’You don’t need to run a six-month project to prove a specific AI use case or a specific AI agent,’ says group chief AI officer

The insurance industry should not be afraid to fail when trying out artificial intelligence (AI).

This is according to Paul O’Brien, group chief AI officer at Davies Group, who told delegates at the Shaping the Future of Insurance conference that Davies has implemented an agentic model to deliver optimisations in the claims process including advisory agents, actioning agents and governing agents.

The conference followed the Which? super complaint from 23 September 2025, which flagged poor claims handling in the home and travel insurance markets.

O’Brien explained that stakeholders should map an end-to-end process with AI agents, saying they should “pilot quickly and [not] be afraid to fail fast”.

He said: “You don’t need to run a six-month project to prove a specific AI use case or a specific AI agent.

“Spin it up fast, prove it works. An AI failure right now [shows] AI can’t do that yet so you don’t throw it in a bin, [but instead] put it on a shelf and then come back to that later. That’s the approach we’re taking to implementing AI throughout our business.”

Keeping pace

Looking to advance the claims process, one example of a project Davies will be revisiting in the future, O’Brien told delegates, is the ability to process images to validate surveyor reports using multimodal AI.

O’Brien explained that the group concept was picked up 18 months ago but had to be shelved as the “capabilities of the image recognition weren’t mature enough, the cost was still high [and it was] still early [days]”.

For smaller brokers without the resources to start the AI journey, he added: “There’s no harm at all in being a fast follower in this space at the moment.

“You don’t have to be the first people to have agentic AI, you can see where the industry is going but my advice would be [to] establish strong partnerships – don’t necessarily try and build your teams internally.”

Insurance Times Fantasy Football