农夫导航

Are two AIGs better than one?

In outlining the work at American International Group (AIG) to integrate generative artificial intelligence, CEO Peter Zaffino said AIG’s intent is to “create a digital twin of our business.”

The “twin” will represent “all key data, processes, business logic and a map of relationships across businesses and functions,” Zaffino said during a conference call with analysts to discuss earnings. The chief executive said ontology will be the foundation to technical reviews, provide claims personnel with insight and tools, enhance decision-making, and improve service to clients.

“You will hear a lot more on this topic in the coming quarters, especially as we make more progress with our rollout,” Zaffino said. “We’ve been building our AIG ontology since we began our work in AI.”

In philosophy, ontology is the study of being鈥攖he nature of existence. In business, ontology is a look at what a business or technology needs and how those things interact with one another. Ontology in AI provides concepts to help AI understand the data it is processing and leverage it to produce operational results.

Zaffino during the call referenced AIG’s partnerships with Palantir, Anthropic, and AWS. As it is explained in a , ontology is about making decisions within an enterprise. AI interacts with ontology.

“Traditional data architectures do not capture the reasoning that goes into decision-making or the action that results, and therefore limit learning and the incorporation of AI,” wrote the company’s chief architect. “Conventional analytic architectures do not contextualize computation within lived reality, and therefore remain disconnected from operations. To navigate and win in today’s world, the modern enterprise needs a decision-centric software architecture.”

The concept of a digital twin allows for tests or modeling of potential operational decisions. Results can be predicted and analyzed through the digital twin.

“Our ontology will create a clear record of any actions taken, which will inform business logic and provide the ability to audit agents’ activities,” Zaffino added.

In terms of its use of GenAI, Zaffino said the insurer started with underwriter assistance for nonprofit business products at the end of the first quarter, and “early results are very promising.” He said AIG Underwriter Assistance is on track to be used for middle market accounts at Lexington 农夫导航 Company in the third quarter and across other lines of business such as all of North America commercial in 2026.

AIG spoke about its use of AI and AIG Underwriter Assistance during its Investor Day event earlier this year. Read .

“We continue to explore opportunities to accelerate our rollout,” said Zaffino during the earnings call, adding that it has harnessed GenAI to support claims. Initial tests on the notice-of-loss process has shown a decrease in processing time from days to hours.

“We’ve been training large language models to extract and organize key insights automatically to enable claims adjusters to make more informed decisions faster than ever in order to fulfill our promise of helping clients when they need us most,” Zaffino said.

Photo: Generated with AI, AdobeStock

Topics AIG