Artificial intelligence streamlines social and health services in the Päijät-Häme welfare region

Like many welfare regions in Finland, Päijät-Häme faces the dual challenge of rising service demand and limited resources. Citizens expect fast, clear access to health and social services, while professionals grapple with growing workloads and administrative pressure. Traditional service channels, especially call centers, are struggling to keep up. With digital transformation high on the agenda, the region saw an opportunity: could AI step in to bridge the gap between human need and system capacity?

Rami Autio, CDO and product owner for digital services, highlights the high expectations for AI across welfare regions. He believes that cost savings enabled by AI and support tools in the coming years may be critical for survival given the current financial pressures.

Päijät-Häme and other regions are facing tight resources, and a key challenge lies in guiding customers effectively. The region spans several municipalities with varied health services. Small teams may receive thousands of weekly calls, and they are legally required to maintain a high service quality. To address this, they decided to add a conversational, natural language–based advisor on their updated website, aiming to reduce pressure on call centers, improve discoverability of services, and steer users toward the right channels.

They selected our AI Agent following a tender process. Autio said: “Twoday’s solution was the most advanced, cost-effective, and manageable over its lifecycle.”

Replacing the older chat solution, the new AI-powered advisor offers many benefits, aiding user information retrieval and alleviating staff workload. It answers questions and guides users to proper services.

“We wanted an interactive conversational style without compromising on speed or quality. In Twoday, we succeeded excellently,” Autio praises.

Anneli Vitikainen, head of AI at Twoday, emphasized their experience: “With over 300 use cases and more than 200 clients, we have the expertise and confidence to deliver AI initiatives.”

Compared to the previous hybrid bot, the advanced solution supports vast data sources and flexible prompt engineering by just one person. It learns from conversations, recognizing patterns and delivering highly tailored responses.

The chatbot is trained using website and other data sources, employing deep learning and natural language processing. Training is continuous, ensuring accurate service without compromising data or patient security.

“When serving such a broad public, it's important that systems are secure and information is expert-verified,” Vitikainen states. The service does not provide medical advice and is not a medical device.

Deployment was far faster and more streamlined than prior implementations.

“Twoday executed the project excellently and we are very satisfied with the development process. The transparent and strictly focused monitoring model for the different phases of the project gave clarity on where we are going.” Autio reports. Project manager Tuukka Heinonen also thanked the region’s professional project team.

Future plans include better support for proxy use and improved information retrieval, leveraging evolving language models to offer multilingual, empathetic interactions. Autio adds that AI will free resources for more demanding tasks and foster a positive, secure experimentation culture.

In his summary: “This project is an excellent example of how AI can streamline public services and enhance customer experience.”