Artificial intelligence is changing the rules of engagement in insurance underwriting – but not everyone is ready to follow the algorithm.
From algorithmic smart-follow MGAs to actuarial modelling tools, UK insurers are increasingly exploring ways to embed AI into operations.
Aviva has become the latest major carrier to trial AI within its Global, Corporate & Specialty (GCS) division. The insurer is testing an AI-powered “Actuarial Agent” through a six-month pilot with insurtech partner hyperexponential (hx). The goal: improve pricing accuracy and reduce friction in workflows.
Karen Dayal, chief underwriting officer for commercial lines at Aviva, said the tool will “augment the expertise of our underwriters and brokers” – not replace it. Aviva has already used hx’s Renew platform to develop 20 pricing models in nine months, slashing development time from over an hour to less than 10 minutes.
Still, industry sentiment is wary. Nearly 70% of underwriters and 67% of actuaries believe AI could displace their roles within five years, according to recent survey data. And four out of five actuaries say they lack the technical skills to work confidently alongside such tools.
Amid this backdrop, new ventures are betting on the benefits. Augmented UW Ltd, a newly launched managing general agent, is targeting the London Market’s smart-follow segment using AI-powered underwriting.
Founded by Daniel Prince, former CEO of Rethink Underwriting, the MGA aims to streamline follow-capacity placement through algorithmic models. Its launch partner, Artificial Labs, will provide the digital infrastructure, with the first risk placements expected later this year.
“The London Market is not broken,” said Prince. “It needs a steady hand to implement new technologies, such as AI, to enhance the way insurance operates.”
The Lloyd’s Market Association has highlighted enhanced underwriting and data-driven decision-making as key market priorities through 2030 – a shift Augmented hopes to be part of.
Despite growing investment, many brokers remain unconvinced about AI’s role in client-facing advice.
While others may test hybrid systems that support, but don’t replace broker judgement, trust remains a decisive factor. Concerns about model “hallucinations” and misinterpretation continue to weigh on frontline professionals.
Even some insurers are preparing for the risks AI introduces. Earlier this year, Chaucer partnered with Armilla AI to develop a first-of-its-kind insurance product covering AI model failures, including drift and inaccuracy.
For now, AI adoption in insurance looks set to accelerate – but on carefully managed terms. Carriers are expanding internal capabilities, MGAs are pushing the envelope, and underwriters are adjusting to new tools.
But the industry’s defining currency remains trust – something not easily replaced by code.
