Prompt Lab #6: Crafting AI-Powered Meeting Summaries That Actually Work

So that you can focus on what matters most ... listening

Image generated using Ideogram in Boodlebox.

Prompt Lab is a sub-section of Cyborgs Writing. If you don't want to receive prompts, you can unsubscribe to this sub-section and still get Cyborgs Writing key posts. See your Substack settings.

---

Hey there, fellow cyborgs!

Today, we're diving into a practical application of AI that's helped me be a better team leader: creating effective meeting summaries.

As the District C Fellow at UNCW and coordinator of the professional writing program, I find myself running more meetings than I'd like. And let's face it, taking notes while trying to actively participate is a juggling act we could all do without.

This is why I’ve been trying out various AI meeting tools that create minutes and summaries.

While the ability to easily transcribe meetings is wonderful, the actual AI summaries leave quite a bit to be desired.

These tools often miss the mark in identifying what's truly important. They'll flag casual comments as action items or overlook crucial decisions buried in the chatter.

The result? A summary that needs more editing than it's worth.

The Solution: A Structured Prompt

Drawing from my experiences with PromptOps and structured content (remember our discussion on modularity?), I've developed a prompt that significantly improves AI-generated meeting summaries.

The key? Narrowing the AI's guesses by providing context and structure.

Here's the three-part structure I use (see the prompt template at the bottom):

1. Agenda: First, I input the meeting agenda. This gives the AI a roadmap of what was supposed to be discussed, making it easier to identify important points. No agenda? Well, that's a problem whether or not you're using AI!

2. Notes: I add brief notes about any important topics that came up but weren't on the agenda. Yes, you're still taking some notes, but they're selective and focused.

3. Transcript: Finally, I include the full meeting transcript.

Pro tip: During the meeting, you can lay "breadcrumbs" for the AI. For instance, saying "Okay, so there is an action item" or "That's a really good point. Let me repeat that to make sure I've got it right." This can help the AI identify key information in the transcript.

Bonus … it makes you a good listener.

Why This Works

This structured approach gives the AI the context it needs to make more accurate judgments about what's important. It's not just sifting through a sea of words; it has a framework to guide its analysis.

The beauty of this structured prompt is its reusability. The structure stays the same; only the content changes from meeting to meeting. This opens up possibilities for automation.

Imagine having your AI automatically send the summary draft to your email after each meeting!

If you've been following our Cyborgs Writing journey, you'll recognize how this approach aligns with our discussions on structured content and modularity.

We're not just throwing words at the AI; we're providing a structured framework that enhances its ability to understand and summarize our meetings effectively.

[GOAL] Provide a concise summary of our meeting for those who did not attend and as a recap for those who did. Focus on key points.

[AUDIENCE] Members of the UNCW District C teamship who coach design thinking at UNCW.

[CONSTRAINTS] Do not include small talk.

[TASK] Use the transcript to develop an email summary. Focus only on what was on the agenda and in my notes. Refer to the agenda to guide your summary.

###

<agenda> insert agenda for meeting </agenda>

<notes> insert any notes you want the AI to highlight </notes>

<transcript> insert transcript </transcript>

What would you change or add about this prompt? I’d love to hear in the comments!