--- title: "Insurance AI from Guidewire" description: "AI in insurance that's grounded, governed, and open. See demos, specs, and the platform behind Guidewire agentic AI for P&C underwriting and claims." url: "https://www.guidewire.com/products/technology/insurance-ai-from-guidewire.md" language: "en" locale: "en" --- # Insurance AI from Guidewire AI in insurance that's grounded, governed, and open. See demos, specs, and the platform behind Guidewire agentic AI for P&C underwriting and claims. Agentic AI for P&C Insurance Embedded where you work, on a trusted platform purpose-built for how you work. [Request a Demo](/en/about/get-in-touch/request-a-demo) ##### ProNavigator for ClaimCenter ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/pronavigator-v2.svg?h=90&iar=0&w=106) ##### ProNavigator for PolicyCenter ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/pronavigator-v2.svg?h=90&iar=0&w=106) ##### ProNavigator for BillingCenter ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/pronavigator-v2.svg?h=90&iar=0&w=106) ##### ProNavigator for InsuranceNow ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/pronavigator-v2.svg?h=90&iar=0&w=106) ##### Concierge ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/concierge-v2.svg?h=90&iar=0&w=106) ##### Agentic Framework ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/agentic-framework-v2.svg?h=90&iar=0&w=106) ##### Developer Assistant for Gosu ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/dev-gosu-v2.svg?h=90&iar=0&w=106) ##### Developer Assistant for Integrations ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/dev-integrations-v2.svg?h=90&iar=0&w=106) ##### Developer Assistant for Jutro ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/dev-jutro-v2.svg?h=90&iar=0&w=106) ##### Predict ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/predict-v2.svg?h=90&iar=0&w=106) ##### Industry Intel ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/hero-carousel/cooperative-scores.svg?h=90&iar=0&w=106) ## Move faster with AI that's trusted for insurance AI on Guidewire improves business outcomes across the entire P&C insurance lifecycle. ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/pc-lifecycle-light.png?h=580&iar=0&w=2170) ### Product Launch and update products quickly with AI-assisted design ### Pricing Shorten rate change cycles with dynamic pricing and rating ### Distribution Build digital experiences with natural language prompts ### Underwriting Triage submissions and write more profitably in less time ### Policy Lower policyholder call times with fast, accurate answers ### Claims Reduce claims leakage with end-to-end handling ## AI on Guidewire is grounded, governed, and open ### Grounded in insurance context Get accurate responses and actions you can rely on because AI operates in the same core system as you. - Define P&C rules and business processes once - Access business, operational, and transactional data without duplication - Review policy, claim, and risk information where it originates ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/pillar-grounded.png?h=293&iar=0&w=343) ### Governed for insurance compliance You decide what AI can do, and where to adjust its performance. - Test and evaluate AI performance before it reaches production - Observe and adjust performance with ongoing evaluations - Access audit traces and data logs to comply with P&C regulations ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/pillar-governed.png?h=319&iar=0&w=390) ### Open to connect and extend Get immediate value: use our tools, bring what you have, build what you need, or connect to third parties. - Use Guidewire AI agents, build your own, or connect to agents you already have - Choose the right AI model for the job, every time - Extend with solutions built by verified partners on Guidewire Marketplace ![](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/pillar-open.png?h=315&iar=0&w=328) ## How does AI work on Guidewire? Explore how AI works and how your teams can use it. Get an overview. See the demos. When you’re ready, we’re here to answer your questions. ### See the benefits Write more profitable business, improve claim cycle times, and give your teams the AI tools to enhance their work. [Learn more](/en/products/technology/insurance-ai-from-guidewire) ### Understand the capabilities Use Guidewire agents, build or connect your own, and control how AI works across your business. [Learn more](/en/products/technology/insurance-ai-from-guidewire) ### Learn about the platform Bring models, agents, P&C context, data, and governance together where the work of insurance happens. [Learn more](/en/products/technology/insurance-ai-from-guidewire) ## Insurance AI at scale ### $326B+ #### Direct written premium represented In the insurance intelligence that informs predictive AI on Guidewire. ### 60+ #### Insurers who are using ProNavigator in production Built for the policy and claims work insurers do every day. ### 41% faster #### Guidewire development with AI assistance Compared with generic AI coding tools. ## AI for the people who run insurance See how AI works for