Travelers Insurance has built its own proprietary large language model, TravelersLLM, to reduce AI costs and improve performance on insurance-specific queries, while still using frontier models for broader tasks. The company generated $49 billion in revenues in 2025 and employs 30,000 people. TravelersLLM was unveiled in June 2026 and delivers better results than commercially available AI models for insurance-related questions, according to the company.
The move comes as AI cost management has become a primary concern for executives. Some companies are adding flexibility to model selection and building their own models to mitigate costs. Mapping tasks to an AI model is a trend across industries to address the high cost of always running queries through the most advanced AI models.
Why Travelers built its own model
Mojgan Lefebvre, EVP and Chief Technology and Operations Officer at Travelers, said the company keeps a close eye on spending. "In general, whether AI or any technology investment, we have a very keen focus on ROI," Lefebvre said. TravelersLLM is cheaper to run than frontier models, and it works alongside them, not as a replacement. Queries are directed to either TravelersLLM or a frontier model depending on the task.
"The approach to the TravelersLLM is, if you don't need to be using the most expensive frontier model to answer the same question, we shouldn't be. The TravelersLLM directly contributes to this ROI-mindedness and saying, 'Use the right model for the right problem,'" Lefebvre said.
TravelersLLM filled a gap that general purpose models couldn't, Lefebvre claimed. The model was trained on millions of company documents and evaluated against tens of thousands of insurance domain questions. It was built with subject matter experts in underwriting, claims, operations, and service management. The model makes institutional knowledge of domain experts available to employees.
Cost savings and strategic flexibility
Using a variety of models reduces dependence on any single external model or vendor, Lefebvre said. "It absolutely improves the economics of using AI at scale for us across the enterprise, and it gives us strategic flexibility," she said. The company runs 70% of its compute in the cloud, a foundation that supports its AI ambitions.
Applications access models, including TravelersLLM, through APIs. For most employees, the underlying models are largely invisible. "For most employees, the underlying models should largely be invisible, and they shouldn't have to think about it," Lefebvre said. "They should simply have the best intelligence available at the point of decision."
Travelers began modernizing its technology foundation more than 10 years ago. The company still operates some legacy systems; modernization was intentional, not comprehensive. That long-term investment enabled the current AI push. "There's no way on Earth that we could be doing what we're doing with AI today if we hadn't been making the investments in modernization and focusing on data and connectivity of our data, its accessibility and quality," Lefebvre said.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
A broader industry shift
Travelers is not alone in tackling AI costs. Snowflake debuted an AI cost management feature this week that dynamically routes tasks to appropriate AI models based on cost and quality. AWS and Oracle implemented AI cost management features and tools this year. The pattern is clear: companies want to avoid paying premium prices for every query.
Data accessibility, usability, semantic layer, ontology, and knowledge graph are critical to connecting data to AI systems, according to the article. TravelersLLM will serve as the foundational capability for agentic AI, the company said.
What comes next
AI has evolved from a tool driving individual productivity to being embedded across company workflows, Lefebvre said. The next step is bigger. "The next frontier is that true transformation where, with AI, you absolutely can start thinking about new business models, products and services," she said.
TravelersLLM delivers better results than commercially available AI models for insurance-related questions, Lefebvre claimed. The model is cheaper to run than frontier models. The company's approach is part of a broader industry trend of using multiple models and routing tasks based on cost and quality.
The article was published on Aug. 24, 2026, by CIO Dive, written by senior news writer Makenzie Holland. It is 4 minutes long in audio format. The article includes a photo of a Travelers Insurance office in Hartford, Connecticut, credited to JHVEPhoto via Getty Images.
Travelers' long-term investment in data modernization enabled its AI capabilities, the article notes. The company's proprietary model is a direct response to rising AI costs. The strategy pairs domain-specific performance with cost discipline.
The article also references a recommended reading: "Surprise AI costs threaten enterprise implementations" by Scarlett Evans, published Aug. 6, 2026.