AI Reset the Board.
Stop Selling Me Who You Were.

The event in my feed

This week it was an "AI build week." Before that, a summit. Before that, a masterclass. Every few weeks another arm of the ecosystem runs an event to promote AI, fills a room, brings on speakers, and posts the numbers afterward.

I'm not against the initiative. A country should push. But I've stopped spending my own time on these, and I'll be honest about why. Too often the body running the event has never deployed the thing it's promoting. It invites people to talk about AI who are, at best, one book ahead of the audience. The number that matters at the end isn't how many businesses can now do something they couldn't before — it's attendance. That's a KPI, not an outcome.

We have run this movie before

Here is why I don't think a short course fixes the real problem.

When short video and livestream hit, the senior marketers — the ones with the awards and the decade of brand experience — mostly didn't get it. Worse, they didn't want to get it. It looked cheap, it looked beneath them, it wasn't how things were done. So they waited. And while they waited, a set of brands that barely existed a few years earlier learned the new format from zero and took the lead. Look at the category leaders today. A striking number of them are names that came up through the shift the incumbents refused to make.

That wasn't a talent gap. The incumbents were more talented, better funded, more experienced. It was a reset they didn't accept in time.

AI is that reset — bigger

Now ask the honest question about AI. Do the top management teams reading a few headlines and sitting through a two-day class actually understand what they're facing? I don't think most do, and I don't think a class can deliver it — because the people best positioned to warn them are the ones getting it wrong at full scale.

Ford cut engineers on the assumption AI could carry the load, then rehired experienced humans to fix quality problems the automation couldn't. Their own hardware VP put it plainly: they thought they could implement AI, feed in the requirements, and high-quality products would fall out the other end — and they were wrong. IBM automated most of its HR requests, then hit the fraction that needed human judgment and moved to expand entry-level hiring again. Forrester found more than half of employers who cut jobs for AI reasons came to regret it.

What the reset actually does

Here's the part that's hard to say out loud at a networking event: AI shrinks the value of experience and accumulated knowledge. Not to zero — but far enough that the gap you spent twenty years building narrows to something much smaller, fast. A model can now hold more domain knowledge than most specialists and apply it on demand. The moat that used to be "I know things you don't" is draining.

Which means, in an odd way, the board has been reset and most of the pieces are back on the starting line together. The advantage no longer sits with whoever has the longest résumé. It sits with whoever adapts fastest — who is willing to be a beginner again, in public, at forty, the way those young brands were beginners at short video.

That is genuinely good news for anyone hungry, and genuinely threatening to anyone whose whole position rests on seniority.

So what should the initiative be?

I'd rather see less promotion and more of the unglamorous thing: help a business actually implement one workflow end to end, with someone who has done it, and stay until it works. One real deployment teaches more than ten summits. If the ecosystem measured events by "how many businesses shipped something that still runs a month later," the whole format would have to change — and it should.

Until then, my advice to any owner sitting in one of these rooms is simple. Don't buy the reassurance. Assume you know less than you think, because everyone does right now, including the people on stage. The reset doesn't care who you were.

Stop selling me who you were in the old world. Show me you can move in this one.

Now or Never.

Published by IMA AI — July 2026.