meeting summary capture?

A summary you can actually act on includes:

• Decisions, stated plainly, not buried in a wall of transcript.
• Action items with an owner, so "someone will follow up" becomes "you agreed to send the deck."
• Open questions that were raised but not resolved.
• A searchable transcript for the details a summary leaves out.

If a tool gives you a transcript but no structured decisions and actions, you are still doing the real work yourself.

Is it legal to record meetings with an AI notetaker?

Recording rules vary by region, and some require consent from everyone on the call. Reputable notetakers join as a visible participant so it is clear recording is happening. Follow your local laws and your workplace policy, and let people know when a meeting is being recorded. The FTC's business guidance is a reasonable starting point for US norms, but check your own jurisdiction.

The real friction: managing another app

The notes are rarely the problem. The problem is:

• remembering to invite the bot to each call,
• logging into a separate dashboard to read the summary,
• and manually copying action items into wherever you actually track work.

Every one of those steps is a chance to fall off the habit.

How folk handles it from your texts

folk treats meeting notes as something that should come to you, not another destination to visit:

• Auto-join, set once. Connect Google Calendar and folk can join your meetings automatically (off by default, you opt in). No per-meeting invite step. See meetings.
• Notes in your chat. After the call, the summary, decisions, and action items land in iMessage, Telegram, or WhatsApp.
• Into memory. folk can fold action items into your memory so they resurface as reminders instead of dying in a doc.
• Ask about it later. "What did we decide about the launch date?" folk searches the transcript and tells you.

You still get a proper notetaker in the call. You just stop babysitting a separate app to use it.

The bottom line

AI meeting notes are worth it when the summary is structured and the notes come to you. The tool that wins is the one that removes the busywork around the recording, not just the note-taking itself.

If you want notes that show up in your texts with nothing extra to open, set folk up with your calendar. It is part of the same assistant that lives in your messages.