Why Insurance Innovation Takes More Than a Great Pilot
October 22, 2026
A shiny AI demo is easy. Getting it to survive contact with a real insurance business is much harder. In part one of this episode of The Reality Layer, Dave Tobias sits down with Lynn Thompson, Partner and Global Head of Strategy and Partnerships at QBE Ventures, to explore why so many pilots stall, what it takes to build a real pathway to production, and why the strongest results come from solving the right problem with the right team behind it.
In part one of this two-part conversation on The Reality Layer, Dave Tobias sits down with Lynn Thompson, Partner and Global Head of Strategy and Partnerships at QBE Ventures, to explore what it takes to turn emerging technology into meaningful insurance outcomes.
Lynn argues that innovation has to start with a real problem, not speculation about what technology might make possible. She shares how QBE has used data, imagery, and AI to help frontline teams identify potential coverage gaps and have better conversations with customers, while earning the trust required to scale those capabilities more broadly.
This conversation also tackles what Lynn calls the “pilot to nowhere” problem. Pilots can launch quickly, but real impact requires a deliberate path to production, clear ownership, strong executive support, and a willingness to iterate several times before value appears.
Dave and Lynn also explore how frontier AI models are changing what insurers can do with large volumes of data, why regulated workflows must remain predictable, repeatable, auditable, and governed, and why the right partner often matters more than the flashiest technology.
Ultimately, Lynn’s message is simple: start with the problem worth solving, build for scale from day one, and choose teams with the judgment and domain knowledge to keep adapting as the technology changes.
What You’ll Learn:
- How data and imagery can help insurers identify coverage gaps and support better customer conversations
- Why underinsurance requires more than simply extending today’s insurance models
- Why innovation projects need strong conviction, executive support, and a clear definition of success
- What Lynn means by the “pilot to nowhere” problem and why so many POCs never reach production
- How current AI models can help insurers process far more data across underwriting and portfolio decisions
- Why regulated AI workflows need to be predictable, repeatable, auditable, and governed
- Why specialized InsurTech partners still matter even as general-purpose AI becomes more capable
- How choosing the right team can matter more than choosing the most impressive technology
- Why deep domain expertise is becoming even more valuable as the cost of building technology falls
Lynn Thompson is Partner and Global Head of Strategy and Partnerships at QBE Ventures, the venture investment arm of QBE Insurance Group, where she leads strategic investments and partnerships with emerging technology companies.
Her career sits at the intersection of large organizations, startups, data, and innovation. Before joining QBE Ventures, Lynn led go-to-market at Data Republic and worked on an AI R&D hub in Singapore. She began her career in management consulting at Deloitte and has worked across the APAC region.
Lynn is also an Advisory Board Member at The Spark, a prevention-focused insurance lab bringing insurers, reinsurers, and technology companies together around major industry challenges, and a Board Member at iluminr, a capability intelligence platform focused on organizational resilience.
If you enjoyed this episode, subscribe, rate, and review it on Apple Podcasts, Spotify, and YouTube Podcasts. Instructions on how to do so are here.
Episode Resources: