AI levels the playing field, as insurers rush to build competitive data advantage

The widespread accessibility of AI models has fundamentally levelled the insurance playing field, as experimental pilot AI programs are starting to be deployed on a production scale — and insurers are seeking to distinguish themselves in their approaches.

“We need to recognise that when ChatGPT came out and it became accessible, it’s in effect a leveller. The standards are rising, but we are all starting from the same level,” says Clemens Philippi, CEO of MSIG Asia at an InsureTech Connect Asia 2025 panel on Thursday (June 5).

Insurers will now have to think closely about what distinguishes their AI model from others—and forward-thinking carriers are now leveraging this foundation as a springboard for a competitive advantage.

Data has become the distinguishing factor, Philippi adds. “We enrich what’s available from GPT with our own data,” he says. “And if we move into agentic AI going forward, we [can] get specific answers and decisions based on a Thai customer mindset or a Malaysian customer mindset.

(Re)insurers and insurtechs have to build a solid data foundation in order to succeed in this new paradigm.

And quality data is crucial, says Mohit Tandon, Chief Operating Officer, Technology at Prudential. “Ultimately it’s garbage in, garbage out. If you have high ambitions but you don’t have good quality, reliable data, then you’re going to struggle with that.”

Data, data, data

The large, complex, unstructured problems are gradually being tackled with AI. Prudential began integrating Google’s MedLM in Singapore and Malaysia in 2024, using it to support human decision-making in the claims process, and allowing it to identify fraud, waste and abuse.

Amazon Web Services (AWS) has worked with insurers in specific areas to build genuine use cases, having worked with AXA to automate 10% of their claims and reduce processing costs by 45%, says Gunish Chawla, Managing Director, Commercial Sector, ASEAN at AWS.

While insurers have a “tremendous amount of data”, there’s no need for them to start consolidating everything from the outset. Rather, Chawla advocated for a targeted approach. “You could choose to go into claims or go into underwriting and build that genuine use case.”

High-quality, well-organised data can help support AI-driven decision-making across various parts of the insurance value chain, from underwriting, claims processing, and risk assessment. But that all starts from good data hygiene, says Richard Cooper, Business Leader for Mercer Marsh Benefits in Singapore.

“We’re finding those patterns, finding those relationships that probably humans are unable to see, and the technology is supporting us,” Cooper says. “From a risk point of view, those data points are giving us much stronger insights into what are the drivers around risk. Better risk means better claims. Better claims mean better premiums.”

Without good data hygiene, insurers face risks, such as data bias, limited functionalities, and privacy. “If underwriters cannot explain how AI made a decision, customers lose trust [in the insurer],” says Atsuyuki Mori, Director, Head of Insurance, Southeast Asia at ABeam Consulting. “If the training data is biased, AI may only work well in very limited situations.”

Strong policies in data privacy are needed, Mori stresses. “Internal, strong policies and clear escalation protocols are needed.”

Underwriters empowered

AI has empowered underwriters and reshaped risk assessments for underwriters, enabling a shift from traditional static underwriting models to dynamic, data-driven approaches.

“Underwriters used to have to make all their decisions based on data,” says Dennis Lee, Principal, Insurance Solutions, ASEAN at Salesforce. “With AI, they are not just making decisions — they’re picking the best decisions. I think that’s really exciting.”

AI has helped underwriters track pharmaceutical developments across multiple jurisdictions simultaneously, says Frank Ahedo, CEO of Further Underwriting International.

“We now have the ability, because of technology, because of AI in particular, to be able to say, what is the FDA in the US approving, what is the EMA in Europe approving, etc, etc, what is available in the markets that we’re actually offering these products,” Ahedo says.

It’s a crucial development, given how medical innovation has been accelerating, he adds. “The FDA was approving three to four targeted therapies in the early 2000s. Now the ballpark is 70 and we would expect that to be 100 a year by the end of the decade.”

And then there’s how AI is simplifying routine work. “We spend an enormous amount of money on medical translations. We spend an enormous amount of money preparing clinical summaries — this is low-hanging fruit,” Ahedo says.

These innovations have also allowed insurers to move towards more flexible, responsive products in the life and health space.

“There’s a big proposition here and we have more flexibility to adjust dynamically,” says Soeren Kruse, Director for Life & Health in Southern & Eastern Asia at Hannover Re. “I would actually welcome a trend where product designs are becoming more sustainable or more need-based.”

The era of five-year rate guarantees is over, Ahedo adds. “Every insurer certainly wants stability. I think technically, that’s going to be a struggle. But the flip side to that is a much more dynamic, versatile product concept.”

Supervisor of the robots

Deliberate efforts to build a culture that embraces AI and positions artificial intelligence as an empowerment tool are also crucial for insurers as they seek to integrate AI into their processes.

“Creating a general culture is key for us to make this a success,” says Philippi. He outlined the insurer’s approach to building this culture—setting up town halls that reassured employees that AI would not replace them; instituting lunch and learning sessions, and creating curricula. “[We] talked to them about all the good aspects,” Philippi says. “We gave access to LinkedIn Learning for people to do a lot of courses on genAI.”

And this cultural change also translates to a practical role transformation for underwriters, says Jeffrey Sheng, Head of IT, Asia Pacific at Sompo Holdings.

Sompo has an AI committee that incorporates data, compliance, risk, and business functions, and conducts training for its employees.

“It’s a change of their roles, a change of their duties. But they are transformed from operator, from a predictor, to a supervisor — a supervisor of the robots,” Sheng says.