The Companies That Win the AI Era Won’t Be the Ones Just Using AI

I’ve watched this same wave roll in three times in 20 years. We’re about to watch it roll in a fourth — bigger, faster, and with most companies still standing on the shore with their backs to the ocean.

First wave: the web.

Businesses discovered the internet and rushed to “get online.” Most did exactly what you would expect: they took the brochure off the coffee table and turned it into a homepage. The catalog became a product page. The phone number got an email address next to it. Digital? Technically. Transformative? Not really.

Second wave: the smartphone.

Every customer suddenly carried a supercomputer in their pocket, and companies responded with “digital transformation” budgets running into the billions. Here is the dirty secret of that era: most companies did not transform anything. They digitized old processes.

The expense report became an app. The bank teller became a mobile deposit. The travel agent became a booking funnel. Faster, cheaper, more convenient? Yes. Reimagined? Almost never.

The winners of that wave understood something different. They crafted experiences. Airbnb did more than digitize hotel bookings; it built a marketplace for living somewhere else. Spotify went beyond digitizing the record store; it rebuilt the relationship between a person and music.

They understood the smartphone was not a smaller computer. It was a new context for human behavior.

But even the winners hit a ceiling. Crafted experience is not the same as personal relationship. Spotify knows what I listen to; it does not know me. Airbnb knows where I have stayed; it does not know me. The systems were beautifully designed for a customer base of millions, not for me specifically.

Third wave: the early AI outliers.

A handful of companies saw the next wave forming behind the one everyone else was riding.

Lemonade is the example I keep coming back to. While much of the insurance industry was busy digitizing claim forms, Lemonade tried to rebuild the relationship itself: AI underwriting, AI claims, a conversational interface, a giveback model.

Lemonade is not the point because it got everything right. It matters because it asked the right kind of question earlier than most of its industry.

It did not simply ask: how do we make insurance more efficient?

It asked: what should insurance feel like if we were building it now?

That should have been the wake-up call. Mostly, it wasn’t.

Fourth wave: now.

AI is no longer scarce. It is no longer a moat. It is a utility, available to every company with a credit card and a pulse.

Customer service will get a copilot. Marketing will generate more copy, faster. Finance will close the books in three days instead of five. Boards will applaud. Consultants will produce slides. Headcount lines will shrink. Everyone will declare victory.

And it will be a disaster.

Not because efficiency is bad. It isn’t. But because the companies sitting next to you are not playing the efficiency game. They are asking a different question:

If we were starting this company today, knowing what AI can do, what would we even be?

That’s the question Lemonade asked in 2015. It is the question almost nobody is asking in 2026. Including, I suspect, Lemonade itself.

Because the cruel thing about being the disruptor of the last wave is that you become the incumbent of the next one. The AI Lemonade built its business on is, by today’s standards, primitive. The interaction model is primitive. The assumption that a customer wants to “talk to an insurance app” is primitive.

Lemonade’s risk now is the same risk Blockbuster had in 2005: believing that being ahead last time means being ahead this time. The wave you rode in on is the wave that drowns you, if you stay on it too long.

Here is the part most strategy decks miss. Speed of adoption is not the lever. Speed of realignment is.

Stanley McChrystal learned this the hard way in Iraq. The most sophisticated military on earth was being outmaneuvered by a decentralized insurgency, not because the enemy had better technology, but because it had a flatter structure and a faster decision loop.

His answer was not to add more process. It was to dismantle the chain of command he had inherited and push decisions to the edge. He created the “Team of Teams.” It worked.

Most companies today are in exactly that position, and they do not know it.

You cannot run a company at AI speed through a structure built for quarterly planning. You will lose. Not slowly. Quickly — and to a company a tenth your size.

The companies that win the next ten years will share a few traits, and almost none have to do with which AI foundation model they license.

I don’t mean a bank app that answers faster, but a financial system that notices risk before I do. I don’t mean a health portal with a chatbot, but a care layer that understands my history, constraints, and next best action. I don’t mean an insurance app that processes claims faster, but a protection layer that sees around corners.

That is the difference.

The companies that get this will look unrecognizable in five years — not because of the technology they use, but because of what they decided to become.

The companies that don’t will have very efficient versions of the wrong business.

The lesson of the last twenty years is the same lesson, told three times: the technology is never the transformation. The transformation is what you rebuild around it.

We are about to be told again.

Most companies will listen just long enough to buy the tools.

The ones who rebuild will own the decade.

---