Claude + Granola: The 2-Minute Email Reply Hack

Replying to emails used to eat my mornings alive. Open inbox, scroll through threads, dig up the meeting notes, try to remember what was actually said, then stare at a blinking cursor for ten minutes. By the time I hit send, half my energy was gone and I hadn’t even started the real work.

Then I stumbled across this post from a LinkedIn creator who basically rebuilt how email replies work in 2026. I was blown away when I saw it. The whole loop takes about two minutes, the drafts sound like you wrote them, and you never leave Claude until it’s time to hit send.

Here’s the breakdown of what the author shared, plus why each step matters.

The 5-Step Workflow

1. Install Granola.ai (it’s free). This tool quietly transcribes every Zoom, Teams, or Google Meet call you join. No bot pops into the meeting, no awkward “recording started” announcement. People talk like humans, and you walk away with the real conversation captured word for word. The original poster’s rule: always open Granola before a call. That’s how you build the context library Claude will pull from later.
2. Connect Gmail and Granola inside Claude desktop. Go to Settings, then Connector, and link both. This is where most people mess up. They connect Gmail and call it done. The author makes a sharp point here: without Granola in the mix, Claude is drafting replies with zero memory of what was actually said in your meetings. It’s guessing. Adding Granola gives it the real source material.
3. Open the email you need to reply to and note two things: who sent it, and which meeting it ties back to. This sounds tiny, but this contributor explains it well. You’re telling Claude exactly where to look in Granola so it doesn’t have to guess. That alone makes the next prompt about 10x sharper.
4. Prompt Claude with full context. The author gives a clean template you can copy directly: “Pull [sender]’s email about [topic]. Search Granola for our meetings on [topic] in the last [timeframe]. Draft a reply that [confirms / answers / declines]. Tone: warm + direct. Length: short.” Notice how every variable points Claude somewhere specific. There’s no fluff for it to interpret.
5. Tell Claude to “save as draft.” It lands straight in your Gmail. Zero copy and paste. You open Gmail, give the draft a quick once-over, and hit send. The whole loop closes without you ever tab-switching out of Claude.

Why this beats writing from scratch

The magic isn’t that Claude writes emails. The magic is that Claude writes emails with the full context of what you actually discussed in your meetings. That’s the difference between a generic reply and one that sounds like you remember every detail of the conversation.

Think about how much of your inbox is follow-ups to conversations you already had. Sales call recaps. Hiring loop next steps. Project updates. Quick “as we discussed” notes. All of that becomes near-instant when Claude can actually see what was said.

The safety tip you shouldn’t skip

This savvy professional drops a warning that I think is gold: don’t grant write access on day one. Start with read-only permissions. Watch how Claude pulls information for a full week. Get a feel for what it sees, what it misses, and how it phrases things. Once you trust the pattern, then you upgrade to write access.

The reason this matters: most accidents happen when someone connects Claude with full permissions on day one and asks it to “clean up my inbox.” That’s how important threads get archived, sent items get weird, or drafts get fired off before they’re ready. Read-only first. Always.

Use cases this unlocks

• Client follow-ups after discovery calls, with the actual pain points they mentioned baked in.
• Hiring loop replies that reference what each interviewer flagged without you re-reading five sets of notes.
• Internal project updates where Claude pulls from your last three standups and writes a single clean summary.
• Polite declines that reference exactly what was offered so the no doesn’t feel cold.
• Reconnection emails to old contacts where Claude pulls the last conversation you had together.

A few things I’d add

Treat your Granola transcripts like a private knowledge base. The more meetings you run with it open, the more powerful every future email gets. Within a month, Claude will draft replies that sound like you’ve been taking obsessive notes for years.

Also worth testing: the same prompt pattern works for Slack DMs, internal memos, and recap docs. The structure (point Claude at the source, define the action, set the tone, set the length) is reusable across basically any written output.

I think the bigger takeaway from this post is the shift in mindset. Writing from a blank page is a leftover habit from before AI could see your context. Now that it can, blank pages are wasted reps.

Check out the full LinkedIn post from the original poster for the exact wording and a few extra notes I didn’t cover here. Worth a read if you live in your inbox.