Travelers' CTO shifts to fewer, bigger AI bets as insurers demand clearer returns
Mojgan Lefebvre, chief technology and operations officer at Travelers Companies, has stopped chasing dozens of AI pilots across the insurance giant. Instead, she's concentrating spending on a handful of high-impact projects that can scale across the company's 30,000-plus employees.
The shift reflects a broader change among technology leaders in the insurance industry. After the initial enthusiasm following ChatGPT's launch in late 2022, CTOs and CIOs are moving away from what Lefebvre calls the "let a thousand flowers bloom" approach-allowing hundreds of small AI experiments to proliferate. Returns on those pilots have been difficult to measure and justify.
"I don't think a thousand little things will add up," Lefebvre said.
Where Travelers is concentrating its efforts
Travelers has narrowed its AI strategy to three areas: improving claims processing, enhancing service management, and strengthening data and analytics capabilities.
In January 2026, the company announced a partnership with Anthropic to give nearly 10,000 engineers, data scientists, analysts, and product owners access to personalized AI assistants for software development and model building. A month later, Travelers launched its AI Claim Assistant, developed with OpenAI, which handles customer questions about claim submissions in natural language.
The claim assistant is already showing adoption. When policyholders report accidents or incidents-known as first notice of loss-roughly 50 percent now file digitally through the Travelers app. Of the remaining half who call, most interact with the AI system. Consumers are showing "tremendous acceptance" of the technology, Lefebvre said.
Measuring what matters
Lefebvre tracks three categories of success. For claims, the company measures how quickly it closes cases. Across other applications, it monitors cost savings and efficiency gains. A third metric is adoption and employee "empowerment"-whether workers are embracing new ways of working rather than resisting the technology.
Without measurement, benefits disappear, Lefebvre said. "Anything that you don't measure can evaporate. You absolutely have to have some commitments in your budgets and your plans."
Choosing partners strategically
Travelers works with both Anthropic and OpenAI rather than consolidating with a single vendor. Lefebvre views OpenAI as strongest in conversational AI, while Anthropic leads in analytical and coding capabilities.
"It's too early in the AI journey to do everything with one," she said. "There are certainly other players, but you also don't want to have ten different partners."
From skepticism to adoption
Travelers launched TravAI, an internal platform integrating multiple generative AI tools with company systems, after ChatGPT's debut. Every employee can access it after completing training. The company intentionally framed the technology as a productivity tool rather than a threat.
"It's taken away that fear that AI is here to take away my job," Lefebvre said.
For insurance professionals managing claims, underwriting, or risk assessment, understanding how major carriers are implementing AI can inform strategy at your own organization. AI for Insurance covers how the industry is applying these technologies in practice.
Technology leaders overseeing AI strategy may also find value in the AI Learning Path for CTOs, which covers enterprise deployment decisions and vendor selection.
Your membership also unlocks:
