Artificial intelligence holds tremendous potential for unlocking real, positive change in healthcare. Yet deployment of AI solutions in clinical settings remains surprisingly limited. Bridging the gap between a promising algorithm and a functional bedside tool requires a “different mindset” — one that moves beyond isolated, controlled experiments and designs systems directly alongside the people who will use them. We sat down with Anouk Wolters, AI Project Lead at KickstartAI, to find out what that looks like in practice.
Written by Berna Syuleymanoglu
As Anouk Wolters, AI Project Lead at KickstartAI, explains, developing an AI solution that makes a difference in healthcare means stepping away from the screen into the chaotic reality of a hospital. “If you only start from a technology perspective and you envision that something might work, that’s not enough,” she said.
“You really have to go the other way around. So, start with the practitioner, start with the nurse, with the doctor, and then really understand their pain points.”
Artificial intelligence holds tremendous potential for unlocking real, positive change in healthcare. Yet deployment of AI solutions in clinical settings remains surprisingly limited. Bridging the gap between a promising algorithm and a functional bedside tool requires a “different mindset” – one that moves beyond isolated, controlled experiments and designs systems directly alongside the people who will use them.
A clear example of this philosophy in practice is KickstartAI’s automated wound-reporting project with Leiden University Medical Center (LUMC), currently underway. We sat down with Anouk to discuss how this methodology works, and how her team is building a system that fits into a nurse’s actual daily workflow.
"You really have to go the other way around. So, start with the practitioner, start with the nurse, with the doctor, and then really understand their pain points"
The initiative grew out of the National AI Challenge 2025, organized by KickstartAI alongside the AI Coalitie voor Nederland (AIC4NL). Nurses are required to document wounds using an internationally recognized framework, but the process today is manual, time-consuming, and prone to subjectivity.
The goal of the project is to ease that administrative burden and improve consistency by automatically generating wound reports from a single photograph.
“It is about helping nurses reduce their administrative workload, and at the same time, make wound administration more consistent and complete,” Anouk said.
To close the gap between algorithm and clinical reality, the KickstartAI team went into the hospital to see, firsthand, the daily struggles nurses face when handling wound reports.
“We tagged along with different nurses to really see at the bedside, or during a patient appointment how this process works in practice, and understand the bottlenecks they experience first-hand,” Anouk shared.
"It is about how technology interacts with people, the guidelines it needs to adhere to, and the governance surrounding it. You have to perceive it as a whole system, rather than just placing a technological solution somewhere just for the sake of it."
Building effective healthcare technology requires a deep commitment to responsible AI. As Anouk emphasized, developers must “never see the technology in isolation,” especially when deploying software within the tightly regulated ecosystem of a hospital.
Instead, success requires looking at the entire socio-technical architecture as a unified system. “You have to look at the system as a whole,” Anouk noted. “It is about how technology interacts with people, the guidelines it needs to adhere to, and the governance surrounding it. You have to perceive it as a whole system, rather than just placing a technological solution somewhere just for the sake of it.”
For KickstartAI, that meant building a genuinely multidisciplinary team. KickstartAI engineers and project leads worked together with Cairelab, LUMC’s AI expert center, which brought experience in developing and implementing AI in a hospital environment, including dealing with complex IT infrastructure and healthcare regulations. Nursing Information Officers (NIOs) were crucial in formulating the key clinical questions and, alongside nurses, were closely involved throughout the project.
Anouk shared the importance of this approach in these words: “During the project, it is important to discuss and validate choices with them, to make sure that we create everything in such a way that it really fits their world”.
This nurse-in-the-loop approach directly shaped the design of the application.
Another key pillar of the strategy is positioning the AI as a supportive assistant rather than a replacement for clinical judgment, so healthcare professionals stay critical and accountable at every step.
“It is always clear that they should have the final say and that the AI system is just an assistant, that it’s there to make suggestions,” Anouk explained. “You want nurses to trust the solution, but you also want them to remain critical and accountable for the final decision.”
"We tagged along with different nurses to really see at the bedside, or during a patient appointment how this process works in practice, and understand the bottlenecks they experience first-hand"
Asked how this technological solution could evolve, Anouk was clear that “this is indeed the first step.” She sees the methodology and application scaling to other hospitals in the Netherlands and potentially across the EU.
Looking ahead at how AI can serve healthcare over the next decade, Anouk pointed to the value of tackling the “low-hanging fruits” first. The long-term possibilities of AI are vast, but its value today can be much more practical: reducing paperwork and improving planning so healthcare professionals have more time and attention for their patients.
That, in the end, is one of the clearest lessons from the project: that the people doing the work need to shape the technology built around it. In this case, that meant working closely with the nurses at LUMC, alongside the NIOs and Cairelab’s implementation experts, and learning from the realities of wound reporting in their day-to-day work.
“We need to give people on the work floor the space to share their experiences from their own perspective,” Anouk said.
Their frustrations, routines, and ideas about what could work better provided the direction for the project. The same principle applies more broadly: AI is more useful when technical development is grounded in the experience of the people it is intended to support. This is why AI4ALL is proud to be sponsored by KickstartAI, which works to make AI benefit people and society across the Netherlands.
About KickstartAI
KickstartAI is a non-profit foundation working on the application of AI across the Netherlands. Together with its partners, it develops AI solutions for real-world challenges, puts them into practice, and shares the knowledge and lessons so that the benefits can reach the wider society.
To discover more about this project and other practical AI solutions, visit the KickstartAI website.
Want to be part of what’s next? Visit our website to learn more about AI4ALL’s upcoming Meetups, the Entrepreneur Program, Masterclasses, and the Mentor Program. Join our movement shaping AI’s future together.