How a 150-year-old insurance company is moving into the AI era

MetLife has been around since 1868. That means the company was operating when electricity was transforming business. Today, the company finds itself in the midst of an analogous shift, this time to AI. Insurance companies are famous for legacy software stacks. The question is how does the CIO push the company into the new era.

I spoke to global CIO Nick Nadgauda last spring about that shift. He had come in the prior year after a two-decade career at Citi, which culminated with him being CIO of the services division. But being a department CIO and a global one is a different level of responsibility, especially coming at such a key time in the company's tech history. Yet he saw similarities between the two positions.

"I faced very similar challenges pushing modern technology, trying to become more technical as an organization, trying to scale out," Nadgauda told FastForward. He says MetLife had begun the modernization process before he took over, and put its money where its mouth is by setting aside funding targeted specifically at updating the company's tech stack.

"The company has established dedicated 'reinvest' funding to support strategic initiatives across the enterprise, including efforts to modernize technology, strengthen data capabilities and deliver long-term business value," he said. It is in year three of a five year modernization effort. 

We talked about what it takes to keep that going and how an insurance company can use AI strategically, while recognizing when humans are absolutely necessary.

Modernizing journey

Even before the ChatGPT moment at the end of 2022, and before Nadgauda came on board, the company was already working with machine learning and automation with an internal tool they call MetIQ, which was helping modernize the company's approach to things like claims processing, call center and the overall data strategy. The maturity of LLMs expedited what they were doing.

"The MetIQ platform started out as something that was more traditional AI, but we were already in this reengineering mindset, and then fortunately, just as we were reengineering, generative AI kicked off, and we were able to inject some of those things into the MetIQ platform and into the business process reengineering efforts," Nadgauda said.

That allowed the company to have a head start at a time when many companies, especially those outside the tech sector, were starting from scratch when it came to incorporating AI into their operations.

He says by paying attention to the processes, which is a message I hear a lot, you can squeeze out efficiency and begin to look at ways the technology can pay dividends.

Like many organizations, his company is seeing some success with AI in programming, particularly around maintenance -- things like fixing bugs and making sure libraries are up-to-date. "We've moved some of our development process to be AI-first, so for things like currency (keeping software updated) and defect detection. AI is the first step in the process," he said.

When it comes to updating code, he says it's moving from writing traditional software languages to describing what you want in plain English, a process that comes with its own set of trade-offs. "Some of these are multi-page prompts, which is not too far off from the amount of time you have to spend writing a piece of the code, or thinking about how you would structure it," he said.

AI trade-offs

One of the challenges Nadgauda faces in his job is what needs to be modernized. There aren't unlimited budgets and executives have to pick and choose what they update and what they leave in place or put off for now.

For instance, he doesn't necessarily have a problem with legacy COBOL programs, which an insurance company of a certain age like MetLife is still going to have on the books. He looks at it from a maintainability perspective, rather than a strict modernization one.

"I don’t necessarily have a problem with COBOL. The real problem in my mind is when you’ve got systems that you can’t support, where you can’t find programmers, you can’t find talent or they’re hard to modify."

But it's more than simple modernization. He also worries about companies missing the true value of AI, and it's not what you might think. "My view, and I feel pretty strongly about this, is that we're on the cusp of a bunch of new business models," he said. "I think as that happens, people need to focus a little bit less on cost takeout and reducing headcount and more on the positives like being able to do new things, and deploy them more quickly."

In the end, this is an insurance company and when it comes to customer-facing situations, that could involve a tough time like filing a life insurance policy claim of a loved one who just died or dealing with the aftermath of a car accident. These are the kinds of things where you have to know when to automate and when to have the human touch.

"I don't see us ever being in a position where you're filing a super important claim where you're talking just to an AI. There'll be somebody there who is trained, empathetic, and can meet you where you're at, and that's never going to change," he said.