Peak3 has released Graphene v4, an AI-native insurance core platform, alongside Graphene Harness, which it calls the first global insurance AI-driven delivery lifecycle. The combination targets the high cost and slow pace of insurance software change by applying AI across the full build-and-maintain pipeline, not just to isolated tasks.
Bill Song, CEO of Peak3, said: "We are disrupting ourselves. Peak3 is moving from selling software and related services to providing the production line that builds it - our entire AI-DLC."
An agent-governed core across lines of business
Graphene v4 introduces the Graphene Agent Platform, a model-agnostic system for building, running, observing and governing AI agents. It ships with a marketplace of pre-built agents for medical underwriting, conversational first notice of loss, intelligent document processing and claims fraud detection. Agents can now reconfigure insurance products, calculations and rules from unstructured documents and natural-language prompts, grounded in the insurer's own data. Configuration work that previously took days finishes in minutes to hours.
The platform is open and headless. Every capability is exposed through APIs, MCP and CLI. Agents act on the core directly under access management, strict attribution and detailed logging enforced at code level, not through prompt-level guardrails. Graphene v4 can also deploy as a lightweight middle layer between AI agents and an insurer's existing systems of record, offering a fast route to agent-ready capabilities without a core replacement.
Functional breadth now spans the complete group health lifecycle including provider network management, commercial lines with co-insurance and reinsurance administered inside the core, indexed universal life and crypto-denominated policies, and native Takaful support alongside conventional business.
An AI production line for insurance delivery
Graphene Harness models the entire software delivery lifecycle as an AI-driven system. Peak3 built digital twins of every key delivery role - business analyst, architect, engineer, tester, reviewer - and connected them into a governed virtual scrum team. The system runs on an AI engineering platform, grounded in a curated knowledge base and connected with AI-driven testing, migration, diagnostics and CI/CD.
Peak3 already builds and maintains Graphene itself on Graphene Harness. The company is now making it available to selected implementation partners, with plans to offer it to clients within 6 to 12 months. Alongside the Harness launch, Peak3 is expanding its forward-deployed engineering teams who build AI-first operations with client teams across both Peak3 platforms and the client's existing stack.
Where Graphene Harness is applied, Peak3 reports a 50% reduction in end-to-end new-feature development and core system implementation cost, with timelines reduced accordingly. The company's stated ambition is 80% within 18 to 36 months. An insurer using Graphene Harness also owns the means of changing its core system, rather than depending on a vendor's delivery organisation.
Why this matters for insurance professionals
The cost of insurance IT has long been tied to headcount and elapsed time. An AI production line breaks that link - the same scope delivered by a smaller team in a shorter cycle. For insurers, this shifts the economics of core system change and reduces dependency on external delivery teams. For professionals working in AI for Insurance, the announcement signals a concrete benchmark: a 50% cost reduction in production today, with a roadmap to 80%. The skills that matter are shifting from traditional configuration toward governing and directing AI agents that do the work. For those on the build side, the AI for IT & Development implications are direct - digital twins of delivery roles are moving from experiment to production line.
Your membership also unlocks:
