In this episode, we revisit a live panel discussion from last year’s Agentic AI event, moderated by Robin Merttens featuring Ian Thompson of IMT Advisory, Sasha Haco, Co-founder and CEO of Unitary, Dr Paul Dongha from NatWest Group, and Nick Williams-Walker, Group COO at McGill and Partners.  

Recorded at a moment when Agentic AI was beginning to dominate conversations across insurance, the discussion explores what the industry might look like three to five years into widespread AI adoption — and what it will take to get there. 

The panel examines where insurers are already deploying AI agents, from claims investigations and underwriting support to real-time risk analysis and customer servicing. But alongside the optimism sits a more cautious conversation around governance, regulation, data complexity and the growing gap between experimentation and operational reality. 

Throughout the discussion, there is a recurring tension between ambition and execution. Some panellists argue the winners will be organisations willing to move quickly with focused, incremental deployments. Others warn that poorly governed AI systems, unrealistic expectations and “fear of missing out” decision-making could create entirely new risks for the sector. 

The conversation also raises broader questions about how insurance itself may evolve as AI capabilities mature — from the rise of highly automated brokers and MGAs to changing expectations around service, talent and human expertise. 

With InsTech and AI Risk returning on 7 July for this year’s expanded Agentic AI event in London, this episode offers a useful opportunity to look back at the predictions, concerns and opportunities shaping the conversation just one year ago. 

In this conversation, the panel discusses: 

Sign up to the InsTech newsletter for a fresh view on the world every Wednesday morning.