SBS Insurance Services has completed a pilot project with UnlikelyAI, testing a new AI-powered platform aimed at improving claims processing.
The trial automated 40% of claims with reported accuracy of 99% and included a full audit trail for every decision, according to the companies.
The project marks UnlikelyAI’s first public deployment following several years of development. The company’s system, which combines neural networks with symbolic reasoning, is positioned as an alternative to large language models (LLMs), which have faced scrutiny in regulated sectors due to challenges around transparency and decision consistency.
SBS selected UnlikelyAI’s platform to address some of these limitations. Unlike LLMs, which the company claims can deliver inconsistent outcomes and lack explainable decision-making, UnlikelyAI’s system is designed to follow rule-based logic. This, the firms say, makes it more suitable for regulated environments such as insurance.
During the pilot, the platform processed claims with minimal manual intervention, while deferring more complex decisions to human staff. SBS reported that the system helped reduce miscategorisation rates by 99% and improved overall claims handling efficiency. According to the companies, the platform’s decision-making process can be traced in full, aligning with regulatory requirements for transparency.
William Tunstall-Pedoe, CEO of UnlikelyAI, said the trial demonstrated that AI can be used in a more transparent way: “This deployment… shows that AI doesn’t have to be a black box. It can be transparent, explainable, and accountable.”
Bazil Crowley, director of Innovation at SBS Insurance Services, said the new system supported the company’s efforts to streamline operations while maintaining service levels: “We’re excited to lead the way for this milestone in InsurTech and are proud to be setting new industry standards for intelligent claims processing.”
What do you think of the use of AI in claims handling? Please tell us your thoughts below