underwriters, adjusters, service teams, billing teams, policyholders and developers, embedded in the Guidewire products where real insurance work happens. Each demo shows a real task, the AI support around it, and the outcome it helps the team achieve. AI use cases Choose a role Underwriters ### ProNavigator for PolicyCenter Focus on the submissions worth writing, make more accurate risk decisions, and drive portfolio growth. Storylane demo of ProNavigator for PolicyCenter https://app.storylane.io/demo/mqsnihiuqw3c?embed=inline Claims adjusters ### ProNavigator for ClaimCenter Get to the facts sooner, spot what needs attention, and move the claim forward with more confidence. Storylane demo of ProNavigator for ClaimCenter, with Intel data via Predict https://app.storylane.io/demo/7dnn4na6tk6s?embed=inline Service teams ### ProNavigator for PolicyCenter Get answers in real time without having to put your customers on hold. Storylane demo of ProNavigator for PolicyCenter https://app.storylane.io/demo/lawodsclwt4j?embed=inline Billing teams ### ProNavigator for BillingCenter Resolve billing questions faster with instant answers on payment plans, invoices, and account status. Storylane demo of ProNavigator for BillingCenter https://app.storylane.io/demo/59nvajggtuzo?embed=inline Policyholders ### Agentic FNOL in Concierge Let policyholders report a loss in their own words and get a complete first notice of loss filed without waiting on hold, 24/7. Storylane demo of Agentic FNOL in Concierge https://app.storylane.io/demo/yikpkq3efvtj?embed=inline Builders and developers ### Developer Assistants Build and ship insurance solutions faster with less friction and higher quality output. Demo of Developer Assistants https://app.storylane.io/demo/sp02qhblxcxf?embed=inline ### Agent Studio Build, test, and deploy custom AI agents for your own insurance workflows, with governance controls built in. Demo of Agent Studio https://app.storylane.io/demo/azxbscetbkyu?embed=inline ## Insurers are already seeing the impact of AI on Guidewire ### Western Financial ProNavigator [Read the story](/en/customers/success-stories/western-financial-group) ![Western Financial Group logo](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/logo-western-financial-group.jpg?h=1000&iar=0&w=1000) Western Financial saved more than 10,000 hours annually with ProNavigator ### Red River ProNavigator [Read the story](/en/customers/success-stories/red-river-mutual) ![Red River Mutual logo](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/logo-red-river-mutual.jpg?h=1000&iar=0&w=1000) Red River puts critical information in the hands of its employees in near real-time with ProNavigator ### Peel Mutual ProNavigator [Read the story](/en/about/press-center/press-releases/20260608/peel-mutual-insurance-company-increases-ai-driven-innovation-and-growth-with-guidewire-cloud) ![Peel Mutual Insurance logo](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/logo-peel-mutual.jpg?h=1000&iar=0&w=1000) Peel Mutual will use ProNavigator to minimize administrative tasks that burden underwriters ### Germania Mutual Insurance ProNavigator [Read the story](/en/about/press-center/press-releases/20260716/germania-mutual-deploys-guidewire-ai-to-put-institutional-knowledge-at-the-fingertips-of) ![Germania Insurance logo](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/logo-germania-insurance.jpg?h=1000&iar=0&w=1000) With ProNavigator, Germania Mutual Insurance was able to streamline internal operations and save employees time with faster information validation ## Frequently asked questions ### What makes AI on Guidewire Cloud Platform different from general-purpose AI? AI on Guidewire is designed for the way P&C insurance actually operates. It works with insurance context, acts within the controls you set, and gives you flexibility in the models and agents you use. AI on Guidewire is grounded, governed, and open. When those three things come together, AI can move beyond generic assistance and support real insurance decisions and actions at scale. For example, an agent reviewing a claim can work from the actual policy, claim history, and business rules, while staying within defined permissions and using the AI capabilities you choose. ### How does Guidewire keep AI grounded in insurance? Guidewire brings AI to the core system of record where insurance context and transaction data already lives. Because that context does not have to be recreated in a separate AI layer, teams spend less effort translating information between systems and AI can work with the same business context used to make the decision. For example, an underwriter reviewing a submission can work with current risk information and the surrounding underwriting context already available in Guidewire. ### Can our AI strategy change as models and technology evolve? Yes. Guidewire’s approach with AI Connect is explicitly intended to allow insurers to easily adapt to a rapidly changing AI landscape, using the right model for the job, every time. AI on Guidewire supports foundation models from Anthropic (Claude), OpenAI (GPT), AWS (Nova), and Meta (Llama), and all inference requests