Last week in Berlin, at the MY WAY conference, EY gave a keynote on the exponential development of AI and what it means for the German Mittelstand. Most of it applies to Swiss family businesses too.

The speaker described a familiar pattern in paradigm shifts. First comes "oh, wow." Then most companies decide they have ten or fifteen years before it reaches them. It reaches them much sooner. By the time they react, it has become "oh, shit," and there are new market leaders.

His core argument was that how far ahead you can see decides how ambitious you can be.

It made me think about when our own projects usually start.

Most renewal work starts under pressure. In our projects, the call usually comes when the core is already declining. Margins are down, a large customer is wobbling, the order book is thinner than last year. That is when the need becomes visible, and that is when budget gets released.

It is also when the options start to disappear.

The window is getting shorter

In April 2025, Microsoft's CEO Satya Nadella said that 20 to 30 percent of the code in Microsoft's repositories was already written by AI. The number is striking because it shows how quickly AI has moved from an emerging technology into the production system of a major technology company.

The keynote's claim was that what happened in software engineering over the last year will reach almost every function within the next two.

I am less convinced the curve stays exponential. Today's models have already been trained on most of the public text that exists. Epoch AI, a research group that tracks AI development, estimates that the stock of human-written text will be used up somewhere between 2026 and 2032. The road to superintelligence may be longer than the keynote suggested.

For a family business, that changes less than it sounds. The tools that exist today are already enough to change how quoting, sales and service work. You do not need an exponential curve for your business model to come under pressure.

Why companies wait, and why it is rational

Waiting is not simply a failure of courage. The core business earns measurable returns today. The next curve needs capital today for returns that may not come for years.

When the core is healthy, renewal looks optional. When the core weakens, renewal becomes urgent. But by then it competes with falling cash flow, cost pressure and fewer strategic options.

That is the paradox. Renewal is easiest to fund when it feels least urgent, and hardest to fund once everyone agrees it is urgent.

The cost of intervention rises as options fall

The keynote's co-speaker, a restructuring expert with fifteen years as chief restructuring officer in Mittelstand companies, gave this a simple shape. In a strategic crisis, operations and financing are still fine, and you need a fitness coach. In an earnings crisis, you need a doctor. In a liquidity crisis, definitely the doctor. Once you have lost control, only the fire brigade helps.

The economics behind the metaphor are simple. The earlier you intervene, the more options you have. The later you intervene, the more each option costs.

His hardest observation: at the early stage, the sense of urgency does not exist yet. Maybe in the leadership team, rarely in the company. By the last stage, everyone understands, and there is little left to choose from.

What starting early means:

→ Name your stage honestly.

Coach, doctor or fire brigade?

→ Put a date on it.

What does the core look like in 2032 if nothing changes?

→ Reserve capital before you need it.

Renewal should not have to compete with crisis spending.

→ Test while the core still pays.

A first experiment is cheapest when failure is affordable.

The keynote ended on agency, not alarm. "You have 80 to 85 percent of what it takes," the speaker told the room. He was talking about robotics. Engineering and manufacturing at scale are already in the Mittelstand's DNA. The same is true for Switzerland. What is missing is mostly software, and the people who build it. I would add one more missing piece: the decision to start before it hurts.

Bottom line: Pressure tells you when to act. By then, some of the best options are already gone.