Fixing a Broken Meeting Notes Workflow

⏱️ Read time: ~3 min

Today I'm sharing a real AI use case I'm working on right now. With Monica, a coach and leadership team development expert from Sweden who's participating in AI Lab for 6 months.

Yesterday, she showed me her meeting notes workflow: she records the call, a custom GPT summarizes it, then she manually edits and copy-pastes action items into Loop.

On paper? A solid AI-powered system.

In practice? Still eating 30 minutes per meeting on cleanup.

Sound familiar? Here's how we approached the challenge.

From "AI That Doesn't Deliver" to a 3-Minute Review Workflow

The Before:
Monica's meeting notes process looked like this:

• Record & transcribe
• Have an AI-Assistant generate a summary that was too general or missed key details
• Manually edit and reformat
• Copy action items into Loop
• Repeat for every meeting

Total time per meeting: ca. 30 minutes of cleanup work.

The Problem We Defined:
Before we touched a single setting, we got clear on what success actually looked like:

• Clean, accurate client-ready summaries 
• Action items ready to transfer (or auto-synced)
• Maximum 3-minute review/editing time per meeting

Not "AI doesn't work for this." Not "try a new tool."

Instead: a specific outcome tied to how Monica actually works.

The Iteration Process:
Once we knew what we were solving for, we moved fast:

• Made live changes in the tool set-up (think: note-taking and summary style)
• Agreed additional settings for Monica to try (e.g., adding custom vocabulary, identifying common meeting types and adding context to the tool) 
• Reviewed the AI-Assistant and identified potential tweaks to the system prompt

We didn't get it perfect in one session. We got it testable and now Monica's running experiments across different meeting types to see what works.

The Messy Middle:
This is where most people stall. They tweak settings randomly. They switch tools. They assume "AI just doesn't work for me."

But transformation doesn't happen in the setup, it happens in the iteration. Monica's logging what works, what doesn't, and what surprises her. We're building her workflow based on real data, not assumptions.

That's the difference between adopting AI and actually implementing it.

Key Takeaway

Stop tweaking tools blindly. Start by defining the problem you're solving and the outcome you're after. Then test, iterate, and refine based on what actually happens.

AI implementation isn't plug-and-play. It's strategic, messy, and iterative. And that's exactly where the magic happens.

Have fun experimenting,
Elena

P.S. Want to see a workflow like this come together live? I'm hosting a free session on this soon - stay tuned for the registration mail. (And if you're curious about the kind of hands-on work Monica and I are doing in AI Lab, just reply and I'll share more.)

"AI is the most powerful tool of our time.

It's not here to replace you. It's here to free you, so you can focus on high-impact work, serve your clients better, and finally get your time back."