Case study

Municipality case study: Kunta-Kati and Riihimäki city

Riihimäki City

The city of Riihimäki is located in the south of Finland, about 90km north of Helsinki. It is home to close to 29,000 inhabitants and, in 2019, was the country’s first municipality to launch a conversational AI-powered virtual agent on the boost.ai multi-city Kommune-Kari platform.

Rebranded as Kunti-Kati for the Finnish market, and developed in collaboration with Accenture, the platform is powered by boost.ai’s market-leading Natural Language Understanding (NLU). This allows local governments to deploy a unique virtual agent, tailored for each municipality, that is connected to a central knowledge hub for all public-sector-related questions on more than 1,500 ready-to-go topics.

Employees from each municipality train and maintain their respective virtual agents whilst simultaneously benefiting from the conversation data and improvements of their neighboring locales also using the Kunta-Kati platform. This is all managed within a no-code user interface requiring zero technical expertise to use.

Kunta-Kati will respond to search queries on the Riihimäki website

Residents of Riihimäki have access to Kunta-Kati 24/7 via the city’s website. The virtual agent can answer questions ranging from education and healthcare to construction permits and more, and is constantly being updated with new information and services to help expand its scope further.

In 2021, the municipality repositioned Kunta-Kati to be more visible on its website. This included integrating the virtual agent into the site’s existing search functionality so that, rather than receiving a page of search results when using the website search bar, the virtual agent chat window would instead pop-up with the relevant information and the opportunity for users to ask follow up questions.

Not only did this simple visibility adjustment make Kunta-Kati more accessible to visitors by providing a dynamic and approachable channel for them to find the information they need, but it also increased traffic to the virtual agent by almost 1,000%.