Back to Insights
AIInnovation SprintsPrototyping

The Skill Isn't Using AI — It's Reading the Moment

Published June 14, 2026·4 min read

Three weeks. Two continents. Two clients. A reminder I needed.

Over the past few weeks I ran innovation sprints with two very different organizations — one in aviation, one in MedTech. Same toolkit. Same facilitator. Two completely different rhythms — and by the end, the contrast had taught me something I keep having to relearn.

The airline: build fast, learn faster

At the airline, we went straight to the airport. We watched how people actually work, and we talked to the teams on the ground. Within hours, I was coding working prototypes based on what we'd learned — and handing them right back to the people who would use them. Not to impress anyone. To learn faster.

Where will this screen live? How do you hold the device? What matters in the 30 seconds you actually have? Each answer made the next version better. By day five, we had a functional MVP. Not a concept. Not a slide deck. A working solution — built with AI, shaped by the people who'll actually use it.

Five days at the airport 1Observewatch real work 2Prototypebuilt the same day 3Hand backto real users 4Iteratesharpen each pass 5MVPworking, not a deck

The MedTech sprint: the discipline to slow down

At the MedTech company, the energy was completely different. We had a motivated team, great people, and AI tools ready to go. By day two, everyone wanted to start building. The pull was real.

But we slowed down. The challenge was more complex, and we needed to understand it properly first. So we kept talking to people. We ran interviews. Sometimes we used AI in parallel — to research and process what we were hearing in real time.

And then it happened. One person shared an insight that had never been written down anywhere. Not in any report. Not in any database. Not in any AI. The room went quiet. Nobody expected it. And it gave us exactly what we needed — the confidence that our direction was right, clarity on where to focus, and a wave of new opportunities we wouldn't have seen without that conversation.

The insight no system had recorded Not in any report Not in any database Not in any AI It came from one conversationAn insight no one had ever written down.

Same team, same AI tools, same energy. What's the move?

You can't read it off the surface — only off the signals. Flip them and watch the right call change.

Stay in the room. Murky problem, and the crucial thing still unsaid. Build now and you'd hard-code a misunderstanding. This was the MedTech sprint — they waited, and the insight that turned everything sat in no report, no database, no AI.
Build — but keep listening. The problem's clear, yet something's still unspoken. Prototype to flush it out: let a rough build provoke the insight the conversation hasn't surfaced.
Frame it first. You've got the key insight, but the problem's still fuzzy. Sharpen what you're actually solving before you spend a day building the wrong thing well.
Accelerate. Clear problem, insight in hand, users reachable today. Build it now and test it the same day — this was the airport. Speed is right because you read the moment.

Reading the moment

Here is what I took away from these three weeks. It is not about choosing between AI speed and human depth. It is about reading the moment. Sometimes the right move is to build fast, test fast, and iterate the same day. Sometimes it is to slow down, stay in the conversation, and trust that the insight you need hasn't been written down yet.

Reading the moment Acceleratebuild fast, test the same day› The problem is clear enough to test› You can put it in front of users today› Each version teaches the next Stay in the roomslow down, keep talking› The challenge is genuinely complex› The answer may not be written down yet› The team wants to build too soon Same tools. The judgment is knowing which moment you're in.

The skill isn't using AI. Everyone can do that now.

The skill is knowing when to accelerate — and when to stay in the room a little longer.