flow through AI Connect rather than being hard-coded into individual agents, so the model behind an agent can be changed via Agent Studio or the Agentic SDK, with zero downtime. Maximum flexibility means adopting a new model does not have to mean redesigning the claims or underwriting logic around it. Insurers can also have different models running for different use cases. For example, an insurer could introduce a new Anthropic model for a specific underwriting use case while keeping an OpenAI GPT model for an existing claims workflow intact. ### How should insurers measure the value of AI? Measure the business outcome that changes, not the amount of AI activity. If AI saves time but the insurer does not write more profitable business, reduce expense, improve recovery, or protect loss performance, the activity has not yet demonstrated economic value. Connecting AI to those outcomes makes it possible to see whether an investment is actually changing business performance. For example, underwriting AI can be measured by new business bound per underwriter FTE rather than by prompts submitted or content generated. Read the white paper: Measuring the Value of AI in P&C Insurance ### Can I use my own AI agents with Guidewire? Yes. You can use Guidewire AI agents out-of-the-box, build your own custom AI agents, integrate AI agents you already have using MCP servers and Cloud APIs, and extend Guidewire Cloud Platform even further with pre-validated AI agents from the Guidewire partner ecosystem. When different use cases can draw on different agentic skills and capabilities, teams can choose specialized AI agents while maintaining a common insurance platform underneath them. For example, an insurer could use a Guidewire agent for one claims process and connect a specialized agent it already uses for another. ### Can we determine how an AI agent reached a specific claim or underwriting decision? Yes. On Guidewire Cloud Platform, every model call, retrieval step, decision and action made by an agent is captured as a unified audit trace, scoped to the permissions of the user who invoked the agent, and customers can store and archive these traces for long-term audit access. For example, if an examiner asks about a claim decision, compliance teams can review the trace of what the agent retrieved and which tools it called. Model versions are also pinned and promoted deliberately through AI Connect by Guidewire, so a third-party provider’s model update cannot silently change agent behavior in production. ## Just the facts: AI on Guidewire Specifications for the AI assistants, agents, and tooling available today. Updated October 2026. [Contact us](#speak-to-an-expert) for the latest capabilities and availability. ### ProNavigator AI assistant with agentic capabilities inside Guidewire core applications, grounded in insurer policies, guidelines, and transactional data [Learn more about ProNavigator](/en/products/core-products/pronavigator) ##### Answers and actions Get insurance-specific answers and completed work without leaving the core application ¹ Restricted Availability (RA) in ProNavigator for ClaimCenter; ² Early Access (EA) in ProNavigator for PolicyCenter; ³ Coming soon ###### Contextual answers Answers questions about the policy, claim, or task in progress, with links to original source documents and notes. ###### Recommended actions Recommends next-best actions on policies and claims and executes administrative tasks when prompted. ###### Claim Summarization¹ Generates claim summaries and ongoing updates in seconds. ###### Policy Change² Executes multi-step policy updates, validates requested changes against rules and product models, and provides AI-drafted change jobs for review and binding. ###### Application availability ClaimCenter, PolicyCenter, BillingCenter³, InsuranceNow ##### Knowledge and grounding Answers come from your documents, your guidelines, and your core system data ###### Insurance knowledge base Centralizes unstructured documents, forms, endorsements, business guidelines, rate manuals, and standard operating procedures into a retrievable knowledge base for ProNavigator answers. ###### Contextual retrieval Produces insurer-specific responses by retrieving policies, business guidelines, and core system transactional data rather than searching the open web. ##### Agents, assistants, and Model Context Protocol Extend beyond standard processes with custom agentic AI development ¹ Included for all ProNavigator customers on Guidewire Cloud Platform; ² Coming soon ###### Agentic Framework¹ Build your own custom AI agents for non-standard workflows and proprietary business processes. ###### Developer Assistants Developer Assistant for Gosu; Developer Assistant for Integrations ###### Model Context Protocol (MCP) server access Claims MCP Access with ProNavigator for ClaimCenter; Policy MCP Access with ProNavigator for PolicyCenter; Billing MCP Access with ProNavigator for BillingCenter²; InsuranceNow MCP Access with ProNavigator for InsuranceNow ### Builder and Developer Assistants Guidewire-aware AI assistance for development, inside the AI coding tools and IDEs your teams