# AI for insurers and health payers Insurers and health payers want lower claims costs and healthier members, without exposing the people in their data. We build privacy-first tools for risk and claims integrity, and we are developing Averra, a digital health twin, for insurers and payers. Last updated 26 September 2026 ## Challenges we help with - Claims leakage, and fraud, waste and abuse that is hard to see in time. - Rising claims costs, and the case for prevention. - Engaging members between claims. - Using sensitive health data without exposing the people it describes. - New obligations for AI used in insurance risk assessment and pricing. ## What is changing for insurers? AI is moving from pilots into claims and underwriting intake, while rules on AI in pricing, conduct and data arrive on fixed dates. What is in use now, and what is coming, with the source for each. ### In use now - **Agents are arriving inside core insurance systems** The main policy and claims platforms released agent frameworks in 2026, Duck Creek in April and Guidewire in August, each with support for the Model Context Protocol (MCP). Agents now work inside claims and policy systems rather than beside them. Sources: [Duck Creek launches Agentic AI Platform for underwriting and claims](https://www.duckcreek.com/resource/press-releases/launch-agentic-ai-platform-underwriting-claims/) (Duck Creek Technologies, 2026); [Guidewire introduces Qusar release to help insurers build and control AI agents](https://www.guidewire.com/about/press-center/press-releases/20260803/guidewire-introduces-qusar-release-to-help-insurers-build-and-control-ai-agents) (Guidewire, 2026) - **Most generative AI is still in the back office** EIOPA's survey of European insurers, published in February 2026, found that most generative AI use cases were back-office tools, such as reading invoices, audio and medical reports, and that most were still at proof of concept, with strong human oversight. Source: [EIOPA survey on generative AI shows swift but cautious adoption among Europe's insurers](https://www.eiopa.europa.eu/eiopa-survey-generative-ai-shows-swift-cautious-adoption-among-europes-insurers-2026-02-02_en) (EIOPA, 2026) - **The FCA wants evidence of good claims outcomes** In its December 2025 review of outcomes monitoring under the Consumer Duty, the FCA asked insurers to monitor settlement values and find the root causes of declined claims. Completing a process is not, on its own, evidence of a good outcome. Source: [Insurance multi-firm review: outcomes monitoring under the Consumer Duty](https://www.fca.org.uk/publications/multi-firm-reviews/insurance-multi-firm-review-outcomes-monitoring-under-consumer-duty) (Financial Conduct Authority, 2025) - **US states expect a written AI programme** States adopting the NAIC's model bulletin expect insurers to run a written programme for AI, with governance and oversight of third-party models and data. Colorado and New York add testing duties for AI and external consumer data in underwriting and pricing. Sources: [Implementation of the NAIC Model Bulletin on the use of AI systems by insurers](https://content.naic.org/sites/default/files/cmte-h-big-data-artificial-intelligence-wg-map-ai-model-bulletin.pdf) (NAIC, 2026); [Amended Regulation 10-1-1, governance and risk management framework](https://doi.colorado.gov/announcements/notice-of-adoption-amended-regulation-10-1-1-governance-and-risk-management-framework) (Colorado Division of Insurance, 2025); [Insurance Circular Letter No. 7 (2024)](https://www.dfs.ny.gov/industry-guidance/circular-letters/cl2024-07) (New York State Department of Financial Services, 2024) ### Coming next - **High-risk rules for life and health insurance pricing** Under the EU AI Act, as amended by Regulation (EU) 2026/1744, AI used for risk assessment and pricing in life and health insurance is high-risk from 2 December 2027. Insurers using it will need human oversight, an impact assessment and explanations for the people affected. Sources: [Regulation (EU) 2026/1744 (Digital Omnibus on AI)](https://eur-lex.europa.eu/eli/reg/2026/1744/oj/eng) (EUR-Lex, 2026); [AI Omnibus enters into force](https://digital-strategy.ec.europa.eu/en/news/ai-omnibus-enters-force) (European Commission, 2026) - **The London market moves on data standards** Lloyd's retired the Blueprint Two programme in March 2026, but the Core Data Record goes on as the London market's shared data standard, with delegated authority and claims versions consulted on in 2026. Sources: [Blueprint Two](https://www.lloyds.com/about-lloyds/blueprint-two) (Lloyd's, 2026); [Data Council launches consultation for Delegated Authority Core Data Record](https://lmg.london/news/data-council-launches-consultation-for-delegated-authority-core-data-record/) (London Market Group, 2026) - **Supervisors will ask about every AI system** US regulators piloted an AI Systems Evaluation Tool for insurance examinations from March to September 2026, and the PRA named AI adoption as a supervision topic for insurers in 2026. Expect structured questions about each AI system you use and who oversees it. Sources: [AI Systems Evaluation Tool pilot project summary](https://content.naic.org/sites/default/files/inline-files/Pilot%20Project%20Summary_1.pdf) (NAIC, 2026); [Insurance supervision: 2026 priorities](https://www.bankofengland.co.uk/prudential-regulation/publication/2026/january/letter-from-gareth-truran-and-shoib-khan-insurance-supervision-2026-priorities) (Prudential Regulation Authority, 2026) - **Solvency II adds climate risk** The Solvency II review, Directive (EU) 2025/2, applies from 30 January 2027 and brings climate and sustainability risk into insurers' risk management, with more exposure and scenario data to prepare. Source: [Directive (EU) 2025/2 amending Directive 2009/138/EC (Solvency II)](https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32025L0002) (EUR-Lex, 2025) Checked against these sources on 26 September 2026. Rules and dates change: we check them again at the start of every project. ## What we offer - **[Insurtech and health