I’ve lost count of the times I’ve bought a slick productivity template, spent a whole weekend setting it up, and then stopped using it two weeks later. The “second brain” idea always sounded great. In practice, keeping it updated felt like a second job.
That’s why I liked this post from an AI professional who skips the paid Notion templates completely. The author’s view is that you don’t need them if you use Claude correctly. Their setup runs on tools most of us already open every morning: Slack, Gmail and Granola. Once those are connected, Claude stops being a blank chat window and becomes a memory you can ask questions.
The process is simple. Here’s how the creator does it, step by step, with the reason each step matters.
1. Connect your everyday tools to Claude
The author does most of their work in Slack, Gmail and Granola. All three are linked to Claude as connectors, which are integrations that let Claude read data from the apps you already use.
Why it matters: A second brain only works if the information is easy to reach. Templates make you copy notes in by hand. Connectors let Claude go straight to the source, so you don’t keep a separate system up to date.
2. Record every meeting with Granola
Whenever the post’s author joins a meeting, they open Granola in the background to record the transcript. They call it the only AI note-taker app they trust.
It also works on their phone, so WhatsApp calls and in-person meetings get captured too. Every conversation ends up in one place where Claude can access it.
Why it matters: Transcripts are the raw material for everything that follows. If half your meetings happen on your phone and never get recorded, you’ve lost half your context. Capturing everything in one tool fixes that.
3. Pull out deadlines, tasks and blockers
After the meeting, this LinkedIn creator opens a new chat in Claude and uses this prompt:
“Use Granola connector, go to my last meeting with [name]. List down the deadlines, tasks, and blockers. But clearly state if something wasn’t said clearly. Do not guess.”
Why it matters: Pay attention to the last two sentences. Telling Claude to flag anything unclear and “Do not guess” protects you from made-up details. AI models sometimes fill gaps with confident guesses. This wording makes Claude point out the gap so you can ask about it.
4. Cross-check against older meetings
Next, the expert asks Claude to check earlier meetings and see whether anything has changed that they might have forgotten.
Why it matters: Projects change slowly. A deadline moves in one call, the scope grows in another, and before long you’re working from old assumptions. Comparing meetings catches those changes before they cause trouble.
5. Teach Claude your writing voice
With all that context in place, the author has Claude draft Gmail replies that sound like them. That starts with a voice profile:
Prompt 1. “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.”
Why it matters: Most AI emails read like AI emails, and people notice. Basing the profile on 30 real sent emails gives Claude concrete patterns to copy, such as your greetings, sign-offs and the words you’d never use. Asking for the result as “writing instructions” makes it reusable in later chats.
6. Draft the meeting follow-up
Once the voice profile exists, the follow-up email takes one prompt:
Prompt 2. “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.”
Why it matters: A good follow-up email covers three things: decisions, owners and the next step. This prompt asks for exactly those. “Keep it as short as I keep mine” stops the usual AI padding, and “Create it as a Gmail draft” means nothing goes out until you check it.
7. Prep before the next call
Before the next call with the same person, this savvy professional asks one quick question:
“What did I promise [name] last time? What’s still open?”
Why it matters: Few things build trust faster than remembering what you promised someone. Thirty seconds of prep means you walk in knowing your open items, so nobody has to remind you.
🚫 The line the author won’t cross
There’s one firm rule in this setup. Never let Claude send Gmail replies on its own without your review. The original poster says that’s where they draw the line.
I fully agree. Drafting is where AI saves you time. Sending is where your reputation is on the line. One wrong deadline or a misread tone in a client email can undo weeks of goodwill. Keeping yourself as the final check costs a few seconds and keeps you in charge.
FYI: It was never about being lazy.
That closing line from the creator stuck with me. The point isn’t to hand off your thinking. It’s to stop spending your energy on recall and admin so you can focus on the conversations themselves.
💡 Tips for building your own version
- Start with one connector: If you’re new to this, connect Gmail first and get comfortable before adding meeting transcripts.
- Ask for consent before recording: Let people know when a call is being transcribed, especially in-person and on WhatsApp.
- Save your voice profile: Keep the output from Prompt 1 somewhere you can paste it into new chats, and refresh it every few months as your style changes.
- Swap the placeholder: Replace [name] with the actual person or company so Claude finds the right meeting.
- Keep the “Do not guess” habit: Add that line to any prompt where accuracy matters more than a complete-looking answer.
What I like most about this workflow is that you don’t need to learn a new app. It connects tools you already pay for and gives Claude enough context to be useful. Check out the full LinkedIn post to see the author’s setup in their own words and copy the prompts for yourself.