AI taking bigger slice of insurers’ tech budgets: DBRS

Though artificial intelligence technology has long been part of insurers’ underwriting modelling, North American insurers have increased the proportion of investment technology budgets allocated to AI — from 8% in 2024 to more than 20% in the next three to five years, says new Morningstar DBRS commentary.

Recent data supports this; the insurance sector is outperforming the global average when it comes to AI implementation, with 99% of insurance organizations using AI in different ways, according to new data from Kyndryl. Plus, insurance is leading globally for employee trust and sentiment toward AI compared to other industries, according to the global IT infrastructure services provider. 

That said, while several areas of the insurance value chain can benefit from AI adoption, the tech also presents risks, say DBRS authors.  

“Although the insurance industry isn’t typically the quickest adopter of leading-edge technologies, it does extensively leverage large datasets of information,” the report reads.  

“Ultimately, companies need to invest in AI to stay competitive; however, at the same time, they must not lose sight of the importance of having commensurate risk management frameworks.” 

AI use cases 

Insurers are increasingly using AI and large language models to boost efficiency and cut operational costs. AI tools can automate repetitive tasks like policy generation, data extraction, and customer interaction summaries, allowing staff to focus on more complex duties.  

“Many insurers have already reported cost reductions from similar efficiency and automation gains,” DBRS writes.  

And based on publicly disclosed use cases by some insurers, AI adoption has a return on investment of roughly three years or more. 

AI also helps identify customer behaviours to streamline sales and reduce acquisition costs. Insurers are also deploying AI-powered chatbots to simplify policy selection and provide investment guidance. 

In property insurance, AI can assess damages from images, estimate repair costs, and quickly evaluate exposure to catastrophes, though human intervention is required for complex or compliance-heavy tasks like claims handling. 

Fraud detection is another key benefit, as AI can flag suspicious claims for review and speed up the payment of valid claims.  

The technology can also create valuable insights for risk selection and pricing by reviewing granular data and improving existing risk models.  

“This may ultimately help insurers write more policies with consistent pricing for similar risk profiles,” says DBRS. “On the other hand, since underwriting decisions have a direct impact on profitability, AI models need to be carefully selected, trained and tested as otherwise mispriced policies could result in very serious reputational and financial implications.” 

The biggest AI risks 

For a risk-adverse industry, it’s true that some AI activities are inherently less risky than others.  

For example, “determining marketing leads based on AI recommendations is generally a low-risk proposition,” DBRS writes.  

On the other hand, overreliance on AI for claim rejections may pose larger legal and reputational risks. “AI models could draw conclusions about which claims to reject based on flawed reasoning that could result in class action lawsuits.” Though it hasn’t yet occurred on the P&C side or in Canada, according to DBRS, there are examples of these suits against health insurance companies in the U.S. 

“But, in our view one of the most serious challenges arises when AI is used extensively in [the] underwriting and pricing of policies as those decisions are directly related to profitability,” says DBRS. 

In those situations, the insurer could be subjected to various costly errors and biases — such as quotes that are unreasonably high or low — if risk characteristics aren’t well represented in the data that trained the AI.  

Additionally, insurers could face risk as the regulatory landscape evolves to accommodate the burgeoning use of this tech.  

Lastly, the large data sets used by insurers’ AI could also expose them to increased cyber risk.  

That risk is higher for smaller companies compared to larger ones, says DBRS. Research by IT company Wipro shows that “even extensive adoption of AI is not always accompanied by data usage policies, which is concerning. Moreover, the survey also shows that smaller companies are more likely to not have data policies even where their adoption of AI is considered extensive.”