Where AI agents meet underwriting slips

Last night’s discussions at Claims 2.0, hosted by InsTech, shed light on how digital platforms are enabling the London market to become more agile and data driven.

Technology is shaking up the London insurance market. The industry features a labyrinthine value chain and a patchwork of legacy systems. Working practices have been slow to evolve, particularly outside of retail markets like home, travel and auto insurance.

Last night’s discussions at Claims 2.0, hosted by InsTech, shed light on how platforms like Guidewire and Reserv are enabling the London market to become more agile and data driven, albeit still in pockets.

At the same time, face-to-face trading is very much in evidence, with many underwriters still passing paper slips across the floor of Lloyd’s of London.

Legacy IT, and the perceived risks of technology change, are ongoing barriers. I heard one story about a mission-critical server, through which billions of value flowed, that was (is still?) hosted ‘in a cupboard’, with staff resorting to spraying the building’s roof with water in a bid to prevent the server from igniting during a heatwave.

The risk of business disruption is becoming more manageable, with AI now enabling large data migrations to be completed within weeks rather than months. Everyone agrees AI will transform business processes, enabling providers to simultaneously deliver accuracy, efficiency and customer service, goals that previously needed to be traded off.

That said, the human element is still the decisive factor, highlighting the importance of effective change management. For example, how can providers inject data and AI into workflows in a way that loss adjusters are willing and able to use?

I took part in an interesting discussion about what it will take to leverage multi-agent systems to automate claims processing. Given the number of parties involved in the value chain, agents are going to need the ability to establish trust, and share trusted data, across organizational and geographic boundaries. I’m embarking on a new project investigating how this can be achieved - message me or connect if you’d like to explore use cases for AI agents operating across trust boundaries in insurance, financial services, healthcare, manufacturing or government.