4.4 Fellow.ai Setup and Coverage Guide

What is Fellow.ai?

Fellow.ai is your personal AI note-taker and meeting intelligence platform. As an architect, it is one of the most valuable tools in your day-to-day workflow. It joins your client meetings automatically, records and transcribes everything, and then uses AI to organize the content into structured notes — summaries, action items, decisions, and more.

Beyond note-taking, Fellow.ai can help you organize your meeting transcripts, draft follow-up emails, create structured documents from call content, and surface what was discussed across multiple meetings with a client. Think of it as a second brain for every client conversation — one that never forgets and is always ready to help you prep for the next call.

How to use Fellow.ai — getting started

Fellow.ai is designed to work with as little setup as possible. Here is how to get it into your meetings:

Automatic join: Fellow.ai will automatically join any meeting where you are invited, as long as the meeting is on your connected calendar and the client has not restricted recording. You do not need to do anything — it shows up, takes notes, and sends you the output when the meeting ends.

Manual join: If Fellow.ai did not join a meeting automatically (for example, if you were added late or the calendar sync missed it), you can invite it manually. Log into Fellow.ai, find your meeting in the meetings list, and click to request that the Fellow bot joins. You can do this even mid-meeting.

After the meeting: Once the meeting ends, Fellow.ai processes the recording and generates AI notes — typically within 15–30 minutes. You will see the summary, action items, and decisions organized and ready to review. This is also the moment the Make.com automation fires to create your cycle planner draft and recap email draft.

What Fellow.ai can do for you as an architect

Fellow.ai is far more than a recording tool. Here are the capabilities worth knowing and using regularly:

• AI meeting summary — automatically generates a structured summary of what was discussed, organized by topic. Saves you the mental overhead of reconstructing a 60-minute conversation after the fact.
• Action items extraction — pulls out every committed action item from the transcript and assigns them to the right people. Review and clean these up — they feed directly into the cycle planner automation.
• Decisions captured — flags key decisions made during the call so they are never lost in a long transcript.
• Full transcript — searchable word-for-word record of the entire meeting. Useful when you need to go back and verify exactly what a client said or agreed to.
• Email recap generation — Fellow.ai can draft a recap email directly from the meeting content and export it to Gmail with one click. This works in parallel with the Make.com automation — use whichever is faster for you.
• Document export — Fellow.ai document widgets can be exported to Google Docs with one click, turning meeting notes into structured client-ready documents.
• Meeting prep — before a call, Fellow.ai shows you the agenda, the notes from the previous meeting with that client, and any open action items. This is your pre-call briefing without needing to dig through emails or HubSpot.

How Fellow.ai connects to the RevGravy automations

When a meeting ends and Fellow.ai finishes processing the notes, it sends a webhook (an automatic notification) to Make.com. This webhook triggers two automations simultaneously:

1. Agenda and Recap email auto-draft — generated from the Fellow summary, action items, and decisions. Posted to the client's Slack channel for the architect to review and send.
2. Cycle planner auto-draft — generated from the Fellow action items and transcript. A copy of the previous cycle planner is populated with new content and linked to a new HubSpot Project.

This means the quality of your Fellow.ai notes directly affects the quality of the auto-drafts. The better Fellow captures the meeting — clear action items, decisions flagged, key topics covered — the better the automation output.

Meeting naming requirement

⚠️ Critical for automations to work. All architect-client meetings must be named using this format:

Customer Name / RevGravy - Bi-Weekly Meeting

Examples: MTMC / RevGravy - Bi-Weekly Meeting, Bluegrass / RevGravy - Bi-Weekly Meeting

Make.com uses the meeting name to identify which client the Fellow notes belong to and route the outputs to the correct Slack channel and HubSpot project. If a meeting is named something like "Customer Name / RG Meeting" or anything that deviates from the standard format, the automation cannot match it to a client and will fail silently — no planner draft, no recap draft.

Coverage gaps — clients where Fellow.ai cannot be used

Two clients at RevGravy cannot use Fellow.ai for specific reasons. For these clients, all automations (agenda draft, recap draft, cycle planner draft) are disabled. Architects handle everything manually.

Do not count these meetings against automation adoption rates in the Ops Dashboard. They are intentional exclusions, not failures.

If Fellow.ai doesn't join a meeting

If Fellow.ai failed to join automatically and you missed the window to invite it manually, here is what to do:

1. Log into Fellow.ai and find the meeting in your meetings list
2. Check if the recording exists — sometimes Fellow joins late and still captures part of the meeting
3. If no recording exists: take notes manually and create the cycle planner and recap manually using the previous planner as your template
4. Check whether the calendar integration is still active — if Fellow.ai has stopped auto-joining multiple meetings, the calendar sync may need to be reconnected

MacWhisper — alternative transcription for Fellow-blocked meetings

Juan uses MacWhisper, a local audio transcription tool, for meetings where Fellow.ai cannot join (primarily BanOpen). MacWhisper records and transcribes locally without any meeting bot joining.