The Power of an LLM Debrief

I spend a lot of time thinking about what works in my classroom and what doesn’t. Sometimes it’s obvious - like when a project catches fire and students can’t wait to dive in. Other times, it’s harder to pin down why something that looked great on paper just didn’t land. I need help to see things clearly. That’s why I’ve always valued getting feedback on my work.

Like most people, I regularly turn to other people for their perspectives and spend plenty of time in self-reflection. But both approaches have their limits. My colleagues are wonderful, but I try to be strategic about when I ask for their time. And while I can spend hours thinking about my own work, I’m still stuck within my own perspective.

That’s why when I read Ethan Mollick’s fantastic book Co-Intelligence: Living and Working with AI, something clicked. His first principle for working with AI is to always invite it to the table, and so I wondered: what if I tried this with an end-of-class reflection? I decided to experiment with an LLM debrief of this year’s 8th Grade Tech Projects class because it had just wrapped up. What it surfaced genuinely surprised me.

Setting the Stage

I quickly realized that for this to work, the LLM needed a lot of context. So I shared everything – my curriculum, my syllabus, even a class-by-class breakdown of how the course actually unfolded (which, like every year, was different from my initial plans). Most major LLMs now have ways to give them this kind of background without cluttering up your conversation. ¹

With all that context in place, I started simply: “What do you see here?” Sometimes the most powerful prompts are the simplest ones. By leaving it that open-ended, I gave the LLM space to surface patterns and perspectives I might have missed on my own. ²

Learning to Dance with an LLM

The magic really happened in the back-and-forth that followed. Of everything the LLM initially pointed out, only about four out of ten things really resonated with me - and that’s okay! I used those points as jumping-off places for deeper exploration.

I’ve noticed that when many people interact with LLMs, they treat the responses as final answers. Maybe it’s because these tools can sound so authoritative, or because we’re not used to pushing back against computer output. But I found myself saying things like “that’s not what I’m looking for” or “that wasn’t quite what I expected” - redirecting, clarifying, and questioning – and that’s when the conversation got interesting.

This kind of dialogue let me blend my teaching experience and classroom knowledge with the LLM’s broader knowledge and pattern recognition abilities. It became more and more compelling as I started to surface and explore possibilities I wouldn’t have considered otherwise.

Unexpected Breakthroughs

I’m sharing this experience in detail because of the results, which were fantastic. Not because the LLM had all the answers, but because our dialogue helped me see my challenges in a completely new light.

Take my ongoing quest to foster collaboration in my classroom. I’d been starting every class with an open invitation for group consultation - a time when students could bring up challenges they were facing and get input from their peers. Sounds great in theory, right? Except nobody ever took advantage of it. Like, ever. Looking back, I can see why asking eighth graders to make themselves vulnerable in front of the whole class might not have been the most effective approach!

Through my conversation with the LLM, we landed on something I never would have thought of: a skills board. It’s a space where students can indicate areas where they’re willing to help others - maybe they’re good with Canva, or know some coding, or have experience with video editing. Instead of putting anyone on the spot, it lets students quietly connect with peers who actually want to help. Even better, as students learn new skills throughout the course, they can add themselves to new areas of the board. It becomes this visible record of growing confidence and capability.

We also dug into my documentation system. I already had some good practices in place - end-of-class reflections, daily “win” goals, and thumb checks for how students are feeling about their work. But something was missing. The LLM helped me envision a more cohesive approach, but it felt a little too involved for this grade level, so I’m still working out the details - and no, I’m probably not using the elementary-school traffic light system it initially suggested!

Why This Feels Different

What makes these LLM debriefs so powerful is how they combine my specific, personal expertise with an incredibly broad perspective. When we discussed the skills board idea, for instance, the LLM drew upon everything from peer tutoring programs to skill-sharing communities to professional mentorship approaches. It’s not just about spotting patterns - it’s about having a collaborator that can pull relevant ideas from across different domains and contexts, helping you see your work in new ways.

I’m starting to see all kinds of possibilities for these debriefs. After an important meeting, you could use one to help process what happened and capture key insights. Following a workshop, you might explore what worked and what could be better next time. Had a particularly tough week? A debrief might help you step back and see the situation more clearly.

The key is giving the LLM enough context to be a useful thought partner. And remember - this isn’t about letting the AI drive the conversation. Push back when things don’t feel right. Ask those follow-up questions. Let your expertise guide the exploration. The LLM won’t mind - in fact, that’s how you get to the good stuff.

Looking Forward

This feels like one of those moments I’ll look back on as a turning point. You know those “before and after” experiences? I think this might be one. The ability to have this kind of thought partnership - to extend our thinking in new ways while still remaining firmly in control of our own professional judgment - that’s pretty amazing.

I invite you to try an LLM debrief in your own work. It doesn’t have to be anything formal - just an experiment in thinking differently about something you care about. Because when you combine your expertise and experience with an AI’s ability to see fresh patterns and possibilities, you might just surprise yourself with where you end up.