The short answer
The best AI for meeting notes in November 2026 is Fireflies.ai, because it transcribes more than 100 languages, says meeting content is never used to train AI models, and can email participants a decline link before the call. Otter.ai is the runner-up for English-speaking sales teams, with Work Mode since 27 October 2026. The best free option is Granola's Basic plan, which records without a bot. Always tell everyone an AI is taking notes.
| Our pick | Fireflies.ai: 100+ languages, never trains on meeting content |
|---|---|
| Runner-up | Otter.ai: Work Mode for prep and follow-up since 27 October 2026 |
| Best free option | Granola Basic: bot-free, audio deleted after transcription, 30 days of history |
| For Notion teams | Notion AI Meeting Notes on Business and Enterprise, no bot, 16 languages |
| Consent | Tell everyone, every time; some US states require all participants to agree |
The best AI for meeting notes in November 2026
An AI notetaker records a call, transcribes it and writes a summary with action items. They all do that. What separates them is how they join the meeting, how many languages they handle, what they do with your recordings and how well they help you tell everyone that a machine is listening.
- Our pick: Fireflies.ai, for transcription in more than 100 languages, a clear consent email with a decline link, and a policy of never training AI models on meeting content.
- Runner-up: Otter.ai, for English-speaking sales and recruiting teams that want help before and after calls, through Work Mode launched on 27 October 2026.
- Best free option: Granola, for bot-free personal notes from your own device, with audio deleted after transcription.
Our pick: Fireflies.ai
Why. Fireflies.ai records and transcribes calls either by sending its Fireflies Notetaker bot or, on the Mac and Windows desktop app, by capturing system audio so no bot joins. It transcribes more than 100 languages with automatic detection, connects to more than 100 tools including HubSpot, Salesforce, Slack and Notion, and has an MCP connection to Claude and ChatGPT so they can read your meeting history. Fireflies says meeting content, including audio, transcripts and summaries, is never used to train any AI models, and that AI vendors such as OpenAI and Anthropic work under zero data retention. Its compliance notification emails participants one hour before the meeting with a decline link that stops Fireflies joining. Pro adds unlimited AI summaries, 8,000 minutes of storage per seat, video recording and downloads.
Who it suits. Teams that meet in several languages, handle client data or want meeting notes flowing into their CRM. Read our Fireflies guide and Otter vs Fireflies.
The catch. For accounts created on or after 1 September 2026, free unlimited transcription only applies while the Auto-join calendar meetings toggle is on; switch it off and you get 3 meeting transcriptions. That nudges you to let a bot into every meeting, which is exactly what makes consent hardest to manage. Bot-free desktop recordings save no audio and label speakers as Speaker 1 and Speaker 2.
Runner-up: Otter.ai
Why. Otter.ai joins Zoom, Microsoft Teams and Google Meet as Otter Notetaker and transcribes live. Work Mode, available to all paid users since 27 October 2026, prepares a brief before each meeting from past calls and open action items, then drafts follow-up emails, updates CRM records and produces documents such as proposals afterwards. Live Assist adds private coaching prompts during calls, starting with Enterprise. On Enterprise, Otter's pricing page also lists AI Chat with Gmail, Google Drive, Notion and Jira.
Who it suits. English-speaking sales, customer success and recruiting teams who want the follow-up done for them. See the Otter guide.
The catch. Otter's privacy policy (effective 16 June 2026) lists training its own AI on de-identified audio recordings and transcriptions, and no opt-out toggle is described. It faces a class action, In re Otter.AI Privacy Litigation, alleging Notetaker recorded people without the consent of all participants; on 13 August 2026 Judge Eumi K. Lee granted Otter's motion to dismiss in part and denied it in part, so claims under the ECPA, CIPA and Illinois BIPA, among others, can proceed. There is no finding of liability. Its pricing page lists six transcription languages, and the free Basic plan caps each conversation at 30 minutes. Our Otter alternatives page lists other options.
Best free option: Granola
Why. Granola describes itself as an AI notepad for back-to-back meetings. It does not join your meeting: it transcribes your computer's audio in the background, so it works with Zoom, Google Meet, Teams and any other meeting app, and with the mobile app for in-person meetings. You type as much or as little as you like, and Granola combines your notes with the transcript into enhanced notes. The free Basic plan includes bot-free device audio capture, AI chat within and across meetings, apps for desktop, iOS, Android and Apple Watch, and Slack. Granola says audio is deleted after transcription on every plan, and that third parties such as OpenAI or Anthropic may not train on your data.
Who it suits. Individuals who want private notes from their own calls with nothing visible in the meeting.