payers](https://gateway-global.co.uk/services/insurtech/)** Our service for insurers and payers. - **[Risk intelligence](https://gateway-global.co.uk/technology/risk-intelligence/)** Risk stratification, risk management and claims integrity. - **[Anonymisation and privacy engineering](https://gateway-global.co.uk/technology/anonymisation/)** Use sensitive data without exposing people. - In development **[Averra](https://gateway-global.co.uk/products/averra/)** A digital health twin designed for insurers and health payers, in development. - **[TPA and claims modernisation](https://gateway-global.co.uk/services/tpa-modernisation/)** Member support, claims and risk review for insurers and TPAs, rebuilt with AI agents. - **[AI assurance](https://gateway-global.co.uk/services/ai-assurance/)** Your pricing and claims AI classified under the EU AI Act, with evidence for supervisors. ## AI agents for insurers and payers Agents for claims, underwriting and customer teams. Each one gathers, drafts and routes; a person decides cover, claims and prices. ### Claims - **Claims intake agent** Takes first notice of a claim by phone, web or email, gathers the details and documents, and opens the claim in your claims system. A person approves: Cover, liability and settlement. It never accepts or declines a claim. - **Claims triage agent** Sorts new claims by complexity and urgency, routes each to the right handler, and says why. A person approves: Handlers can change any routing, and customers who may be vulnerable go straight to a person. - **Claims integrity review agent** Flags claims whose patterns may point to fraud, waste or overcharging, with the evidence behind each flag, for your counter-fraud team. A person approves: Every flag. A person decides what happens next, and no claim is refused on a flag alone. - **Recovery and subrogation agent** Finds settled claims where another party may be liable, and prepares the recovery file. A person approves: Whether to pursue a recovery, and every letter to the other party. ### Underwriting - **Underwriting submission agent** Reads broker submissions, from emails, schedules and slips, into structured risk data, and checks each against your appetite. A person approves: Every underwriting decision, term and price. - **Policy wording agent** Compares policy wordings and endorsements with your standard wording, and marks each difference for review. A person approves: An underwriter or wording specialist signs off every change. - **Bordereaux and delegated authority agent** Checks coverholders' and MGAs' bordereaux against your binding authority and the market's data standards, and lists every exception. A person approves: The delegated authority manager clears every exception. ### Health payers - **Prior authorisation agent** Gathers the clinical documents a health payer's rules ask for, prepares each prior authorisation request and tracks it to a decision. A person approves: A licensed clinician makes every medical necessity decision. The agent never refuses a request. ### Customers and conduct - **Policyholder service agent** Answers policy and claim status questions by phone, chat and email, in the caller's own language, and passes complex cases to your team. A person approves: Anything about cover, a claim decision or a complaint goes to a person. - **Customer outcomes agent** Reviews calls, letters and complaints for signs of poor outcomes or unfair treatment, and gathers the evidence for your Consumer Duty reporting. A person approves: The findings, and what goes to your board. Every agent is logged, tested before launch and after every change, and given only the access its task needs. [How we keep agents in control](https://gateway-global.co.uk/ai-agents/#governance) ## The systems they connect to An agent is only useful inside the systems your teams already work in. These are the platforms and standards insurers and brokers run on. - **Policy and claims platforms**: Guidewire InsuranceSuite, Duck Creek, Sapiens, Majesco, Salesforce Financial Services Cloud. - **Broker and agency systems**: Applied Epic, Acturis, Open GI, SSP. - **Data standards**: ACORD standards (AL3, XML and its JSON standards for APIs), ACORD delegated authority standards, Core Data Record and MRC v3 (London market). - **Fraud and claims data**: CUE and the Motor Insurance Database (MIB), Insurance Fraud Bureau, Cifas, Verisk ISO ClaimSearch. Access to each is through its provider's own membership rules. - **Health payers**: HL7 FHIR and the Da Vinci guides, X12 837 and 835 claims, DHA eClaimLink in Dubai, DoH Shafafiya in Abu Dhabi. We connect through each system's own interfaces, its open standards and the Model Context Protocol (MCP), and confirm each connection at the start of a project. Naming a product here does not mean we are its maker's partner. ## Standards and safeguards - People review consequential decisions: our tools inform them. - Designed around UK GDPR, including its rules on automated decision-making. - Designed with the EU AI Act in mind, which classes AI for risk assessment and pricing in life and health insurance as high-risk. ## Evidence - Gateway Global has filed three UK patent applications, including one on controlling where AI processing requests can go. [See our patent portfolio](https://gateway-global.co.uk/technology/patents/) - With Prime Edge AI, every device attests itself in hardware and routine processing stays on the device. [Read how it works](https://gateway-global.co.uk/technology/prime-edge-ai/) ## Questions people ask **Is Averra ready for insurers yet?** Not yet. Averra is in development. It is designed for insurers and health payers, and you can register your interest. **Do your tools decide claims?** No. They point to patterns for a person to review. Decisions about people stay with people. **Can you work with anonymised data?** Yes. Anonymisation, pseudonymisation and redaction are part of our privacy engineering work. ## Tell us about your portfolio. [Talk to us about your portfolio](https://gateway-global.co.uk/contact/insurers/) --- Source: https://gateway-global.co.uk/industries/insurance/ Updated: 2026-09-26