already use ##### Developer assistants Accelerate Gosu, Integration, and Jutro development ###### Developer Assistant for Gosu Writing, refactoring, and optimizing Gosu code ###### Developer Assistant for Integrations Building, changing, and maintaining Integration Gateway flows ###### Developer Assistant for Jutro Building digital experiences on the Jutro Digital Platform ##### Tools and Guidewire context Connect Guidewire knowledge to supported AI coding tools and development environments ###### AI coding tools supported Claude Code; OpenAI Codex; Cursor; Gemini CLI; GitHub Copilot CLI; GitHub Copilot ###### IDEs supported IntelliJ; Guidewire Studio ###### Accessible by MCP Guidewire Cloud APIs; proprietary programming languages; data models; curated libraries; design patterns ##### Usage and data controls Use existing AI service credentials while keeping Guidewire code and prompts outside foundation-model training ###### Bring your own key Developer Assistants are MCP servers that are connected in the insurer’s existing AI coding agents using your own service-provider credentials; Guidewire does not charge for token consumption using Developer Assistants. Tokens are consumed and charged by your third-party AI service provider. ###### Code privacy Data retrieved from Guidewire Cloud Platform by Developer Assistants is protected by our AI security protocols. Users are still responsible for adhering to their own company data policies and security protocols when using third-party AI service providers. ### Agentic Framework Managed runtime, low-code studio, and TypeScript/JavaScript SDK for building, running, and governing custom AI agents on Guidewire Cloud Platform ##### Agent build layer Create, deploy, and manage AI agents that know P&C insurance ###### Agent Studio Low-code visual interface to build, test, deploy, and manage custom AI agents. ###### Agentic SDK TypeScript/JavaScript toolkit to build agents via code and integrate them with existing engineering practices and CI/CD pipelines. ##### Agent core layer Enable AI agents to plan and execute the work ###### Task composition Decomposes high-level goals into subtasks, routes dynamically from intermediate outputs, and validates data against strict JSON schemas before passing it to the core system. ###### Memory management Preserves active session state across interruptions and indexes long-running workflow history so agents can resume later using metadata-filtered context such as claim ID, jurisdiction, or line of business. ###### Agent and subagent orchestration Delegates discrete workflow steps from a coordinating agent to specialized subagents and passes summarized results between steps. ##### Contextual awareness layer Keep AI agents grounded in real insurance context ###### Insurance knowledge base Retrieval-augmented generation (RAG) with metadata filters for state, policy type, effective date, and line of business, returning retrieved information with citations. ###### MCP servers & tools Exposes Guidewire Cloud APIs, external databases, and partner systems as MCP tools; agent actions run through Guidewire core validation pipelines and business rules. ##### AI gateway and governance layer Manage how models are accessed, routed, tested, and controlled ###### Foundation models Anthropic Claude; OpenAI GPT; AWS Nova; Meta Llama ###### Gateway controls Centralized model routing, token limits, usage metering, and automated failover through AI Connect. ###### Offline evaluations Evaluations against insurance-specific custom datasets and ground truth using Recall@K, Precision, F1, and nDCG. ###### Production evaluations Automated checks of responses and planned actions for safety, groundedness, and relevance; failures are blocked and routed to a human. ###### Audit tracing Unified trace of every model call, retrieval step, and tool call, scoped to the permissions of the human user that invokes the AI agent. ###### Observability Datadog evaluation scores and step-by-step audit trace trees for agent reasoning and production performance. ##### Security and data controls AI agents operate within the insurer’s single tenant isolation and are bound by role-based access permissions ###### Access controls AI agents run in the insurer's isolated cloud tenant and inherit the acting user's single sign-on (SSO) and role-based access control (RBAC) permissions. ###### PII redaction Personally identifiable information is redacted from agent activity before it reaches audit logs. ###### Model training restriction Data sent to a model is blocked from being used for model training. ### Predict Predictive modeling and real-time scoring using your own data, third-party models, and Guidewire Industry Intel models [Learn more about Predict](/en/products/analytics/predict) ##### Embedded insights Predictive scores delivered inside PolicyCenter and ClaimCenter ###### Predict for Profitability Embedded analytic insights in PolicyCenter for rating and risk selection ###### Predict for Claims Embedded analytic insights in ClaimCenter for severity segmentation, litigation potential, claim routing, and subrogation likelihood ##### Data and model