Solutions catalog
Browse and filter artificial intelligence solutions in the insurance industry.
124 solutions
AllianzGPT is Allianz Group's internal generative-AI chatbot combining large language models with secure integration of internal data, hosted in Allianz's Azure cloud. It launched on 1 September 2023 on GPT-3.5; by early 2025 it offered GPT-4o for chat, DALL-E for images and a separately ring-fenced DeepSeek version. A Standard tier serves employees for general productivity, and about 60 Premium ('Ask Me Anything') instances support functions such as audit and risk consulting. Allianz reported in February 2025 over 60,000 active users, 95% weekly engagement among active users and more than 10 million prompts (self-reported), aiming to reach all 158,000 employees. A German financial news report of April 2026 said Allianz plans to launch AllianzGPT 2.0 in 2026 as a platform open to all employees.
by AXA (Direct Assurance, France)
FR
Direct Assurance, the AXA Group's French direct motor insurer with over one million customers, signed a three-year strategic partnership with Tractable in May 2025 to deploy Tractable AI Triage. The computer-vision system assesses vehicle damage from photos submitted by policyholders within minutes, so that Direct Assurance can estimate compensation and give the customer an immediate settlement offer. The announcement was published by the vendor and discloses no metrics; no confirmation of live use or claim volumes was found as of October 2026.
Generali France uses Shift Technology's cloud-based fraud-detection solution Force, which scores claims in real time by combining thousands of internal and external data variables and alerts investigators to suspicious cases before payment. The partnership began in motor insurance in 2016 and was extended to property claims (multi-risk home, property and business policies), with automatic alerts from November 2019; about 107,000 of roughly 125,000 annual P&C claims fall within scope. Before deployment, 0.7% of property claims were identified as fraud against a 2% benchmark for leading insurers. Generali France later reported annual fraud-detection savings rising from about EUR 20m to about EUR 40m after the 2020 switch to real-time alerts, noting that not all of this comes through Shift (ActuIA, 2023).
Zurich UK uses Quantexa's Decision Intelligence platform for real-time claims fraud detection. As claims are submitted online or by phone, it maps the people, addresses, businesses and solicitors linked to a claim, checks them against CUE, the Insurance Fraud Register and Companies House data, and gives handlers a risk rating and scenarios for deciding on referral to the fraud team. Announced in October 2025 as Quantexa's first UK insurer partnership, it went live that month across six lines including motor, property and casualty, with Guidewire, Microsoft and Databricks as further vendors. Celent reports GBP 4m of net-new fraud found in the first nine weeks and about twice the legacy system's fraud indemnity impact. Winner of the 2026 Celent Model Insurer Award for Data, Analytics and AI.
Baloise uses Insiders Technologies' AI products (OvAItion document recognition and classification, the OmnIA automation platform, the GenerAItor no-code AI builder and Intelligent Invoice Management) for input management. About 140,000 incoming documents per month, some up to 1,000 pages and arriving via paper, email, chat, app and portals, are classified into more than 250 document classes and read, including claims notifications and supplier invoices. Baloise has used Insiders software since 2010; all six input-management tenants were migrated to the SaaS Insiders Cloud within 18 months (reported March/April 2025). Baloise reports that business requirements are now implemented three to four times faster. Baloise has been part of Helvetia Baloise since December 2025.
Generali France has rolled out Microsoft Copilot since 2023 with Avanade: all employees have Copilot Chat and 3,700 use Microsoft 365 Copilot, with 70% adoption and about 15 prompts per user per week (Microsoft, September 2025). Its Cognitive Factory builds AI and automation with Copilot Studio and Azure OpenAI; 17 use cases were in production in 2025. Examples include an NLP voice front-end that handled 1.3 million calls (about 30% resolved by customers themselves), intent-based routing of 1.2 million emails, an AI agent for home-insurance subscription, document recognition and about 300 RPA robots (2.1 million operations in 2024). In a 2025 staff survey, regular users reported saving 1 to 2.75 hours a week.
