AI in claims: start with the problem, not the technology
By Rob Stewart - 25th September 2026
At the CILA Conference this week, I joined a panel discussing the role of AI in claims and loss adjusting, and where it might take the profession over the next few years.
One theme kept coming up for me: rather than starting by asking where we can use AI, we should start by asking where the friction is.
The people closest to a claim usually know exactly where that friction sits: repetitive administration, rekeying information, searching through lengthy files, reviewing documentation, completing QA or simply working out what needs to happen next. Those are the areas where I think AI has the most to offer.
For straightforward, transactional claims, there is real potential for technology to drive more of the process. With complex losses its role is different, helping professionals digest information, identify issues and make better-informed decisions. A phrase I used during the discussion to sum this up was “automate certainty, assist ambiguity”.
We saw a good example of this in our own operation recently. Our handlers work with Haste, an AI agent that runs within Synergy Cloud, the claims platform owned and developed by Claims Consortium Group. On a travel claim where thunderstorms had caused a missed connection, the handler asked Haste to draft a response explaining to the insured why her flight refund had been deducted from her additional costs. Haste pulled the payment records and claim notes and confirmed the figures behind the settlement, including the refund, the excess and the balance. It also noticed that the underlying costs had been logged in US dollars while the payment had been made in pounds, and flagged that the conversion hadn’t been explicitly recorded rather than assuming it was correct. That gave the handler a clear, customer-friendly explanation to work from, along with a specific point to check before anything was sent.
AI can recommend an outcome, draft correspondence or flag something a handler should look into. But when a decision materially affects a customer, particularly around coverage, repudiation, vulnerability or other grey areas, accountability still sits with the person and the organisation making that decision. The FCA’s Consumer Duty requires firms to put their customers’ needs first¹ and to act to deliver good outcomes for retail customers², and that obligation doesn’t pass to a piece of technology. We should never reach a point where the explanation is simply “the AI declined the claim”.
Used well, AI gives us the opportunity to make claims quicker, more consistent and better controlled, which is better for claims professionals and customers alike.
The technology will keep changing quickly. The challenge for our profession is to stay curious about what it can do, while holding on to the judgement and expertise that make it worth using in the first place.
