Customers have moved on. Most products haven’t.

At Framna, we work on this question with our partners every day, and inside our own company too.

In September I took it to the main stage at Future Product Days in Copenhagen. The line behind me read: “You don’t just compete with products in your category. You compete with the smartest experience your user had today.”

Expectations are no longer set only by your competitors. They are set by whatever your customers used last, and increasingly that is an AI assistant that understands a question asked in their own words. Then they open an app that only offers the choices it was built with.

So at Future Product Days we put the question to the industry in a survey: how far have companies and their product teams come with AI, and how are they moving through what I like to call the shift? It is the most significant paradigm shift digital products have seen in decades.

This article is the first in a series of films, articles and masterclasses over the coming months, because we would rather work this out with the companies facing it than on our own.

Between sessions I talked with designers, engineers and product leaders, and I sat down with Don Norman, a legend in our industry. He wrote The Design of Everyday Things and is widely credited with coining the term “user experience” while at Apple in the 1990s.

We asked 530 people who design, build and lead digital products, at Future Product Days and across Framna’s network, in September 2026. Most are designers and product leaders. Two in three work in companies with more than 250 employees, and 45% in companies with more than 1,000. The answers are self-reported, so read them as the view from inside product teams.

You can read the full analysis here or see the survey results here. What follows is my reading of it, as CEO of Framna, where more than 500 of us design, build and run digital products for brands that millions of people use every day.

55% of the people we asked say AI has raised what their users expect, including 21% who say it has raised it dramatically. Only 42% say their product has adapted, including 14% who say completely. Almost one in four say it has hardly adapted, or not at all.

I see the gap every time I open my banking app. My daughter, my father and I all bank with the same bank. She is 20, I am 52 and he is 76, and our lives could hardly be more different. Yet the app shows all three of us the same screens. The bank knows each of our histories in detail, and its app behaves as if we had never met.

Some people have stopped waiting. A tool called BankMCP already, by its own description, connects bank accounts to an AI assistant. If the assistant becomes the place where people manage their money, the bank keeps the accounts and loses the relationship.

The bank still holds what the newcomers lack: years of history. As Don Norman put it when we spoke, the bank may know all your activity, but the teller doesn’t. An AI teller will know your history and can use it when it talks to you. That is what I mean by an intelligent product: one that learns about the individual user and keeps getting better every time they use it.

My bank is not the exception. We see the same gap in many of the categories we work in, from energy and travel to retail and healthcare. Years of customer history are only an advantage if companies put them to work before someone else does.

Companies feel the pressure. 71% say AI is making them rethink their product strategy, 37% from the ground up. For 81%, that rethink is under way right now. Only 13% say it is largely done.

Today, the most common AI capability is conversational input, live for 43%. The plans point straight at the gap in my banking app: personalized experiences top the list of what teams will add in the next twelve months, at 45%, followed by insights and guidance, and AI that carries out tasks for the user. Adaptive interfaces are live for only 17%, while 35% plan to add them.

The money is following, if cautiously. 29% say their product budget is growing, against 7% who say it is shrinking, and another 26% have a flat budget and are expected to do more with it. And 32% say it is too early to tell where the budget is heading, or do not know.

Building is where teams have moved fastest. Nine in ten say at least some of their code is now written with AI, and for one in five, agents write it first. 64% say they ship more than before.

But only 23% say the way their team works has changed completely, and so far the gains show up mainly as more output. 43% say they have become faster than last year, and 30% see better quality. Most of the rest see no difference yet, and one in five cannot say whether quality has changed at all.

Getting something new in front of users is another matter. For just over half, it still takes a month or more. Only 6% manage it in days, and 21% say it varies, which to me usually means it is not being measured.

Asked what slows them down, teams pointed to familiar places. Risk and compliance came top, with legacy systems close behind. Data readiness and the speed of decision-making followed. Skills and capacity came lower, and culture came last. That surprised me. I suspect culture hides inside the other answers: how fast a company decides is culture too.

None of these brakes is new, and no tool removes them on its own. Risk and compliance, like the speed of decision-making, is about how an organization says yes.

Put the brakes next to the roadmap and the pattern is hard to miss. The most common AI today is a chat window, which can sit on top of a product without touching customer data, old systems or a risk sign-off. What teams plan next, personalized experiences and products that adapt to the individual, needs all three.

Don Norman has watched the same thing happen with one technology after another, and he does not blame the tools.

Customers will not give anyone ten years. Digital products spent the last decade improving in small, steady steps that most companies could keep up with. AI has changed the pace, and expectations now move faster than most organizations.

The tools matter. AI can scale business impact, but only when the organization around it is set up to use it. Most companies have the same models. What sets them apart is how fast they turn those models into products their users choose. That takes three disciplines that depend on each other and run as a loop.

Capture finds where to win. Build turns that direction into a live product. Operate keeps the product evolving, and what it shows feeds the next round of Capture. With every round through the loop, you know more about your users and your product, so the product gets better each time. A faster build is wasted on the wrong idea, and the right idea goes nowhere if every release waits for a sign-off. This is where I would start.

Every company has to go through the shift, and so do we. It is complex, and it is both scary and exciting. We have come a long way, and we are still in the middle of it.

We are taking our own medicine across all three disciplines. We decide where to put our effort on data rather than gut feel. Framna Build, our own software factory, is becoming how we build every product. And the way our teams work and decide is changing with it, as roles evolve and AI fluency becomes the baseline in every craft.

That is also changing what we are to our partners. We still build great products, but increasingly we help companies through the shift itself. It is only with all three disciplines in place, working as one system, that AI turns into real business success in the age of intelligence.

The next time my daughter, my father and I open our banking app, I hope it knows which of us it is talking to. The companies that cut their decision time will get there first, and set the bar for everyone else.

We share what we have learned from our own shift and from the companies we work with: where intelligent products create value, and what it takes to change how products are built and how decisions are made.

In a session with your leadership team, we present our point of view and go through what the findings mean for your organization.