AI in healthcare: from tasks to transformation

AI is changing healthcare, but how, exactly? That question was at the heart of a recent evening discussion with healthcare professionals, technologists, and industry experts. The event combined a keynote from Petteri Raejärvi, a Medical Director of Digital Systems at Terveystalo, with roundtable conversations exploring AI's practical realities, opportunities, and challenges in healthcare.

Petteri’s perspective: where medicine meets technology

Petteri opened the evening with his vision for AI in healthcare. Having worked extensively in telemedicine, he shared how AI can help clinicians focus on what matters most: the patient.

He outlined three stages of AI adoption:

1. Task automation: Simple but impactful uses of AI, like summarizing patient records or drafting appointment notes. Standalone tools help with individual tasks but rarely touch the bigger workflow.
2. Appointment automation: AI becomes embedded in the workflow. Ambient scribing, for instance, can generate appointment notes in real time, giving doctors more face-to-face time with patients.
3. Healthcare pipeline automation: The long-term vision: a system where AI supports patients across the entire care continuum, guiding interventions and improving outcomes while professionals remain central to decision-making.

Petteri highlighted Terveystalo’s work, which integrates AI tools directly into workflows, ensuring smarter coordination and better data handling. He also shared examples from digital medicine, where AI has helped reduce high-risk prescribing and freed up time for doctors to engage more meaningfully with patients.

“AI only creates value when it fits the workflow, not when we try to make it fit old ways of working,” Petteri emphasized.

Roundtable insights: exploring AI in practice

After Petteri’s talk, the discussion broadened to include the participants’ experiences and questions about AI in healthcare. Some of the key themes that emerged:

Trust and reliability

Can large language models (LLMs) be trusted to make medical inferences? Participants stressed the importance of validation, transparency, and safeguards. Suggested strategies included:

• Building compliance and explainability into applications with clinical quality validators.
• Collecting validation data to ensure AI outputs match expectations.
• Leveraging public-private collaborations to responsibly scale health data use.

Where AI adds value 

AI isn’t uniform across healthcare. Telemedicine, diagnostics, and chronic disease management are seeing immediate impacts. Successful development requires collaboration between providers, tech companies, and researchers.

Adoption and human factors 

Technology alone isn’t enough. Adoption depends on human behavior:

• Early adopters embrace tools naturally.
• Others need personal guidance, clear value, and workflow alignment to engage.
• Behavioral and cultural change is as important as software deployment.

Practical considerations

AI is powerful, but it’s not always the answer. Sometimes optimizing workflows, improving data management, or enhancing service design can have a more immediate impact. Voice interfaces and ambient scribing were highlighted as promising tools that reduce friction in care delivery.

What we learned

Several lessons stood out from the evening:

1. Start with real problems. AI works best when it addresses actual clinical and patient needs.
2. Integrate, don’t isolate. Embedding AI into core systems ensures it complements, rather than complicates, workflows.
3. Validate rigorously. Transparency, reliability, and compliance are essential for trust.
4. Engage users personally. Adoption is about people, not just technology.
5. Think long-term. The ultimate goal is smarter, more efficient healthcare pipelines, not just task automation.

The evening painted a clear picture: AI in healthcare holds immense promise, but success depends on thoughtful design, careful validation, and a human-centered approach. Technology alone won’t transform healthcare, but used wisely, it can help clinicians focus on what matters most: the patient.