What I Got Wrong Running My First AI Workshop

So I wrote about running my first scenario workshop at Cowboy back in December. I was excited, noted a few things I'd do better. But now that I've had time to sit with the feedback, I think the real lessons are different from what I thought.

The thing I keep coming back to is that I talked too much. Each exercise was supposed to be a quick setup and then let people go. Instead I over-explained, added context, made sure everyone understood the framework. Classic. By the time we got through three exercises, we'd run out of time for the fourth one — which was the one where you actually decide what to *do* about the scenarios you just built.

So we did this whole journey from horizon scanning to tensions to scenario planning... and then stopped right before the action part. The irony of that is not lost on me.

But here's what surprised me in the feedback: people didn't care that much about the exercises themselves. What they valued was the space. One person said the most precious moment was just "the first board, the exchange of everyone's questions about the future — it's interesting to know what each person thinks." Several mentioned the team coming together. Someone who doesn't like unstructured creative sessions admitted they held on and actually enjoyed it.

I designed an AI workshop. They experienced a team moment that happened to be about AI. That's a meaningful difference.

The other thing that came through clearly: they wanted it to be more practical. "We should see the basics first." "Something where you visualize an aspect of your job and find a way to use AI to make life easier." The philosophical part was appreciated, but it wasn't enough on its own.

One colleague raised the tension between Cowboy's green vocation and AI running on fossil fuels. His suggestion was to use AI for prototyping, then once the process is mature, convert agents into simple scripts that are less resource-hungry. That's more strategic than anything I had prepared. The best insights came from the participants, not from me.

Next time: less talking, more doing. And exercise four happens no matter what.