Is AI Really the Solution to Your Problem?

On 3 December, 117 participants gathered at the Health Innovation Centre of Southern Denmark, where the AI Competence Unit in Regional IT and the RAIN network (AI in Healthcare) focused on how technology, both with and without AI, can support daily operations in the healthcare system. 

The participants represented a wide range of professions: clinicians from various specialties, administrative staff, project managers, consultants, IT and data professionals, as well as leaders and support functions.
What united them all was a clear curiosity, and a better insight into the available possibilities and how these might be applied in practice. 

Needs Before Technology: Setting the Framework for the Day

The network meeting was introduced by Thomas Schmidt, Head of Digital Innovation in Regional IT, who set the stage for the discussions. He reminded participants why defining the needs must precede choosing the technology, stating: 

AI is automation, but automation is not necessarily AI. We should base our decisions on needs and frameworks first, then evaluate technological options. Whether we’re dealing with specialised or generative AI, everyday applications are becoming increasingly common - allowing us to make smart choices when selecting solutions.

This message naturally set the tone for the rest of the programme, highlighting both AI and other technological approaches. 

Automation in Practice: When AI is Not the Answer

Several presentations demonstrated how automation can, in fact, be a powerful solution on its own.

Robotic Process Automation (RPA): A Reliable Everyday Tool

Peter Engholm, automation specialist in Regional IT, introduced participants to Robotic Process Automation (RPA) as a digital assistant tool capable of performing rule-based tasks with high predictability. He explained the differences between AI and RPA and showcased examples of various digital robots already in use in the region, supporting everything from documentation to data validation. 

The presentation demonstrated how automation can free up time in departments dealing with heavy, repetitive, and administrative workflows. 

Automation Applied to Clinical Workflows

Next, Maria Sølvberg, Head of Secretariat at the Orthopaedic Surgery Department, Odense University Hospital, and Mathias Jul Jørgensen, RPA developer in Finance and Data at Odense University Hospital, presented a solution that left a strong impression: A robot that ensures the correct input of technical data for operations.

Previously, error rates were as high as 50%, complicating booking processes and prolonging surgical workflows. The new robot has significantly reduced errors, improved data quality, and delivered more stable patient pathways. This case highlighted how value is created when clinical needs and digital competencies are closely integrated. 

"AI – Yes or No?” A Necessary Critical Perspective

In the third presentation, Sabine Morris Delhez, physician and PhD researcher at Odense University Hospital and University of Southern Denmark, explored the considerations that should precede the adoption of AI. She identified key challenges, including automation bias, lack of generalisability, and unclear accountability. She encouraged participants to ask fundamental questions:

• What is the goal?
• What is the hypothesis?
• What needs to be tested?
• Is it worth the investment?

Her presentation prompted reflection, emphasising that AI is not always the best choice and that simpler technological solutions may often be preferable. 

Generative AI in Daily Use: SydChat and Responsible Applications

The AI Competence Unit introduced participants to generative AI and the region's proprietary chatbot platform, SydChat. 

Kathrine Holtz Paarup, Project Consultant in the AI Competence Unit, explained the principles behind language models, including pitfalls like hallucinations and automation bias. Participants engaged in hands-on prompting exercises and learned how generative AI can support tasks such as idea generation, drafting text, communication, and structuring documents – all without processing patient data. 

FlowPriority – Smarter Use of Data

Kenneth Bo Pedersen, Chief Physician at the Department of Medical Gastrointestinal Diseases at Odense University Hospital, and Kristian Kvist, Chief Consultant in Documentation and Management Information, presented FlowPriority: a solution providing clinicians with a data-driven overview of patient populations. 

They demonstrated how systematic and dynamic use of existing data can yield significant benefits. In one outpatient clinic, FlowPriority reduced administrative workloads by 5–7 hours per week and freed up 2–4 hours per week for physicians, which could instead be dedicated to patient care.

The presentation illustrated how data-driven solutions can deliver substantial value without involving advanced AI.
For more information about FlowPriority, you can contact Kenneth Bo Pedersen at kenneth.bo.pedersen@rsyd.dk. 

Focusing on Needs First

The day's final presentation by Anne Hagelskjær Asmussen, Special Consultant at Health Innovation Centre of Southern Denmark, tied the discussions back to the core message of the event. 

Using tools such as user journeys, reverse brainstorming, and prioritisation models, she showed how to identify the right problem and why this is crucial for selecting the right solution. 

A key point was that while users often know their wants, they may not necessarily understand their needs. 

From Needs to Solutions: Reflections on the Day

The day provided participants with the opportunity to explore various technological approaches, discuss their own challenges, and gain insight into technical, organisational, and needs-driven perspectives. 

It also served as a space where the AI Competence Unit’s technical expertise and RAIN’s network framework enabled clinicians and other professionals to access knowledge, advice, and practical experiences that can be difficult to obtain during a busy work schedule. 

Both units contributed to a day that bridged the gap between technology, practice, and needs. 

A huge thank you to all participants and presenters for an engaged and inspiring afternoon, demonstrating that the path to the right solution starts with the need and AI is just one of many possibilities.