My honest take on AI in mid-market business.

1. Most AI projects fail on execution, not technology.

I see it all the time. A consultant's report. A roadmap. Sometimes a prototype. Then reality bites: nobody actually builds the thing. The consultant gives advice and walks away. That's not a weakness in these businesses. That's the gap I fill.

2. Using AI tools just speeds up mediocrity.

Copy that says nothing. Analysis nobody questions. Decks that could belong to anyone. AI isn't the problem. The problem is most people never learned how to direct it, so the output never rises above the question.

3. AI workshops without working results are money down the drain.

The usual routine: two days of inspiration, plenty of applause, then back to business as usual with a handbook gathering dust. My workshops end with something that runs the Monday after. Not just good vibes.

4. GDPR doesn't kill AI. It's a quality benchmark.

Companies that write off US cloud tools wholesale and then dodge AI for years haven't looked properly. Powerful, compliant systems exist: self-hosted, on German servers, with proper data processing agreements in place. Hiding behind GDPR usually means you weren't going to start anyway.

5. AI won't rescue mid-market companies. It amplifies whoever's already delivering.

Efficiency magnifies what's already there. Good processes get faster. Bad processes produce more bad work, faster. AI isn't a turnaround tool. That's why every audit I run starts with one question: what's already working? Then we look at what should speed up.