Beyond the Algorithm: AI as a Business and Organizational Challenge
Artificial Intelligence (AI) is often framed as a technical phenomenon—a matter of algorithms, data pipelines, and computing power. But this framing tells only part of the story.
Claim the future
Norwegian insurance company Gjensidige are now implementing AI-agent Eva to handle claims from clients directly. They are aiming for Eva to handle up to 70 percent of claims from individual customers in the future.
Research centre DIG (Digital Innovation for sustainable Growth) at NHH invited partners and industry for an update on "The Invisible Hand of AI" in Oslo in April. It was time to take a look at how AI is already shaping plans of major Norwegian corporations.
Vivi Kofoed is Executive VP for claims in Gjensidige, and talked about their plans for how the company will utilize the possibilities the new technology represents. She said that post-pandemic, they have seen a significant rise in the overall number of claims the company receives. She pointed out three factors driving the rise in claims; electrification, reopening and health.
- We still have problems with electrification and don’t get it right, even with all the help. We are getting more claims related to cases from international travel, and we see an increase in health-related claims from people wanting to get better sooner, she said.
In total Gjensidige saw a 30 percent rise in the number of claims and set a record for ongoing cases. Should this increased burden be met with increased staff numbers, costs could spiral out of control.
The answer for Gjensidige was the first engaging claims agent, Eva. Customers now have one access point, available around the clock. The ambition is for Eva to handle 70 percenty of all claims.
- This makes it possible for the Gjensidige and our staff to fulfil our promise of being there for the customers when they need it the most, Kofoed said.
Eva, Sofie and Frank
In addition to Eva, there are two other AI-agents now also working for Gjensidige. Sofie is in charge of interaction between members of staff of the various departments within Gjensidige, making sure all internal communication related to casework flows smoothly. The aim for Sofie is to cut costs by 70 percent. Frank is the platform for communication and handling with all outside partners needed to handle cases, such as surveyors, builders and other trades.
Vivi Kofoed said it was a conscious decision to give the AI-agents human and not more technical names.
– We want it to feel normal and natural to work with these agents and not make them sound scary, she said.
– We are now in the middle of an historic re-skilling and re-training of all staff working in the claims department, Kofoed told the workshop.
Self-conscious AI-agents
Professor Eirik Sjåholm Knudsen at NHH and DIG revealed how AI-agents and applications are developing into a stage where they are aware of context as well as self-conscious. But he also finds that the current AI-applications are often wrong, but always certain, as they are programmed to be creative. – So far we often find it difficult to make the applications solve the actual problems we want them to solve. We need the people who owns the problem to take the driving seat in future development of AI-applications and -models, he said.
Virtual work force – not a tool
However, the direction is clear, according to Eirik Sjåholm Knudsen.
– AI will become a multiplier of the workforce, rather than a tool, he said. He added that the current approach is trying to reign in AI, rather than training it.
– We are not working in ways that use AI most efficiently but rather using AI to make our existing ways of working more efficient.
Regulations and skills are part of the problem
Advicers and accountants from consultancy KPMG as well as Norwegian tech sector industry association Abelia are both confirming the impression that here are some obstacles to overcome before releasing the full potential of AI.
Regulations are making it difficult for organizations to train AI-agents on current customer data, as use of these data is very limited.
Many in Norwegian industry are also experiencing a shortage of expertise and skills in the application and use of AI. This leads many industries to keep implementation of AI on hold, even if it is accepted that a transition to AI is necessary and will become important.
10 AI-predictions for 2025
2024 was the year artificial intelligence (AI) dominated the media, the stock market, and consumers' daily lives. A search for "artificial intelligence" in Aftenposten yields 1,189 results. Generative AI, such as OpenAI’s ChatGPT, experienced the fastest adoption rate ever. The value of AI companies like Nvidia, Meta, and Alphabet soared.