Last night, at my entrepreneurs’ club. Entrepreneurs Organization. Coffee, a quick catch-up, and within ten minutes the talk was about AI. As it almost always is these days.
Next to me a manufacturer. “Tim,” he said, “my nineteen-year-old son built something over a weekend that we got a thirty-grand quote for last year.” He laughed a little uneasily.
Further along, someone with a webshop. Showed me on his phone how he gets AI to write his product copy on the morning train. “Saves me half a day. Every day.”
And there was one who said it was all a bit overblown. One. The rest feel exactly the same. A wave is coming. And just as many opportunities.
At our company that wave is no longer a forecast. It is here. And it went fast: in three months.
What shifted
We had three developers with their hands full. Today it looks different. A lot of development is moving to our consultants. Sounds odd. It is not.
Because they sit at the table with the client. They understand the problem best. And with AI they now turn that problem into software themselves. The detour via a separate developer gets shorter. Sometimes it is gone.
Bugfixing shifts too. To the first line of our support. That works, because our first line is also our second and sometimes third line. We do not work with juniors. Whoever picks up the phone here knows what they are talking about.
Three months. That is all it took.
How we do it
We let Claude write code and review itself. On security too. Then a developer looks it over. And honestly? So far little bad comes out.
The craft shifts. Less typing yourself. More thinking clearly about what it has to do. And then checking it is right. That is not a lesser job. It is a different one.
The best part is the people. Our developers want to move with it. Not out of fear, but because they can see the fun in it. And our consultants are unanimously enthusiastic - they are all doing AI engineering. That is not a given. I am proud of it.
Do we pass this on to the client?
Then the real question lands on the table. Do we keep the gains for ourselves? Or pass them on?
For me it is not a long discussion. It cannot be otherwise. If we do not, someone else soon will. Cheaper, faster, just as good. And then you end up at the back. Not because your product is worse. But because you held on to your old business model too long.
Not passing it on feels safe. It is the opposite.
Do I have no worries, then?
I do. I would love to see good European alternatives emerge. For Claude Code. For Office. For our infrastructure. I already see promising initiatives, and that gives me hope.
Until they are here, we keep it tidy: GDPR-compliant Enterprise contracts at Claude and Amazon Bedrock. Well arranged. But I keep watching, because dependence is a risk too.
What is not an option is standing still. The wave is coming either way. You let it work for you, or you wait until it rolls over you.
We have chosen.
Want to know what that shift means for your business? Read why custom is affordable for mid-market now, or how it plays out on the floor in AI in manufacturing.