Europe’s discussion about AI often starts in familiar territory: adoption rates, investment, company valuations, productivity gains and the continent’s position relative to the United States and China. Instead, we should reframe how we’re viewing success with AI and focus on how it is truly helping and solving critical problems.
Healthcare offers some of the clearest examples, bringing together Europe’s strengths in scientific research, clinical expertise and regulation, while creating opportunities for AI to deliver outcomes that can be measured. Earlier diagnosis, safer procedures, quicker access to treatment and better prevention are improvements in people’s lives.
Shortening the journey from discovery to treatment
Drug development is a field where time and cost have traditionally been major barriers. Researchers can spend years testing possibilities before identifying a viable treatment. Generative AI and laboratory automation are beginning to change that equation.
Iktos, a Paris-based startup, combines generative AI with automated laboratory robotics to design and test molecules more rapidly. By helping researchers identify promising candidates earlier, its approach can compress parts of drug discovery from years to months.
Every month removed from development potentially brings a new treatment closer to patients. It also demonstrates where Europe can compete: at the intersection of advanced research, scientific expertise and commercial application.
Putting specialist knowledge where it is needed
Technology can also address healthcare inequalities by moving expertise rather than requiring patients or professionals to move. Consider surgery: safe and timely procedures remain inaccessible to millions of people worldwide. In many cases, the challenge is not simply a shortage of facilities. There is also a shortage of specialist knowledge and the ability to coordinate it across distances.
UK-based Proximie uses a digital platform to connect operating rooms, enabling surgical teams to see, support and learn from colleagues elsewhere in real time. A specialist can assist another team with an unfamiliar procedure without physically being in the same hospital. The goal is to extend specialist knowledge to where it is most needed.
Making innovative treatments easier to reach
Scientific progress only matters if patients can benefit from it. Yet even when potential treatments exist, navigating clinical trials, expanded-access programmes and other routes to pre-approval therapies can be complicated.
Amsterdam-based myTomorrows uses AI to help physicians and patients navigate this fragmented landscape. With more than 300 million people worldwide living with conditions without an approved treatment, and many thousands of therapies in development, the challenge is finding the relevant option and connecting the right patient to it quickly enough. That is precisely the kind of complex information and logistics challenge where AI can have practical value.
Moving mental healthcare from reaction to prevention
Mental healthcare provides another example of where better measurement can change the conversation. Relapse is common, but deterioration may develop between appointments, making early warning signs difficult for clinicians to detect.
French startup Callyope uses AI to analyse speech samples alongside medical records, helping identify potential signs of relapse and generate alerts for clinicians. Its platform supports mental-health professionals in monitoring patients remotely and at scale. The promise is changing the timing of intervention: identifying deterioration earlier, when support may be more effective, rather than waiting until a crisis becomes visible during a scheduled appointment.
A common thread across different companies
Iktos, Proximie, myTomorrows and Callyope, all part of the AWS Pioneers Project, operate in different parts of healthcare, but they point towards the same idea. They work at different stages of the journey from scientific possibility to patient outcome: discovering treatments, helping clinicians deliver care, connecting patients with options and identifying risks before they become crises. That is a more meaningful definition of AI progress than adoption.
Europe should build on the advantages it already has: strong research institutions, deep clinical expertise and a regulatory framework capable of creating trust in high-stakes technologies. Healthcare is where these strengths come together most clearly. Regulation should not be viewed only as a constraint; when designed well, it can become part of Europe’s competitive proposition by establishing confidence in how AI is used.
The central question is not whether Europe is adopting AI quickly enough; it is whether adoption is producing outcomes that truly matter. If Europe wants a more credible measure of success, it should look beyond the number of systems deployed and the value of the market. It should ask whether diagnoses are happening sooner, whether surgery is becoming safer, whether patients can reach treatments more fairly and whether care can become more preventative. Those are harder metrics to reduce to a single ranking. But they are far closer to what winning with AI should truly mean.
