Why AI Adoption Stalls Even After the Demo Impresses
18 set 2026
Four Claude meetups since July, and the rooms keep getting fuller.
Four Claude meetups since July, and the rooms keep getting fuller.
Everyone comes to see what Claude can do. On Wednesday at Adnovum in Zürich, we heard more stories.
The demo worked. Nobody used it.
The board approved. Then 3 months later, the same person was still copying the report by hand.
Dr. Andreas K. Janoschek described that moment at the Claude for Financial Services meetup in Zürich this week.
Hands up if it never happened in your own company...I see this all the time (it was often the case before AI, with any other tool). It is the part that gets forgotten. A lot of energy goes into building the tool, and much less into the person who has to use it every day.
→ Can you name the person who will run this workflow in 12 months?
Not the person who builds it. The person who uses it every morning.
In a study of Swiss independent wealth managers he quoted, all of them use AI. Only 11% run something at scale.
Dean Marti from SIX told the same story with a factory.
A hundred years ago, one big steam engine sat in the basement. Long belts carried its power up to every machine. So the machines had to stand wherever the belts could reach.
When electricity came, electric motors replaced the steam engine. But they kept the belts. Almost nothing changed.
Then Henry Ford put a small motor into each machine and arranged the floor in the order the work flows. That is when productivity jumped. When they actually rethought the process, not only fixed the holes.
Stefania Stan from UBS added a point on money. Running AI after launch usually costs more than people plan for. So true!
Thank you and congrats to the great event Adnovum Karl Richard V. Melinda Ferreira Zoltán Mészáros and everyone contributing to the event, it was great to be with you!
Who runs yours in 12 months?
What this means for leaders rolling out AI in their own teams
The pattern from that Adnovum room matches what I see inside the companies I advise. Everyone can point to the tool. Almost nobody can point to the person who owns making it part of daily work, six months after launch.
In my Private AI Advisory work, the first question I ask a leader is never which AI tool to pick. It is who on the team will still be running this in a year, and what happens to their job description when they do. Most leaders have not answered that question before they greenlight the build. That is exactly what senior leaders actually need from AI in 2026, and it is rarely the tool itself.
Three patterns I see repeatedly:
→ The tool gets a champion for the launch, not for the maintenance. Once the person who pushed the rollout moves on, so does the habit.
→ Nobody redesigns the workflow around the tool. Like the belts in Dean Marti’s factory story, the process stays the same shape it always was, just with a new engine bolted on.
→ The cost of running AI after launch is treated as a rounding error, not a budget line. Stefania Stan’s point about post-launch cost applies here too: adoption needs ongoing attention, not just a launch budget.
The 11% figure from the wealth management study is the part I keep coming back to. Everyone in that survey uses AI. Only one in nine run it at any real scale. That gap is not a technology problem. It is a leadership and ownership problem, and it is exactly what I work through with leaders one to one.
If this resonates and you want a practical guide, I have a free AI Tools & Tips Guide that walks through the tools I actually use as a coach and leader, with the workflows behind them, so you are picking tools with a plan for who runs them, not just what they do. Leave your email here and I will send it over: https://www.thechangerepublic.com/free-resources