How to send expert meeting follow-ups with AI.

I use this to send the follow-up that actually advances the deal, fast, every time. The chain is Granola for what was said, Airtable to log it, Gmail to send it. Granola captures the meeting; Claude reads those notes and writes a follow-up that proves I listened - their words, the specific pain, a useful next step - logs the outcome to the CRM, and drops the email in Gmail ready to send. I reach for it the moment a call ends, while it is fresh, for every meeting that deserves a follow-through.

How I use this, and what for

I use this to send the follow-up that actually advances the deal, fast, every time. The chain is Granola for what was said, Airtable to log it, Gmail to send it. Granola captures the meeting; Claude reads those notes and writes a follow-up that proves I listened - their words, the specific pain, a useful next step - logs the outcome to the CRM, and drops the email in Gmail ready to send. I reach for it the moment a call ends, while it is fresh, for every meeting that deserves a follow-through.

What the AI actually does

The leverage is the expert follow-up without the lag. The follow-ups that win are specific, value-adding and sent within the hour - and that is exactly what slips when you are busy. Claude turns the Granola notes into a follow-up that references what they actually said, adds something useful, and proposes a clear next step, then logs it and stages it in Gmail. You go from "I should follow up" to a drafted, logged, ready-to-send email in minutes - and it never reads like a template.

The notes are reliable because of how the capture works. Granola does not join the call as a bot - it runs on your own device, listens through the system audio, and merges what you typed with what it heard, so you finish a meeting with real notes already written. If the call lived in Fireflies instead, that works the same way: it records, transcribes and makes every meeting searchable, so AI has a clean transcript to build the follow-up from.

The tools and APIs you need

• Granola
• Fireflies
• Airtable
• Gmail
• Claude

The right prompts

Whole chain from a finished call

```
My call with Acme just finished - it is in Granola. Read the notes, then:

1. Write an expert follow-up email in my voice - reference what they
   actually said, add one useful thing, propose a next step with a date.
2. Log the outcome and next step to the ATS Airtable base.
3. Create it as a Gmail draft to my main contact there for me to send.
   Keep it short and human, no template feel.
   ```

Add value, not just a recap

```
For the Acme follow-up, do not just summarise the call - they were worried
about ramp time, so include one concrete point on how meritt shortens
ramp, and link the relevant proof. Then draft it in Gmail.
```

How to do it

Find the meeting in [Granola](https://www.granola.ai/)

- the most recent call with that company.

Have Claude read the notes

and pull what matters: the real pain, what they cared about, what was agreed, the next step.

Write the expert follow-up:

open by referencing something specific they said, add one genuinely useful thing (a resource, a relevant proof point, a clear answer to their concern), and propose the next step with a date. Short, in your voice, no template smell.

Log it to [Airtable](https://www.airtable.com/)

- outcome, next step, stage, next-step date - so the CRM is current.

Draft it in [Gmail](https://workspace.google.com/products/gmail/)

to the right person, ready for you to review and send.

Review and send

- one read, then out, while the call is still warm.

See the difference

Let AI carry the heavy lifting; you keep the judgement and the final say.

How you'll know it's working

• A follow-up that proves you listened, sent within the hour, not days later.
• It adds something useful - it is not a meeting recap.
• The CRM is updated in the same move, not as a separate chore.
• It reads like you wrote it, because it was in your voice.

Questions people ask