It happened again. A new model was announced, and my LinkedIn feed is overflowing with superlatives. "Revolutionary." "Paradigm shift." "Everything will change." I scroll on.

Not because I find AI uninteresting – quite the opposite. I've been working intensively with it for years, I advise companies on it, I implement solutions. But that's precisely why I have an allergic reaction to the hype. Because I know what comes next.

The Hype Cycle Only Goes One Way – Down

Gartner calls it the "Peak of Inflated Expectations". I call it simply: the pattern. A new AI tool appears, everyone is excited, companies invest hastily, and six months later a managing director tells me why "AI didn't work for us".

Except AI did work. Just not the way it was promised. And not for the use case it was deployed for without proper thought.

"The tool isn't to blame if you try to drive a screw with a hammer."

What I Do Instead

I start with questions. Not answers. What problem needs solving? What data exists? Who will actually use the system – and how? Is AI even the right tool here, or is there a simpler solution?

Sometimes the honest answer is: no, AI isn't worth it here. That may cost me a contract. But it strengthens my credibility – and the credibility of my clients with their own teams.

40 Years in IT Taught Me One Thing

Technology comes and goes. The problems it's meant to solve remain. I've watched client-server architectures rise and fall. The internet. Cloud. Mobile. IoT. Every wave brought genuine innovation – and genuine excess.

AI is different from all previous waves. It is more powerful, more versatile, and – yes – also more disruptive. But that makes sobriety more important, not less.

Those who assess today, with a cool head, what AI can and cannot do – who set realistic expectations while still being bold enough to pursue real implementations – those are the people who will still be here in three years. The hype surfers will be riding the next wave by then.

What This Means for Your Organisation

You don't need an AI evangelist. You need someone who tells you what's possible, what it costs – and what you're better off leaving alone. Someone who knows the difference between a fine-tuned model with genuine value and an expensive ChatGPT wrapper.

That's not modesty. That's professionalism.

— Michael Jung, Mainz, February 2026