How insurers can use generative AI to strengthen risk management

The second in a series on how the insurance industry now he sees Gen AI and its potential.

Insurers applying Gen AI to risk management are finding a variety of relevant applications of the technology for that purpose, according to industry professionals and observers.

Christophe Testi chris@creatives

De-risking has many facets that are ripe for Gen AI, according to Jessica Groopman, senior innovation advisor at Intentional Futures, a digital strategy and innovation consultancy. “This includes creating risk profiles, analyzing huge amounts of data and identifying risks and assessing underwriting eligibility or ineligibility, analyzing regulatory compliance and assisting with compliance reporting or monitoring,” she said.

In addition to identifying risks, Gen AI can help mitigate risks in the form of predictive maintenance, which is also used in other industries, Groopman added. It uses AI to analyze Internet of Things (IoT) devices and predict equipment failures, which can then recommend specific steps that will reduce complaints (and therefore reduce risk).

Similarly, applying Gen AI in the context of parametric insurance can leverage these technologies to reduce risk, Groopman explained. “Parametric insurance, which allows insurance access to developing markets or for previously uninsurable risks, can use sensor data, real-world data, to cover insurance or at least prevent claims processing in areas such as natural disasters or issues of increasing risk and resilience in agriculture. , in food security and in the energy sector,” she said.

The amount of risk data that insurers must sort through and analyze presents a challenge to be able to use that data for risk management, said Kimberly Harris-Ferrante, distinguished analyst vice president, Gartner.

“Will it be too overwhelming for a human to look at? It will slow down or clog the system,” she said. “We hear a lot about knowledge transfer, how I bring my new, less experienced subscribers up to speed with the experience of those who may age and leave the workforce.”

Insurers working on risk management frameworks can take practical steps to evaluate using AI, according to Lauren Finnis, head of business lines, insurance consulting and technology, North America at Willis Towers Watson (WTW). “Once you’ve got your committee together that will really assess the risk and the opportunity and how you might use AI,” she said, “then (it’s) coming to educate your workforce and put the appropriate guardrails in place, making the work to Teach your workforce how AI could benefit them and do it in a governed way.”

Nationwide finds that corporate and operational risks are the biggest risk concerns to manage, according to Todd Lukens, senior vice president and chief information security and technology officer at the carrier. For underwriting actuarial functions, the firm either generates its own models or partners with third parties.

“When we think about using AI, Gen AI particularly in the context of Nationwide’s enterprise, we make sure that any data we provide, whether it’s internal intellectual property or otherwise, has the appropriate protections up to and including how we anonymize that data?” said Lukens. “How do you make sure that before we release that data into the model, we’ve removed all the sensitive data elements? That’s all part of what we’ve called our AI data operations. We think about the data pipeline and how it’s fed into these different AI capabilities.”