New Protocol Enables AI Agents to Negotiate Commercial Deals Autonomously

Developers have released an open protocol designed to enable artificial intelligence agents to conduct business negotiations without human oversight, addressing what creators describe as a missing infrastructure layer for automated commerce. The Agent-to-Agent Commercial Negotiation (A2CN) protocol specifies standards for mandate verification, session management, offer exchange, and transaction recording between autonomous software systems. The specification includes cryptographic proof mechanisms for authority validation and dual-signed records to prevent disputes. The protocol targets enterprise procurement scenarios where AI agents are already generating tens of millions in savings for Fortune 500 companies, according to published case studies from firms including Pactum, Fairmarkit, and Zip. These systems currently operate with limited interoperability, requiring custom integrations for each supplier relationship. A2CN addresses eight identified failure modes in agent-to-agent commerce, including the absence of shared terminology for commercial terms, lack of standardized authority verification, and no neutral mechanism for proving agreement terms. The protocol uses W3C Decentralized Identifiers and JSON Canonicalization Scheme signatures for authentication. The reference implementation includes 202 passing tests and adapters for existing procurement platforms. Creators describe the protocol as aiming to become invisible infrastructure similar to TCP/IP or payment processing networks, where transactions flow through standardized layers without end-user awareness. Enterprise adoption will require supplier onboarding, which the protocol addresses through a session invitation mechanism allowing cold-start negotiations without pre-deployed endpoints.