First AI Agent Identity Codes Issued: Agent BI Enters the Governable Era

On July 21, 2026, at the Beijing Zhongguancun Exhibition Center, a seemingly routine standards conference accomplished something landmark: the first batch of AI Agent identity codes were officially issued.

Over 200 enterprises claimed their dedicated nodes — Alibaba 1688, Meituan, Didi, Xiaomi, China Southern Power Grid, and all three major telecom operators among them. Every AI Agent now has its own "digital ID card."

What does this mean? It means the "wild west" phase of AI Agent development over the past two years has officially entered a new era of registerable, traceable, and auditable governance. For the enterprise BI sector, this is not a story to skim past.

From "Usable" to "Governable": A Gap That's Been Overlooked

Over the past 18 months, nearly every BI vendor has been telling the same story: ask data questions in natural language, let AI write your SQL, run attribution analysis, and generate reports. The "usability" of Agent BI is no longer in question.

But what enterprise CIOs and CDOs worry about has never been "can it work?" — it's three sharper questions:

• Who is calling? When an agent accesses your core operational data, is its identity verifiable? Which vendor built it, what version, and what permission boundaries does it operate within?
• How do they collaborate? An enterprise may run data analysis agents, approval agents, and customer service agents simultaneously. How do they "talk" to each other? With MCP, A2A, and proprietary solutions coexisting without unified interfaces, collaboration remains theoretical.
• Who is accountable? When an agent autonomously makes a business decision — say, adjusting marketing budget allocation for a region — is the behavior auditable and traceable?

These three questions are precisely what the new national standard attempts to answer.

GB/Z 185: Seven-Layer Closed Loop for Agent Registration

AI — Agent Interconnection (GB/Z 185.1–185.7—2026) is China's first standard system covering the full lifecycle of AI agents. Seven parts form a closed loop:

• Overall Architecture → Identity → Trusted Management → Capability Description → Intelligent Discovery → Multi-party Interaction → Tool Invocation

The logic is clear: first establish identity (who are you) → then describe capabilities (what can you do) → enable discovery and interaction (how do others find you and collaborate) → finally standardize tool invocation (how do you operate external systems).

The identity code system uses a layered identification scheme similar to internet domain names — enterprises select a unique abbreviation as prefix, one code per agent, ensuring uniqueness and traceability from the source.

The companion AIP (Agent Interconnection Protocol) V2.1 release addresses six challenges in one package: trusted access, identity authentication, capability discovery, interconnection collaboration, settlement transactions, and behavior auditing. The source code is already open-sourced on the AtomGit community.

What This Means for Agent BI

Back to the enterprise data analytics scenario. A typical Agent BI usage chain looks like this:

A business user asks a question in natural language → the data analysis agent interprets the intent → calls data sources or APIs to retrieve data → executes analysis logic → returns visualization results or recommended actions.

Along this chain, the national standard brings three direct changes:

First, trusted identity. An agent is no longer a "black box." Through the identity code system, enterprises can definitively know: this analysis agent accessing my CRM data belongs to which platform, what version, and has passed which security certifications. This is the prerequisite for data security compliance.

Second, composable capabilities. When agent capability descriptions and discovery mechanisms are standardized, multiple agents within an enterprise can truly collaborate — the data analysis agent detects an anomaly in reports, automatically triggers an attribution agent to dig deeper, then links with a notification agent to push conclusions to relevant decision-makers. From "point intelligence" to "chain intelligence."

Third, auditable behavior. Standardization of the tool invocation layer means every step an agent takes — which tables it queried, which models it used, which judgments it made — leaves a traceable record. This is especially critical for heavily regulated industries such as finance, healthcare, and government services.

Quick BI's Governable Agent BI in Practice

Quick BI has already made systematic moves toward governable Agent BI.

Smart Q (Quick BI's natural-language analytics assistant) follows a design philosophy highly aligned with the national standard: behind every natural language query, there is explicit identity authentication, permission control, and operation logging. What you asked, what intermediate reasoning the AI performed, and what conclusions it returned — the full chain is traceable.

On multi-agent collaboration, Quick BI's skill-based architecture is already exploring an "agent registration → capability declaration → on-demand invocation" model: a data insight skill can be called by agents in other business systems through standard interfaces, rather than through tight coupling.

• Recognized in the Gartner Magic Quadrant for Analytics and Business Intelligence for 7 consecutive years
• The only Chinese BI vendor with continuous presence on the quadrant
• A position earned not just through analytics capability, but through depth of understanding of security, compliance, and governance in enterprise scenarios

From "asking data questions" to "governing the process," from "single-agent answers" to "trusted multi-agent collaboration" — this is the critical step for Agent BI to move from demo-grade to enterprise-grade.

What Happens Next

The standard is currently published as a "guidance technical document," reflecting an agile standardization approach during the industry cultivation phase. But the signal is already clear:

• Beijing serves as the launch city, with plans to cover 120 key cities nationwide
• Over 90 primary and secondary schools in Beijing's Xicheng District have already integrated AI agents into pilot programs
• Beijing University of Posts and Telecommunications has embedded agents into its full academic administration workflow

For enterprises, now is the time to examine your Agent BI solution. If you are selecting or upgrading a BI system, ask one more question: is its agent capability governable?

Verifiable identity, describable capabilities, auditable interactions — these three criteria will shift from "nice to have" to "entry requirements." The second half of Agent BI has just begun.