The State of AI in Insurance: Which Use Cases Are Delivering Real Value?

Watch the Roundtable

Summary

The pace of AI adoption in insurance is accelerating, and so is the pressure to show what it's worth.

Drawing on exclusive data from the Evident Use Case Tracker, this roundtable brought together Alex Baldenko (Head of R&D at MassMutual), Jodie Wallis (Global Chief AI Officer at Manulife) and Matt Gorman (SVP & Chief AI Officer at Travelers) to discuss the latest trends shaping insurers’ AI strategy in 2026 and beyond.

Key Discussion Topics

Prioritize structural transformation over incremental gains: Insurers are moving beyond incremental point solutions, such as internal productivity tools, toward structural transformation by prioritizing the major themes that create the greatest value. This process is driven by top-down conviction from senior business leaders using a VC-like due diligence. Manulife, for example, is focusing resources on several major themes, including: underwriting, distribution, and virtual assistance to achieve its goal of $1B in AI:ROI.

Scaling requires foundational "plumbing" and governance: The shift to core insurance processes in areas like underwriting and claims, requires significant foundational infrastructure, or "plumbing," which includes data modernization, robust governance layers, and building trust with stakeholders and regulators. Moving to advanced agentic AI systems for core functions is a fundamentally different conversation than rolling out basic GenAI productivity tools, which in turn, requires reimagining how the business works in the long-run.

Scaling thresholds must be designed upfront: The decision to scale a pilot depends entirely on the specific use case, especially for customer-facing solutions where individual behavior cannot be controlled. Before building, organizations must define and design the scaling requirements, including a high threshold for accuracy and critical infrastructure components. This infrastructure includes real-time consistency checks, failovers, and kill switches, which must be “baked in” from the start.

Measure value beyond productivity: While productivity gains dominate reported outcomes (75% of disclosed use cases), this metric often understates the true financial value captured, as revenue uplift and customer satisfaction are more difficult to report publicly. True financial value should only be counted based on cost efficiency (total or unit cost reduction). Productivity gains are better treated as a "gateway" to engage the workforce in AI adoption. To tell the full value story, senior business leaders must define measurable business strategy objectives that AI can help accelerate.

Adopt an agnostic "build now, buy later" strategy: Given the rapid changes in foundation model capabilities, insurers are building on a model-agnostic foundation to avoid the risk of vendor lock-in. This "build now, buy later" strategy involves building non-differentiating capabilities that are not yet available from enterprise software vendors. In turn, this requires a mindset of building solutions with a comparatively short shelf life (12–18 months) that are planned to be replaced as new capabilities customized to the needs of the industry become available over time.