Where does an AI transformation start?

An AI transformation starts with one question: how much time do you have before the AI natives disrupt you? AI natives are companies built on artificial intelligence (AI) from day one.

I learned this approach from Dominik Stein, Head of Growth at the investment firm EQT, when he was my guest on the DEALMACHER podcast. EQT's portfolio spans AI natives such as Lovable and Harvey, growth companies, and classic buyouts far away from AI. So EQT needs an answer that works for all three. I summarised it in my CEO Briefing on how EQT moves its portfolio toward AI.

One question comes up again and again in EQT's portfolio discussions: “This whole new AI world is all great – but I have 50 developers and a 15-person product team. What exactly do I do tomorrow?”

Dominik's first answer is the time question. Whether a company calls itself AI-first comes later. The time you have drives everything that follows: how fast you move, how radical you get and how much risk you take.

How much time does a company have before AI natives disrupt it?

The time a company has depends on its market, and the answer differs widely. Dominik gives two examples:

• Website-building software: a company that sells traditional website-building software and now competes with Lovable has twelve to eighteen months, in his view. After that, customers build their sites by prompting instead.
• Software for school photographers: an EQT portfolio company from Berlin sells vertical software for school photographers. It has more time, because few AI investments chase that niche.

More time still means moving aggressively. The company with more time only sequences its steps differently.

Dominik also draws a sharp line by founding year. To him, every company founded before 2023 is almost a legacy company. It was built before ChatGPT and still runs on processes from before AI. A company founded in 2024 gets the new technology stack for free. Everyone else has to rebuild while running the business.

How do you get teams to actually work with AI?

You put people next to your teams who have already done the rebuild. EQT answers the Monday-morning question with people.

It sends in trainers. These are operators who have done the rebuild themselves in the last twelve months, often in exactly this kind of software company founded before 2023. The process runs in waves:

1. First teams: the trainers take the first two or three teams and convert them to the new way of working.
2. New team shape: small units with one developer and one product person, plus sales people who bring in customer input. They ship in two-week sprints.
3. Next teams: the converted teams then help convert the next ones.

The principle underneath: a few people teach themselves everything from what they read. Most teams need someone standing next to them who has already made the mistakes.

What is Reverse Public Shaming?

Reverse Public Shaming is Dominik's name for peer pressure that pushes companies toward AI. EQT runs forums across its portfolio. In these forums, chief financial officers (CFOs) meet other CFOs, and chief technology officers (CTOs) meet other CTOs.

In each round, EQT picks the two or three companies furthest along on AI. They present to their peers: this is what I built, this is what it did. Two reactions follow in the room:

1. First reaction: why am I not there yet?
2. Second reaction: if they can do it, so can I.

Then people call each other, exchange notes, and small networks form across the portfolio.

Why does peer pressure move people more than pressure from the top?

Peer pressure moves people because the example is close to their own situation. The company that presents is a peer, for example a company from the Black Forest region that had nothing two years ago. That closeness turns the first reaction into the second.

I put it in two short lines in my CEO Briefing on Reverse Public Shaming: “Pressure from above gets nodding. Pressure from peers gets phone calls.”

The same mechanism helps AI-native founders who sell to large companies. One referenceable peer in the buyer's world beats ten case studies from somewhere else.

Does an AI transformation work for every company?

An AI transformation does not work equally for every company, and Dominik is honest about that. Some EQT portfolio companies have lifted productivity so far that 30 to 40 percent of capacity was freed up or redeployed. Others accelerated revenue sharply through new products. And others struggle.

The outcome depends on the industry. It depends even more on the people and on how fast they truly adopt. One thing has changed everywhere in Dominik's view: AI is topic number one in every board meeting.

How can a CEO without EQT's resources lead an AI transformation?

A CEO can copy the structure. The structure is the part that transfers, and it does not need EQT's resources:

1. Ask the time question honestly: how much time do you have before AI natives disrupt you? The answer sets your speed and your risk.
2. Put someone next to your teams: bring in a person who has done the rebuild, instead of circulating articles. EQT starts with the first two or three teams.
3. Create rooms for peers: let the people who are furthest along present to their peers what they built and what it did.

The results will be uneven. The structure gives your teams a fair chance to be among the ones that move.