Sharpening how a municipality works with AI

This summer we're stepping into Gemeente Arnhem to run a training programme on working effectively with AI, not as a novelty, but as a practical tool for the daily work of municipal staff. The heart of it is a simple, memorable framework we developed specifically for this context: KORT (Kader, Opdracht, Resultaat, Toon). The idea is to give civil servants a structure they can actually hold in their heads, so that a good prompt becomes second nature rather than a guessing game. Public-sector work has its own demands, clarity, tone, accountability, and the training is built around those realities rather than generic "prompt engineering" tips lifted from somewhere else.

Before bringing it to Arnhem, we wanted to be sure it held up in practice, so we ran a pilot at the studio in Leiden with a small, hand-picked group. That session was less about delivering the material and more about pressure-testing it: watching where people hesitated, where the framework clicked, and where our examples landed flat. The feedback was genuinely useful. We trimmed the parts that overexplained, leaned harder into hands-on exercises, and reworked a few examples to feel closer to the actual letters, notices, and advisory documents that municipal staff deal with every day. What's going to Arnhem is a noticeably tighter, more confident version of what we started with, and that's exactly what the pilot was for.