Perfecting customer service with Machine Learning



Aarhus Borgerservice


Public Administration

Helping out a neighbour

The Danish citizen service in Aarhus Municipal, commonly referred to as “Borgerservice” in Danish, operates as a comprehensive public service facility aimed at delivering a diverse array of services to citizens in Aarhus.

Borgerservice’s fundamental objective is to facilitate convenient access to government-related services and information for residents. The range of services provided by Borgerservice spans from inquiries about hobbies/sports to crowdfunding, day care, and elderly home services. Consequently, addressing queries often demands a nuanced and extensive knowledge base, requiring first-line respondents to provide instant and informed assistance.

Aarhus municipality office have always had a keen eye when looking to improve their services. They have an innovative way of thinking and working with data – which is why they were first mover within customer service, introducing an agent assist solution in “Borgerservice”.

Borgerservice chose the pace

Onboarding of new employees is one of the most time-consuming tasks of Borgerservice since a new employee needs time to adjust and learn the service items in order to provide the best possible care.

We therefore create a solution Agent Assist – which is a personal assistant to the customer care employee. The Agent Assist solution – with the nick-name “Anni”, optimizes the onboarding process for new employees. The solution is easy to use to use and has additionally reduces the average waiting time. All summed up in three overarching goals:

  • Deliver the best possible service to the citizens calling in
  • Faster onboarding of new employees
  • Reduce the number of calls made to the back-office experts

We always want to provide the best service, to us that means that you don’t have to wait several minutes on the phone to get help. The solution is a brilliant tool that helps us provide better service to the citizen and it allows us to also use guides that aren’t used as often.

Janni Søvang

Project & system responsible, Borgerservice Aarhus

Machine Learning to the rescue

We delivered a speech-to-text solution that helps customer service agents find their support articles based on the conversation between a service agent and a customer. The solution is able to swiftly identify the right articles for the given scenario, but also suggest similar articles or relate other subjects based on the conversation.

This makes it significantly easier to onboard new employees as new hires are less required to know all articles right away and instead be guided by the solution. As previously mentioned, the machine learning aspect will result in a more efficient and user-friendly solution.

Trifork’s mission is to make the world a better place, together with Borgerservice Aarhus we have been able to be first mover in making automatic speech recognition in danish, but there are still plenty of options to explore.

Jens Peter Hedegaard

VP, Trifork

Continuously improving

The system is build on Triforks generic Smart Service Center platform, that with intensive use of Artificial intelligence converts voice to text and intelligent keyword extraction. The solution is bound with microservices and containers that can be deployed in the Cloud and On-Premise.

The solution automatically learns and adapt from the agents usage and feedback. The user interface is web-based and interacts and blend in with Borgerservices existing web-based tools. The platforms microservice architecture makes it easy to connect to other call center solutions and user interfaces.