development Import data, build or bring models, and manage model development ###### Predictive models Create, manage, and evaluate custom models from imported datasets; import third-party models; integrate Guidewire Industry Intel models ###### Supported algorithms Blazing Text; CatBoost; Decision Trees; Explainable Boosting Machine; Forward Stepwise GLM; Hugging Face; Individual GAM-H2O; Individual GLM; Random Forest; Text Mining; TF-IDF; XGBoost ###### Model development tooling Auto-bucketing; AutoML hyperparameter tuning; chaining models; exploratory data analysis; model comparison; pipeline construction; post-bucketing analytics; real-time scoring; variable management ##### Monitoring and governance Track model behavior and produce evidence for model risk and compliance ###### Model evaluation metrics Deviance; Gini coefficient; model estimate and output scores; RLLS point estimates; signal, noise, and selection scores; consistency; spread ###### Performance monitoring Transaction and model performance metrics, score distributions, and input-value deviation monitoring ###### Model governance framework Machine learning and statistical model governance for reliability, ethics, and regulatory compliance, based on fairness, accountability, transparency, and proactiveness ### Industry Intel Prebuilt, market-validated claims models built from pooled P&C data contributed by more than 115 insurers [Learn more about Industry Intel](/en/products/analytics/industry-intel) ##### Industry data Pooled insurer data and federated learning used to develop industry predictive models ###### Pooled industry data P&C data from more than 115 insurers representing $326B in direct written premium, used to build prebuilt, market-validated predictive models and scores ###### Federated machine learning Analytics code operates in the insurer’s system locally and returns only mathematical insights; raw sensitive data does not move or get extracted to another system. ###### Model governance framework Reliability, ethics, and regulatory compliance based on fairness, accountability, transparency, and proactiveness ##### Claims models and scores Segment and triage claims on day one, without building a model first ###### U.S. Auto Collision Subrogation Intel Subrogation recovery likelihood for collision claims ###### U.S. Auto Bodily Injury Intel Attorney likelihood; litigation likelihood; payment likelihood; estimated severity; estimated loss ###### U.S. Workers' Compensation Intel Legal likelihood; medical payment likelihood; medical severity tier; indemnity payment likelihood; indemnity severity tier; overall loss tier ###### Application availability Claims Intel models are embedded directly in ClaimCenter and InsuranceNow through Predict with Analytics Manager. ### The complete P&C platform built for the AI era More than 570 insurers already trust Guidewire to give them the agility to engage, innovate, and grow efficiently. Core operations, transactional data, workflows, APIs, and enterprise security standards exist together in one place. This is the foundation for a reliable, governed platform for agentic AI that doesn't require insurers to recreate that context anywhere else. [Learn more about Agentic Framework](#agentic-ai) [Learn more about security and governance](#security-governance) ![Guidewire Cloud Platform: Applications and Agentic Framework sit on Application Services, Data Services and Platform Services. People and agents use Applications; agents use the Agentic Framework; developers and agents work across every layer; Marketplace connects to all layers.](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-page/platform-marketecture.png?h=564&iar=0&w=1185) ## Harness agentic AI with insurance-grade control See the tools insurers use to build, connect, evaluate, and manage AI agents on Guidewire Cloud Platform. Guidewire Agentic AI capabilities Choose a capability Build agents ### Agent Studio Build AI agents that have insurance expertise out of the box. Keep your AI agents and data under the same roof. Build custom agents using a low-code visual interface, or a pro-code toolkit to compose complex P&C workflows; whichever fits the task and the team building it. Then pair it with the AI model and MCP server best suited to the work, and deploy it to production with minimal friction. ![Agent Studio authoring canvas](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-agentic/agent-studio.png?h=1014&iar=0&w=1920) Connect LLMs ### AI Connect Switch AI models with zero downtime. AI Connect routes every AI agent's requests through a single gateway. That means as better models come out, your team can move as fast as the AI landscape does. Pointing your AI agents to a new model only takes a few minutes, giving you the flexibility to choose the best LLM for each job, anytime. ![AI Connect integrations catalog](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-agentic/ai-connect.png?h=1014&iar=0&w=1920) Evaluate agents ### Testing and Evaluations Test and evaluate every AI agent end-to-end. AI agents are scored offline against insurance-specific golden datasets before deployment, then monitored