I used to think meeting notes were just part of the job. You sit there, half-listening, scribbling bullet points, and then spend another 20 minutes after the call turning them into something useful. Sound familiar? I felt seen when I read this post from an AI professional who admitted they repeat themselves to their AI tools almost 17 times a day. That tiny friction adds up fast.
The creator shared a workflow that basically turns Granola.ai into a shared memory for Claude. And once I understood how the pieces fit, I was honestly a little annoyed I hadn’t connected them myself. So let me break down what this LinkedIn user actually built, step by step, because each move has a clear reason behind it.
The core idea: stop repeating yourself
The whole workflow solves one problem. Your AI tools have no memory of your meetings, so you keep feeding them context over and over. The original poster’s fix is to capture every meeting automatically, then plug that context straight into Claude. No more re-explaining what was decided or who owns what.
Here’s the part I respect most: the author refuses to use meeting bots.
It doesn’t join in any of my meetings. I hate bots in the meetings. Please don’t do it. I focus on the meeting rather than remembering context.
Granola runs quietly in the background instead. It works on Zoom, Meet, and Teams (the creator uses Meet), and it never announces itself in the call. That’s a small detail, but it keeps meetings human while still capturing everything.
The setup, step by step
What I like about this expert’s approach is that every step earns its place. Nothing is filler. Here’s the flow they laid out:
- Download Granola from granola.ai.
- Treat it like a notepad, similar to Apple Notes.
- Open it in every single meeting, non-negotiable, so nothing slips through.
- Let it capture all transcripts with timestamps automatically.
- In Claude, go to Settings, then Connectors, then add Granola.
- Now anything you draft with Claude already carries the context from your meetings.
- Run the voice-profile prompt once (below) to teach Claude how you actually write.
- Save that output as project instructions so it sticks around.
Step 5 is the hinge that makes the whole thing work. Connecting Granola to Claude means your transcripts become searchable context. Step 8 is the quiet genius move: saving the voice profile as project instructions means you set it up once and never touch it again.
The one-time voice prompt
This is the prompt the creator says you run a single time to build your writing profile. Copy it exactly:
Read my last 30 sent emails in Gmail. Write my voice profile: how I greet, how I sign off, my sentence length, my go-to phrases, the words I never use. Format it as writing instructions to follow when drafting as me.
I think this is the smartest piece of the whole system. Instead of guessing at what “sounds like you,” it pulls from 30 real emails you already sent. That’s why the drafts come out sounding like a person and not a corporate template. As the author put it, this is why your emails will actually sound like you.
The after-meeting follow-up prompt
Once the setup is done, this is the prompt you paste after any meeting. The original poster reproduced it clearly, so here it is exactly as shared:
Pull my last meeting from Granola. Draft the follow-up using my voice profile: what we decided, action items with owners, one clear next step. Keep it as short as I keep mine. Create it as a Gmail draft to the attendees.
Notice how specific this is. It asks for decisions, action items with owners, and one next step. That structure keeps follow-ups short and useful instead of a wall of text. From there, you open the Gmail draft, drop the attendees in the To field, and you’re basically done.
The weekly accountability check
There’s one more prompt this savvy professional throws in, and it’s my favorite for staying on top of loose ends. Once your meetings live in Granola, you can ask Claude:
Which action items from this week’s meetings are still waiting on me?
That single question turns a week of scattered calls into a clean to-do list. No digging through notes. No forgotten promises. The context is already there.
Why this matters
The magic isn’t any one tool. It’s the connection between them. Granola remembers the conversation, your voice profile remembers how you write, and Claude stitches both together into a draft that’s ready to send. The person who shared this said their recaps just became a mere to-do list, and that’s exactly the payoff.
A few things I’d flag if you want to try it:
- Consistency is the price of entry. The author calls opening Granola in every meeting non-negotiable, and they’re right. Miss a meeting and that context is gone.
- Do the voice profile properly. Pull real sent emails, not drafts. The quality of your output depends on the quality of that sample.
- Save it as project instructions. This is what separates a one-off trick from a system you actually keep using.
I love how the creator framed the mindset behind it. As they put it, they’re not lazy, they just hate doing the small stuff that AI can handle. That’s the whole point. Offload the busywork, keep your attention on the actual conversation.
If you run a lot of meetings, this is one of the cleaner note-to-follow-up systems I’ve seen shared lately. Go check out the full LinkedIn post for the creator’s exact setup and give it a shot before your next call.