AI speeds up service needs assessment in Helsinki by over 50% – more time for customer work

"Our client assistance team seeks information from large masses of information that helps employees provide the right kind of service to the elderly at the right time. Based on a survey conducted during the experiment, hours per week may be spent on information retrieval, as information can be found across different sources, is updated quickly, and legislation is multidimensional. We wanted to find out how AI-assisted information retrieval can support the work of the client assistance team and help serve the needs of senior citizens better," says Jessika Seikkula, Project Manager at the City of Helsinki.

The experiment tested a generative AI assistant with limited data. Helsinki's service needs assessment for the elderly has 200 employees, and each employee can have hundreds of clients. 

"The employees who participated in the experiment reported that the AI assistant had sped up information retrieval by up to 50 percent. In the experimental phase, the correctness of the answers was further verified against the original source, which slowed information retrieval. Previously, searching for information could take four hours a week, for example, and now it takes half as long. The benefits and potential are obvious, but of course, improvements must be made before the actual deployment," says Minna Häkämies, Head of Data Insight at Digia.

The time saved for information retrieval is significant. For example, if one employee saves 2 hours per week, that means an additional 60 hours of customer work in a unit of 30 people.

"A similar application can be used for any job that is based, for example, on the interpretation of instructions or the details of legislation. It can also be used to guide employees who are becoming familiar with finding up-to-date and correct information," Häkämies adds.

AI from visions to the everyday lives of city residents 

Helsinki has achieved promising results in the Kokeilukiihdyttämö project. In autumn 2025, 14 experiments were carried out, of which Digia participated in three. Previously, the cooperation had already tested, e.g., AI image recognition in recognizing errors in the register, and AI could complete weeks of work in 10 minutes.

"The idea is to take AI from visions to the everyday lives of city residents. We want to deliver better-quality services. With the help of the experiments, we can more easily investigate the direct benefits of using AI for employees and city residents. The experiments provide valuable lessons for the development of the city's digital services," Seikkula adds.

Further information:

Jessika Seikkula             
Project Manager, City of Helsinki     
jessika.seikkula(@)hel.fi 
    
Minna Häkämies    
Head of Data Insight, Digia
p. 040 516 6108, minna.hakamies(a)digia.com