Last week, Tim and I attended DMEA 2026 in Berlin. We spoke with many people, and kept having similar conversations:

"We're building a TI messenger that's better than the others." "We have the most accurate voice-to-text solution for clinics." "Our tool now also meets standard X."

All important “building blocks”. All valid.

What's stayed with us since: in many of these conversations, the actual building wasn't really the topic. Few people spoke with us about the vision for the patient experience. Or about the healthcare system that should ultimately emerge. Or about what we want to build for people in the end.

We kept coming back to the same image: as if we were all playing building blocks together. Each person holding up their piece. The green 4-block, the red 6-block, the blue 8-block. Everyone looking for the matching connector. And the question of what we are actually building tends to come up later in the conversation, if at all.

The real issue isn't the technology

The industry isn't short on solutions. It has excellent building blocks. What's often missing is the blueprint.

Prof. Dr. Kristina Sinemus captured this in her keynote with an analogy that's stayed with us: the electric light didn't come from improving candles.

Exactly that. Transforming healthcare takes more than better blocks. It takes the courage to think something fundamentally new — with the patient at the centre.

Human-centricity, in that sense, isn't a nice-to-have. It is the blueprint. Without it, you may end up building something that technically works — but isn't really made for anyone.

We saw this play out across two concrete fields at the conference.

Field 1: Interoperability

Systems don't talk to each other - not technically, but semantically. Medical staff spend a large share of their time on administrative work that requires no clinical expertise. Duplicate entries, manual transfers, searching for information that already exists somewhere. Just not where it's needed right now.

AI can help build a bridge here. It doesn't connect systems magically. But it understands semantic context. "Heart attack" and "myocardial infarction" mean the same thing. For machines, that wasn't a given until recently. It's changing now.

What really stays with us though: technology isn't the biggest obstacle. It's more about two questions of mindset.

First: do we really understand the people we are building for? Not users in the abstract. The doctor in the ICU. The nurse on the night shift. The patient explaining their condition for the third time. Every role has a different journey, different pain points, different moments where bad software does real harm.

Second: how collaborative are we really? Those who build moats and hoard data win short-term advantages. Those who open systems and create interfaces win a long-term healthcare system that actually works.

Taking both seriously is still hard for many players in the field.

Field 2: AI in practice

The second field where this question is being decided is AI. Here the gap between theory and practice shows up especially clearly.

AI systems often reach 98% accuracy in controlled environments. In the real world, those numbers drop. Because doctors remain skeptical and often don't follow the recommendations. That's not a failure of the technology. It's a signal for where we should be looking.

The human factor is currently the biggest bottleneck. Even the smartest algorithm is of little use if the person holding the stethoscope doesn't trust the screen. The move from black-box AI to transparent, supportive tools is the logical next step.

Two observations from the conversations:

It's worth distinguishing between use cases. For administrative tasks, AI is ready for broad deployment. For core medical decisions, caution is warranted. Treatments shouldn't be derived from an LLM alone.

Digital literacy will soon essentially mean AI literacy. Training medical staff to work with AI is becoming a baseline.

What we take away

The technology is ready. Or at least readier than ever before. What's slowly shifting is the mindset. Pausing for a moment before reaching for the next block, and asking what we're actually building.

What makes us hopeful: the readiness for change is growing. Slowly, but noticeably. And some of the conversations we had at DMEA were already heading in that direction.