Millennial Specialty Insurance (MSI) is overhauling how it builds and monitors insurance products as AI models and new property data sources introduce fresh risk-management challenges. Andrew Dalton, the company's chief product officer, said the carrier is focused on keeping AI systems accurate over time while using them to speed up product design.
Dalton said the analytics and modeling space is changing quickly, and MSI is bringing in new data sources while keeping close oversight of how those inputs affect underwriting decisions. The central concern, he said, is managing the risks that come with relying on AI and handling property data at scale.
AI drift demands constant monitoring
A key operational challenge is "AI drift," the tendency for models to lose accuracy as real-world patterns shift. Dalton said MSI responds with continuous monitoring and testing of its systems to confirm they stay reliable as conditions change.
"We have to monitor and test these systems constantly," Dalton said, emphasizing that validation is not a one-time step but an ongoing part of the product lifecycle. That discipline applies to both the models MSI builds and the third-party tools it adopts.
Anthropic partnership compresses design timelines
MSI's partnership with Anthropic has changed the pace of product development. Dalton said work that once took months now gets done in days with AI assistance. The speed gain, however, does not remove the need for human judgment.
"Human expertise remains crucial for interpreting data and evaluating model outputs," he said. That distinction matters for insurance teams deciding where AI can replace manual work and where it should only augment it. For professionals exploring how these tools apply to their own operations, AI for Insurance training can clarify which tasks are safe to automate.
Data relationships remain the edge
Dalton expects AI to become a competitive differentiator in insurance, but not because the models themselves will be unique. He said modeling capabilities will become more commoditized over time, meaning the winners will be companies with strong domain expertise and the ability to use AI tools effectively.
He pointed to unique data relationships and distribution partnerships as lasting advantages that competitors cannot easily copy. That view suggests insurers should invest in proprietary data sources and partnerships rather than assuming off-the-shelf AI alone will separate them from rivals. Executives weighing these strategic trade-offs may find AI for Executives & Strategy resources useful for structuring the decision process.
Why this matters for insurance professionals
For insurance professionals, the practical takeaway is that AI adoption is not a one-time implementation project. MSI's approach points to a workflow where models are continuously tested for drift, human reviewers evaluate outputs, and proprietary data relationships create the real competitive moat. Teams that build those monitoring and review processes now will be positioned to move faster than competitors who treat AI as a set-and-forget tool.
Your membership also unlocks:
