Insurance and the AI race: are you in or out?
Insurance
Readingtime: 4 min
“We have reached a good level. We probably have the most potent artificial intelligence (AI) solution in production. It fits well with the insurance market needs and delivers value. That places us among the front-runners, but that’s not what’s important,” said Arndt Gossmann, chief executive officer of DGTAL, discussing the company’s position in the digital landscape.
As the reinsurance industry undergoes significant transformation, Gossmann highlights the urgency for companies to engage with AI technology.
DGTAL has been ahead in the AI space since its launch in 2021. Gossmann noted: “When it comes to putting an AI product in the re/insurance claims assessment workflow we are at a certain level and there are not many of us.
“We’ve not seen anybody who has a comparable level, but it is, of course, a journey. A common complaint we keep hearing from insurers is that they still don’t see immediate return on their investment from AI.
“That’s because expectations are misaligned. Aiming for full automation is premature as AI alone is not yet able to consistently make sound decisions. But AI is extremely powerful in uncovering all the information one needs to make much better decisions.
“Implementing human-machine collaboration delivers greater efficiencies and shows immediate results.”
Gossmann emphasised that the rapid evolution of AI presented unique opportunities and challenges for the entire re/insurance industry.
Looking at the complete data
Elaborating on the integration of AI into handling catastrophe claims with DGTAL’s AI tool, DRILLER, Gossmann explained how the tool acted “as a co-pilot” for claims handlers, portfolio managers, actuaries and auditors.
“It allows them to monitor and explore insurance claims portfolios in detail and almost in real time while looking at all unstructured data and combining it with structured data,” he said.
“Only 2 to 5 percent of the available information applies to the structured data of insurance companies. All the rest is unstructured, which makes it difficult to analyse,” Gossmann continued.
“If you can systematically have an overview of 100 percent of your information, it provides deep insights into how to deal with it. For emerging risks such as cyber, where we lack good historic data we’ve learned that we have to wait for sustainable trends in the data.
“If you only have structured data, it needs two to three years until actuaries can make a sound prognosis.
“However, combining qualitative data from claims with transaction data, you can understand trends and react much earlier.”
Operational efficiencies are a key benefit of AI, although Gossmann explained that AI works only with machine-readable data, meaning there is a need to make data AI-ready first.
“All our AI tools have a strong front engine that translates all unstructured data into structured data,” he noted.
“Analytical AI allows you to work with these metadata, structured, and unstructured data in a more progressive way, which primarily helps to get fast insights.
“Regulators are keen to understand the possibilities of AI to reach better conclusions.”
“Generative AI (Gen-AI) is coming up now and will benefit operational procedures even more. It will add another raft of efficiency gains—it is much more flexible, allowing for better manoeuvrability within AI systems.”
Gossmann quoted a Harvard Business Review article reporting Accenture research, which concluded that Gen-AI will augment operational efficiency dramatically, and in insurance more than other industries
Adopting AI one day at a time
Climate-related risks are becoming more prominent, and Gossmann highlighted AI’s role in compliance and adaptability. “The spectrum of what you can do expands, and regulators are keen to understand the possibilities of AI to reach better conclusions,” he said.
“AI can empower decision-makers and controlling entities. One regulator said it wanted to know which questions it can ask going forward, showing a clear interest in state-of-the-art technology.”
“AI is still new to the whole sector: there’s not much difference between Europe, the US and Asia, although our experience shows that there seems to be a bit more curiosity in the US about trying out AI tools.
“We so far follow European clients to the US and deliver pilots they want to implement in their American operations,” Gossmann explained.
“AI’s successful implementation is about adoption—not just top-down mandates. The trend is to develop focused use-cases that support and deliver value to the people using them from day one.
“We need to support curiosity in these projects and for pilots to happen, which helps start the transformation of the entire organisation.”
Gossmann noted challenges and barriers to AI adoption, stating: “US companies are curious, particularly in operational departments. They want proof of concepts before diving in, but once they see it working, they want it immediately.
“In re/insurance, if you wait too long to start, you will miss out on the opportunity to create a competitive advantage that you might need,” he concluded.
Arndt Gossmann is the chief executive officer of DGTAL. He can be contacted at: a.gossmann@dgtal.io
For more news from the American Property Casualty Insurance Association (APCIA) click here.