in production against safety rubrics. It's a system you can measure and modify, step by step, so every agent keeps getting better. ![Evaluation run results](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-agentic/testing-evaluations.png?h=1014&iar=0&w=1920) Manage agents ### Agent Management Observe and control every agent from within a single console. Give your AI development teams a view of your agents in production. Every LLM call, retrieval step, and tool used becomes part of a single, unified trace that captures the full reasoning path behind an AI-driven decision. When a regulator or auditor asks why an AI agent did what it did, your compliance team will already have the answer. ![Central AI Tower console](https://edge.sitecorecloud.io/guidewiresodb06-guidewire0f2e-productioncd91-5186/media/project/guidewire/guidewire/components/run-ai-agentic/agent-management.png?h=1014&iar=0&w=1920) ## AI security and governance ## Inherit a secure foundation insurers already trust Guidewire Cloud Platform is the managed enterprise platform used by the largest number of insurers, globally. It processes and stores vast amounts of sensitive industry data with enterprise-grade security. Shared responsibility model Guidewire is responsible for the security of the platform, orchestration layer, and base configuration of our applications. Customers are responsible for securely using the platform, retaining full ownership and management of their specific configurations, custom code extensions, and integration interfaces. Certification and attestations Guidewire Cloud Platform holds the certification stack security teams expect from an enterprise SaaS provider, including SOC 1; SOC 2 Type II; ISO/IEC 27001; ISO/IEC 27701; and PCI DSS (ROC). ## Customer data is handled with care No training, zero retention Customer data passed to third-party AI model providers such as Anthropic, OpenAI, AWS, and Meta via Amazon Bedrock is never used to train or improve that provider's foundation models. These providers are contractually bound, and technically configured, not to retain customer data after the inference request is served. Centralized routing All inference requests flow through AI Connect and are routed through Amazon Bedrock rather than being hard-coded into individual AI agents, so controls are enforced in one place and the model behind an agent can be changed by configuration. PII redaction PII (personally identifiable information) is redacted before data is sent to audit logs. Regional data residency Isolation zones bind each customer tenant to specific regions, so data processing and storage stays within its respective jurisdiction. That keeps data aligned with regional obligations such as GDPR in EU zones, PIPEDA in Canada, and APPI in Japan. Tenant data stays within a single isolation zone, and multi-region deployments are implemented as one tenant per zone, so data does not silently cross regional boundaries. ## AI is governed for compliance, with human oversight On Guidewire Cloud Platform, audit-ability is treated as a runtime property, producing the artifacts that regulators and internal audits expect. The following controls exist specifically to address risks that are unique to AI, such as hallucination, unsafe inference, prompt injection, and model drift. Glass-box observability Every model call, retrieval step, and tool call is captured as a unified workflow trace that can be stored and archived by customers for long-term auditing access. Human-in-the-loop guardrails Agentic outputs in Guidewire applications are evaluated against safety, grounding, and relevance rubrics. Outputs that fail any rubric are blocked and routed to a person to comply with human oversight requirements of the EU AI Act and the NAIC AI Model Bulletin. Bounded autonomy AI agents act through validated core APIs and cannot bypass business rules or core system logic. An agent can’t reach a database directly, or skip the validation pipeline, due to its inherited access control. Inherited access control AI agents run under Guidewire Hub single sign-on (SSO) and the customer’s existing role-based access controls (RBAC), so an AI agent operates within the same permissions as the human user who invokes the AI agent. Evaluated before production Guidewire customers can run offline evaluations to score AI agents against their custom datasets before deployment, allowing insurers to catch hallucinations early and validate model version changes before they ever reach production. Controlled model versioning Model versions are pinned and promoted deliberately with AI Connect, so a third-party provider's update cannot silently change agent behavior. The same input doesn’t produce different outputs without a tested, intentional version change. Guidewire customers are encouraged to conduct a direct assessment of the AI capabilities against their own security frameworks via documentation review, requested audit reports, and collaborative deep-dive sessions with Guidewire technical experts. ## Speak to an expert See a demo of AI on Guidewire. Or tell us what you are trying to figure out and we will send it to the right person.