Summary

AI maturity is rising across the insurance industry, and the competitive landscape is shifting fast. As once-differentiating capabilities become more widespread, insurers face growing pressure to move beyond efficiency gains and scale AI in ways that improve core outcomes across underwriting, pricing, and claims.

To mark the launch of the 2026 Evident AI Index for Insurance, this session brought together Akhil Lalwani (Allianz UK), Andreas Bayerstadler (Munich Re), Tony Marron (Liberty Mutual Insurance), and Alexandra Mousavizadeh (Evident) to unpack the latest results.

Watch the roundtable

Meet the Speakers

Key Discussion Topics

1. Culture and strategy come before technology
AI should accelerate business strategy, not define it. That requires a culture of experimentation, a quality (not just efficiency) lens, and close collaboration with domain experts. Technology alone doesn't drive transformation without the operational and cultural foundations in place.

2. Talent is shifting from building foundations to scaling deployment
Insurers are democratizing AI access via internal GPT tools and large-scale training rather than restricting it to specialists. The consensus: the gap isn't building AI, it's adoption, which depends on trust and bringing the whole workforce along. One flagged gap: a missing "business translator" role bridging technical and business expertise.

3. Cost and pricing uncertainty are driving a flexible build-vs-buy strategy
Vendor pricing is seen as a "future cost risk," with a shift toward demand/consumption-based models. This is pushing insurers toward model-agnostic flexibility rather than committing to one stack. Regulatory uncertainty adds to the case for flexibility. There was also agreement that models are becoming commoditized - the real value lies in proprietary data and context.

4. Governance and trust are the real bottlenecks, not technology
Responsible AI frameworks (human-in-the-loop, privacy, accountability) were framed as a "necessity to earn the license to continue to innovate." Governance is working if it keeps evolving, not if it's static; fail-fast experimentation is expected, with timing being key ("things that don't work today may work in six months").

5. The next frontier is end-to-end transformation, not point solutions
The shared call was to move "from automation to transformation", which requires technical, operational, and cultural change together, and genuine willingness to redesign whole processes rather than bolting AI onto a single step.