Inside the ClaimSorted $13.3m round — what actually changed
ClaimSorted just closed a $13.3m seed round founded in 2024 by Pavel Gertsberg and German Mikulski and operates as a tech-driven third-party administrator (TPA). The thesis is narrow: wrap AI around the slowest, most manual steps in the claims journey — intake, document verification, eligibility checks, triage, fraud flags — and keep skilled human adjusters for complex decisions.
The October 2025 seed (reported variously as $13.3m and €11.4m depending on FX timing) funds three things: expanding headcount, deepening the AI layer, and scaling into the US and EU alongside the UK. Management has publicly targeted a 25% cost-per-claim reduction within 12 months of deployment, measured via NPS uplift and unit economics. That's a sharper, more auditable KPI than most insurtech pitches and it's exactly the framing FCA supervisors respond to.
The broader signal matters more than the round size. Atomico leading at seed — typically a Series A/B fund — tells you serious capital is betting that TPA-as-a-software is a category, not a feature.
FCA's 2026 claims-handling crackdown: why AI just became the default
The FCA confirmed in early December 2025 that it will not introduce AI-specific rules; existing frameworks (Consumer Duty, SM&CR, operational resilience) are the mechanism. What changed is the supervisory lens. The FCA's 2026 insurance priorities put claims handling and service quality at the top, with home and travel claims singled out for poor consumer outcomes. Guidance on audit trails and human-in-the-loop protocols is expected in 2026, and the FCA's AI Lab and regulatory sandbox are open for firms testing new claims workflows.
Translation for operators: if your claims journey can't prove fairness, timeliness and explainability, your board should assume a skilled-person review is on the menu. AI isn't the risk — the status quo is. Manual claims queues, inconsistent adjudication, and weeks-long paper trails are what trigger Consumer Duty failures. Automation with the right audit layer is the defensible position.
The Aviva benchmark: £60m saved, 80 models, one playbook
Aviva is the public proof point nobody can dismiss. Across 2024–2025 the group deployed more than 80 AI models inside its claims function. Disclosed outcomes:
- Liability assessment time on complex motor claims cut by 23 days.
- Routing accuracy to the right team up 30%.
- Customer complaints down 65%.
- Motor claims transformation alone saved £60m+ in 2024.
That's not a pilot. That's a line item in group results. Smaller carriers and MGAs used to hide behind 'we don't have Aviva's data science team' — but that argument dies when a Y Combinator-backed TPA will run the stack for you and a Marketing OS will run the demand layer on top.
Your 5-step roadmap to turn AI claims automation into paid results
- Map your slowest claim type. Pick one line (motor FNOL, travel baggage, SME property) and measure current cycle time, touch count, and complaint rate. You need a baseline the board can argue about.
- Instrument the triage layer first. Document classification, fraud signals, eligibility checks and first-touch routing are where AI wins 60–80% of the time savings with the lowest regulatory surface area.
- Build the audit trail before the model. Every AI decision needs a reason code, a confidence score, and a human override route. This is what an FCA skilled-person review asks to see.
- Tie KPIs to Consumer Duty outcomes. Speed-to-payout, fairness of decision rates across segments, and complaint volume are the metrics that make AI spend defensible at both the audit committee and the supervisory meeting.
- Turn results into distribution. Once the cycle time drops, your marketing team has the single most valuable asset in insurance — a faster, fairer claims story that brokers, aggregators and corporate partners will actively promote.
What this means for insurance marketers and MGAs
Step 5 is where most insurers lose the round. You ship a faster claims experience and then describe it in a PDF. The brokers don't notice. The aggregators don't re-weight. Price-comparison rankings don't move. Your cost-per-quote stays flat while the product has demonstrably improved.
The 2026 playbook for insurance growth teams — run on a platform like Anjin — has three layers:
- Proof content: case studies with real cycle-time numbers, broker testimonials, and compliance-ready claims journey explainers. This is what wins panel placements and broker-of-choice decisions.
- Programmatic distribution: landing pages per line of business, per broker vertical, per region — matched to the exact queries your target partners search. ('ai claims handling uk', 'tpa with automation', 'fca consumer duty claims'.)
- Operational feedback loop: your marketing team should see the same cycle-time dashboard the claims ops team sees. When a metric moves, a campaign ships. When a complaint driver is fixed, a broker email goes out.
Doing all three with a traditional agency stack means five tools, two retainers and a slow internal approvals queue. You won't keep up with a ClaimSorted-powered competitor on that operating model.
Anjin: the Marketing Operating System for insurance growth teams
Anjin is the Marketing Operating System — a single platform where AI agents plan, produce, publish and measure marketing work that used to need a team of ten. For UK insurance teams in 2026, that means:
- Content agents that turn a single claims-data brief into 40 broker-ready pages, each tuned to a line of business and a real search query.
- Compliance-aware copy: Consumer Duty language, clear-and-fair disclosures, and plain-English claims explainers produced at pace without ping-ponging drafts to legal.
- Programmatic SEO for the long-tail queries MGAs, brokers and price-comparison partners type when they're deciding who to place business with.
- A measurement agent that ties organic pageviews, broker inbound, and quote volume to a single dashboard — so your team can argue for investment with numbers, not anecdotes.
Agencies will be the first to run Anjin at scale for their insurance clients, but the platform is the system — the agencies are the launch audience. In 2026 the winners will be the insurance brands running the whole marketing stack on one AI-native layer, not the ones paying three retainers to keep up.
The £888 Lifetime License — Offer Closing Soon
Lifetime access to Anjin for a one-time payment of £888. Not a subscription. Not a seat. Not a trial. One payment, unlimited use, for as long as Anjin exists.
The average marketing team spends £888 in about three working days on tooling, freelancers and coordination software. You're buying the platform that replaces most of it — once.
This price will not be offered again once we close our early-access cohort.
Claim your £888 Anjin lifetime license →Founders, agency owners and in-house marketers — this is how you run marketing at AI speed without the team, the burn, or another year of waiting.
Sources: Insurance Edge, Vestbee, Crunchbase, FintechGlobal, PwC UK 2026 insurance priorities, Pinsent Masons AI regulation guide, McKinsey Aviva claims case study, FCA AI approach page
