Fingrid currently uses AI especially in expert work, but it is also being developed to enhance business processes and customer interactions.
“AI can be used to streamline routine tasks so that people can focus on where their expertise is truly needed,” says ICT Development Manager Hanna Kuusi, who coordinates Fingrid’s AI development.
Generative AI is available to all Fingrid employees. The tool is helpful in areas such as text drafting and processing, documentation, software development, and information structuring. Uptake has risen to over 80%.
The first AI agents are already working in areas such as translation and IT support.
AI also supporting the power system
Many of Fingrid’s AI solutions are still in the development or pilot phase. For example, a generative AI tool is being integrated with the case management system so that information from different sources can be found with a single search in the future.
Fingrid uses AI in customer interactions, one example being the Datahub chatbot, which allows customers to request information about missing metering data or how to log into the customer portal, for example.
AI is also being piloted for processing crossing statements. Its role is to produce a preliminary opinion, which an expert then reviews, edits if necessary, and approves before sending it. The aim is to reduce the time it takes to process statements and to reduce the burden of routine tasks for statement handlers.
“The goal is to target maintenance more flexibly and identify risks in good time.”
In addition to expert work, Fingrid has piloted AI in support of the power system operations. One key development is the use of machine learning to predict failures of substation equipment.
“Monitoring systems continuously collect information on the status of equipment, and models analyse anomalies that may indicate disturbances. The goal is to target maintenance more flexibly and identify risks in good time,” Kuusi explains.
Helen develops AI as part of the team
Other players in the energy sector are also using and developing AI for their own needs. The sector relies on vast amounts of data, which AI is excellent at processing.
“The energy sector has been using AI for a long time, especially to forecast production, consumption and networks,” says Mikko Muurinen, Head of Data & AI at Helen Group.
At Helen, AI has become a crucial part of the development of energy systems and customer experience. Initially, the focus was on analytics and forecasting, but AI is now being used to generate new insights, for example about energy saving opportunities, based on historical consumption patterns.
“AI is used to make systems more efficient and reduce emissions.”
“It is used to make systems more efficient and reduce emissions, as well as to make predictions, whether they focus on the minutes ahead or a horizon of several decades,” says Muurinen.
In customer service, AI deepens customer understanding and enables a more personalised service. This can be seen, for example, in AI agents embedded in websites and AI-assisted customer service. By 2027, Helen aims to resolve more than half its customer service interactions with the help of AI.
According to Muurinen, AI will be built directly into Helen’s processes so that it will start to function more and more naturally as part of expert teams.
A common direction for the energy sector
As the energy sector develops AI for various needs, collaboration between actors becomes more important; questions about AI reliability, governance, and the role of humans are shared across the entire energy sector.
“Together with Fingrid, we have promoted matters such as AI development models and exchanged ideas on more technical issues. We have also considered how different agents and assistants can be used safely and appropriately,” says Muurinen.
High-quality and sufficiently comprehensive data is the basis for the development of AI, and Helen and Fingrid have also had many discussions about this.
“Reliable data is essential for building AI solutions that better serve the entire energy sector,” says Kuusi.