Zurich UK piloted Sprout.ai's Automated Policy Checking for property claims from December 2020 to February 2021. The engine uses natural language processing and knowledge graphs, trained on policy wordings and documents from over 20,000 past claims, to check claims against policy cover and surface context such as Ombudsman data, while claims handlers keep the decision. Zurich reported policy-checking accuracy above 98% and settlement decisions cut to under 24 hours (April 2021). A later Sprout.ai customer story featuring Zurich UK's Head of Claims Digital Experience describes integration into Zurich's claims systems for end-to-end property claims automation, but no production scale or post-pilot metrics have been published.
AXA Secure GPT is AXA Group's internal generative-AI service giving employees and business applications secure access to large language models. AXA engineers built it with Microsoft in about three months on Azure OpenAI Service, in a dedicated, privacy-compliant cloud tenant, and launched it in July 2023 for 1,000 AXA Group Operations staff. Employees use it to generate, summarise, translate and correct text, images and code, supported by over 100 department-specific prompts. By April 2024 it had reached AXA Group Operations staff in 14 countries and 140,000 employees had access (Microsoft, vendor-reported). AXA Switzerland confirmed internal use in February 2024. Mistral models are being integrated. In July 2026 AXA described Secure GPT as its internal LLM service for all employees while starting a progressive global roll-out of Microsoft 365 Copilot.
Group-wide partnership announced by Allianz SE and Anthropic on 9 January 2026. Claude models are to become part of Allianz's internal AI platform, free for all employees, and Claude Code is already used by thousands of Allianz developers. The partners are developing custom AI agents for multi-step workflows, from intake documentation to claims processing in motor and health insurance, with employees taking over sensitive or complex cases. They are also co-developing AI systems that log every decision, rationale and data source for traceability. At announcement the claims agents and logging systems were in development; no live deployment, unit or metrics have been published since.
Zurich's AI-powered CRM for agents was built in-house by ZCAM, Zurich's wholly owned AI and analytics company, with development starting in January 2024. It consolidates customer and policy data and past interactions in one interface following a 'three-click rule', adds tone analysis and message feedback for agents working by phone, email and WhatsApp, and recommends the most suitable product and next steps using an approach modelled on Spotify's recommendations. It is integrated with Microsoft Outlook and Salesforce. Business Insider (May 2025) reported use in four unnamed Zurich markets and internal early estimates of over 70% shorter servicing times. Distinct from Microsoft 365 Copilot for Sales used in Zurich Commercial Insurance.
by Zurich Insurance Group
Global
Zurich's Commercial Insurance business uses Cytora's agentic-AI risk digitisation platform as an upstream, headless intake layer for underwriting: it turns unstructured, multilingual broker submissions into structured, decision-ready risk data and returns outputs in local languages, feeding Zurich's Z-Horizon+ workbench (a configured Microsoft Dynamics 365). Zurich reported a 90-day rollout to five countries; in May 2026 a Zurich manager cited 11 countries live and 22 targeted by year-end, with 20+ markets planned within 16 months (Mexico is among them). Reported results (May 2026, Zurich and Cytora): manual triage time down 80% (75 to 15 minutes), digitisation accuracy up from 70-80% to 98%, and intake straight-through processing up from 10% to 95%.
ZFinder is a fraud-detection model developed in-house by Zurich Seguros (Spain) that combines big data and machine learning, comparing each claim with thousands of similar claims and improving with every analysis. Zurich applied it first to home insurance (announced June 2018), with plans to extend it to other lines under the supervision of its internal investigation team. Zurich reported that automatic fraud detection doubled after launch and that in 76% of claims investigated via ZFinder a loss adjuster's work was not essential. The most common fraud types it found were simulated claims (47%), claims for excluded cover (22%) and concealed pre-existing damage (17%). No newer figures were found.