The Agent-to-Agent (A2A) Manifesto
A living manifesto
These principles are our guiding principles, not a frozen doctrine. As we build real agentic insurance rails with our Founding Partners, this manifesto will continue to evolve and sharpen.
We believe a profound shift is happening
The digital economy is shifting from search to delegation. Consumers are no longer looking for information; they are instructing AI agents to execute complex tasks on their behalf.
But while AI is moving fast, regulated industries - like insurance, investments, and lending - can't run on probability. A hallucinated fact is a mistake; a hallucinated insurance policy is a regulatory and customer failure.
To bring the world's most vital financial safety nets into the agentic era, we must establish a new foundation of trust. We, the undersigned, commit to building the Agent-to-Agent (A2A) economy on these five core principles:
I. Create certainty from ambiguity
Language models are brilliant at conversation, but not built to calculate risk. Regulated products require absolute certainty. We believe AI should orchestrate the customer journey, but only deterministic, carrier-verified pricing engines should generate the quote. We commit to a zero-hallucination standard for financial data.
II. The customer's right to safe delegation
Customers have the right to delegate their financial administration to AI assistants without the fear of silent mis-selling. We believe that an AI agent acting on behalf of a human must be bound by the same standards of clarity, transparency, and duty of care as a human broker.
III. Insurer sovereignty and risk integrity
AI platforms will become the new distribution channels, but they must not become proxy underwriters. We believe carriers must retain absolute control over their risk appetite, pricing logic, and how their products are communicated. The infrastructure connecting AI to insurers must protect the carrier's core IP, not dilute it.
IV. Compliance as code, not friction
Regulation is not an obstacle; it is the bedrock of consumer trust. We believe that in the agentic era, compliance must be encoded directly into the routing protocols. Immutable audit trails must accompany every AI-negotiated transaction, proving exactly how a policy was constructed.
V. Open protocols over walled gardens
The future of regulated AI distribution will not be won by a single closed platform. It requires a shared, canonical language - a universal protocol that allows any authorised AI agent to communicate safely with any insuring entity. We commit to standardising the infrastructure that connects intent to execution.
Our commitment
The agentic future is inevitable. How we protect the consumer and the carrier during this transition is not. We are building the rails for a compliant, automated, and infinitely scalable financial ecosystem.
Join us in establishing the standard.