This week we were in Gothenburg for D-Congress, Scandinavia's largest gathering for those working in e-commerce and those delivering the software and services needed to make it work, efficiently, expansively and engagingly. A broad, eclectic crowd, which makes these days far more interesting than when we met in a similar constellation 15 years ago. Music, film and Olympic gold on the programme. Festival with good bands instead of a long sit-down dinner. Slick, loud and good in every way.
So, what did we talk about this year? Hard to say, really, but AI and agents were everywhere. No company and no speaker can resist being AI-ready, AI-powered or Home for agents.
Everyone talks about AI and agents. Many are genuinely astonished by what's possible. But collective understanding is thin, the shared mental model is vague and nobody is really stepping up to lead the way.
Conversations jump from question to question: is Claude really better than ChatGPT? (read more here). Can you put a button in a PIM that instantly creates better product copy and a new image? What happens to all the SaaS tools now that the cost of building custom solutions, including UI and integrations, is approaching zero? And more philosophically: what do we do now that the world is splitting between those who are curious and forward-leaning and those who become frightened and resistant?
The train is leaving. Not everyone has a ticket. And not everyone even knows it's time to head to the platform. What happens to inequality and democracy when value creation concentrates in the hands of those who manage to learn a new language that changes faster than education systems and organisations can keep up with?
That last question stuck with me. The one about the platform.
Previous technology shifts had a natural lag. It took time to roll out broadband, time to teach organisations to use ERP systems, time for new roles to take shape. That lag was breathing room. Societies adapted, education systems adjusted, those who fell behind caught up.
That breathing room looks smaller now. The frontier labs release new models every other week. Each release compresses the cost of things that used to require specialist expertise: code, analysis, legal, strategy. Someone who understands how to work with these systems today will, in six months, have an advantage measured not in knowledge but in embedded workflows, trained intuitions and accumulated context. It is hard to catch up with something that is accelerating.
This applies to individuals. To companies. To countries.
The most uncomfortable part is not the inequality itself. That has existed before. What's uncomfortable is that value creation is starting to concentrate in a layer that is invisible to most people. Not in factories or offices or even in code, but in the understanding of how to steer systems that in turn steer everything else. Those who cannot see the layer do not even know what they are missing.
And your mental model does not change by reading about AI. It changes by using it.
Our advice: get your own experience, fast. It is the only thing that actually changes how you think. Seminars and reports give you words for what you already believe you think. Sitting with a problem you actually care about and solving it with AI gives you a new reference point.
The difference between those who understand what AI does and those who do not is rarely about intelligence or interest. It is about how many hours you have spent actually using the tools on real problems.
That is exactly what our workshops are built for. Hands-on, with real problems from your own working life. We meet you where you are, whether you are new to AI or already working agentically and want to understand what comes next. No slides.
And look: prompting in a project feels powerful today. But it looks primitive once you have seen the next season. Or the one after that.