The catch. Basic keeps only 30 days of meeting history; unlimited history, Notion and HubSpot integrations and follow-up emails need the Business plan. Granola trains on anonymised user data unless you opt out in Settings (Enterprise has training off by default). And because no bot appears, nobody knows you are transcribing unless you tell them.
Already paying for Notion?
Notion's AI Meeting Notes, on Business and Enterprise, transcribes and summarises meetings without a bot. You start it with /meet, the desktop app captures your microphone and system audio, it supports 16 languages and allows 10 hours per user per day. If your team already lives in Notion, the notes land next to the project they belong to, which is often worth more than a dedicated tool. See our best AI for note-taking page.
Consent: tell people, every time
Recording laws differ by country and state, and this is not legal advice, but the practical rule is the same everywhere: tell everyone that an AI is recording and taking notes, and get consent where the law requires it. Otter's help centre explains that some US states, naming California, Illinois and Florida, require every participant to agree, and suggests applying that standard to every call. Fireflies says that if you turn off its consent and notification options, you are responsible for informing participants and complying with recording laws. Granola recommends asking permission before transcribing others.
Each tool gives you a way to do it, and the bot-free ones need it most:
- Fireflies: Settings, Compliance Notification sends an editable email one hour before the meeting with a decline link, and pins a notice in the meeting chat. On the desktop app, Settings, Meeting Preference, Recording has a spoken consent message, an on-camera watermark and a chat message (macOS only), all on by default. Leave them on.
- Otter: turn on Send pre-recording emails under Account Settings, Meetings, and every calendar guest gets an email 5 to 10 minutes before. Enforcing it for the whole workspace is Enterprise only. The host can remove Notetaker like any attendee.
- Granola: turn on its transparency features: an automatic message in the meeting chat when it starts transcribing (Granola's help pages say this works only on some meeting apps on macOS, and not on iPhone) and the Granola Watermark on your video. Workspace admins can switch them on for everyone.
Also check where notes go. Otter, for example, can auto-share notes with attendees; you can stop that under the gear icon on the home calendar, Default audience for shared notes.
How to choose an AI meeting notetaker: three questions
- Bot or no bot? A visible bot tells everyone a recording is happening. Bot-free capture is less intrusive but puts the duty to tell people entirely on you.
- Which languages do you meet in? Fireflies lists more than 100; Otter's pricing page lists six; Notion lists 16. Check your languages before you buy.
- What happens to the recording? Fireflies says it never trains on meeting content; Granola deletes audio and lets you opt out of training; Otter trains on de-identified recordings with no toggle. For client or HR calls, this should decide it.
A meeting notes workflow, end to end
- Before the call: make sure the consent email or notification is on, and add a line to the invite such as "An AI notetaker will record and transcribe this call; reply if you would rather it did not."
- At the start: say it out loud. Remove the notetaker if anyone objects.
- During the call: say decisions and owners clearly ("So Priya owns the draft by Friday"). Notetakers capture what is said, not what is meant.
- Straight after: read the summary and fix names, numbers and owners before it goes anywhere.
- Follow up: ask the notetaker's chat for a follow-up email with decisions and action items, and edit it in your own words.
- Later: ask across meetings, for example "what has the client asked for in the last three calls?"
Our upcoming guide on how to transcribe a meeting with AI and the meeting notes prompts go further.
Prompts to copy
Summary that people will read
Summarise this meeting for people who were not there: the purpose in one line, the three decisions made, the action items with owner and due date, and any open questions. Keep it under 150 words.
You get: A short summary a busy colleague will actually read
Follow-up email in your voice
Draft a follow-up email to [attendees] from this meeting. Thank them briefly, confirm the decisions, list the action items with owners and dates, and propose [next step]. Plain, friendly tone, no jargon.
You get: A ready-to-edit follow-up email
Check what was actually agreed
List every commitment anyone made in this meeting, quoting the words used and the timestamp. Mark which ones have a clear owner and date and which are vague.
You get: A quoted list of commitments that shows which are firm
Track a client across calls
Across all meetings with [client] since [date], what have they asked for, what concerns have they raised more than once, and what did we promise? Give the meeting and date for each point.
You get: A cross-meeting view of one relationship
Prepare for the next meeting
I am meeting [person or team] about [topic] on [date]. From our previous meetings, list open action items, unresolved questions and anything they said they would come back to us on.
You get: A pre-meeting brief built from past calls
Spot the missing decision
Read this meeting transcript and tell me which topics were discussed at length but ended without a decision or owner. Suggest one question to settle each.
You get: A list of loose ends to close before they cause trouble