On 26 June 2026, the National Standardization Administration of China (SAC) held a press conference to announce the release of seven national standardization guiding technical documents on AI agent interconnection, designated GB/Z 185-2026 Artificial intelligence - Agent interconnection. 

 Drawing parallels to mainstream Agentic AI protocols such as MCP and A2A, the series introduces the Agent Interconnection Protocol (AIP), a new protocol framework to address persistent domestic challenges including AI agent silos, lack of traceability, and high innovation costs. 

The AIP adopts a governance-centric approach focused on equipping AI agents with traceable, verifiable identities across the full lifecycle of agent interaction, complementing existing MCP and A2A. 

The seven-part standard series comprises: 

1. GB/Z 185.1-2026 Artificial intelligence - Agent interconnection - Part 1: General architecture  

2. GB/Z 185.2-2026 Artificial intelligence - Agent interconnection - Part 2: Identity code 

3. GB/Z 185.3-2026 Artificial intelligence - Agent interconnection - Part 3: Identity management 

4. GB/Z 185.4-2026 Artificial intelligence - Agent interconnection - Part 4: Agent description 

5. GB/Z 185.5-2026 Artificial intelligence - Agent interconnection - Part 5: Agent discovery 

6. GB/Z 185.6-2026 Artificial intelligence - Agent interconnection - Part 6: Agent interaction 

7. GB/Z 185.7-2026 Artificial intelligence - Agent interconnection - Part 7: Agent tool invocation 

Part 1 establishes the top-level technical framework for agent interconnectivity. Parts 2 and 3 lay the trust foundation by issuing "digital IDs" to each agent, enabling identity verification and mutual recognition. Part 4 provides capability profiles for transparency on functional scope, while Part 5 sets up supply-demand matching. Parts 6 and 7 define collaboration rules and tool invocation protocols, closing the operational loop. 

Together, the standards form a closed-loop, full-lifecycle system. Standardizing architecture and interaction rules allows enterprises to reuse components, reduce custom development, and shorten time-to-market. Unified identity authentication and end-to-end traceability mechanisms also strengthen the institutional foundation for trusted, secure cross-domain interactions. 

Although they remain guiding documents, the standards are positioned as key enablers of China's AI+ Action Plan. SAC officials noted that regulators are accelerating mandatory standards on agent identity codes, alongside parallel work on specific areas such as agent auditing and trading, to quickly establish a credible rules framework. These standards are also expected to be submitted to ISO and IEC, with the aim of elevating them from Chinese national rules to internationally recognized norms. 

For foreign stakeholders, GB/Z 185-2026 carries significant national strategic weight and signals the direction of upcoming mandatory standards. All AI agents deployed in China will need a traceable identity code to achieve compliance and secure market access. Enterprises should track the series closely and prepare their systems for technical alignment with China's evolving AI governance requirements.