Why we're bringing the claims engine to insurance.
Record claims volumes are colliding with a shrinking, ageing workforce — and the market has already settled on the model it wants: AI with a human in the loop. That's the engine we've spent two years building.
We've just been named 2026 North West Technology StartUp of the Year and 2026 North West Overall StartUp of the Year — recognition earned in mass litigation and collections, where we automate the document- and process-heavy work behind millions of claims. Insurance claims operations have the same shape: high volume, document-heavy, compliance-bound, and under pressure on both cost and headcount. So we're bringing the same engine to it.
The squeeze is real, and it's structural
UK motor insurers paid out a record £11.7bn across 2.4 million claims in 2024, with the average claim up 13% (Association of British Insurers). Property payouts hit a record £6.1bn in 2025 (ABI). Claims inflation has run well ahead of general inflation for years. None of this is a spike to wait out — it's the new baseline, and it lands first on the claims function.
And the people to handle it are leaving
Roughly one in five claims adjusters leaves the profession each year, and each departure takes about six years of experience with them (Deloitte). Around a quarter of the insurance workforce is expected to retire within a decade (Chartered Insurance Institute), and 72% of insurers say labour shortages are already hurting claims efficiency (Gallagher Bassett). You cannot hire your way out of a structural talent gap while volumes climb. The only durable answer is to make each handler dramatically more productive.
The market has already chosen the model
Three quarters of UK financial-services firms now use AI, and insurance is the highest-adoption sector at around 95% (Bank of England / FCA, 2024). But the same survey is precise about how: 55% of AI use involves some automated decision-making, and only 2% is fully autonomous. The market doesn't want a black box — it wants automation that does the repetitive work and keeps a person on the judgement calls.
That 55% / 2% split is the whole argument. The industry has decided it wants AI with a human in the loop. That is exactly how Saber was designed — not as a hedge, but as the operating model.
Where Saber fits
Saber sits above the systems insurers already run — case-management like Proclaim, policy and claims cores like ACE — rather than replacing them. It reads the correspondence and documents, extracts and classifies, runs the routine steps, and surfaces the exceptions with the evidence attached for a handler to approve. Deployments reach value in weeks, not the multi-month timeline of a core replacement. On the claimant side, the $2bn valuation of EvenUp and the rise of Supio prove the thesis at scale; what those tools do for legal demand packages, Saber does for the insurer's own high-volume claims and correspondence — a track record most overlay vendors can't point to.
Built for how insurers actually buy
Claims leaders don't buy on capability alone; they buy on whether it survives their risk and compliance review. Saber is built for that conversation: every action is logged for 100% QA and a complete audit trail, which speaks directly to Consumer Duty and ICOBS expectations on prompt, fair claims handling. The human-in-the-loop design aligns with UK GDPR rights around automated decisions, and we operate to the security baselines enterprise buyers expect — with a simple commitment that your data is never used to train our models. Regulation is cited by over half of UK insurers as a brake on AI adoption; our model is designed to release that brake, not test it.
Press enquiries
Interviews, commentary, data or assets — tell us what you need and we'll come back within one business day.