Zetrix AI and CAICT Launch Know Your Agent Service

Zetrix AI Berhad and the China Academy of Information and Communications Technology (CAICT) have launched a Know Your Agent (KYA) service designed to establish verifiable digital identities and delegated permissions for autonomous AI agents. The framework aims to help businesses, governments and digital service providers verify who an AI agent represents, what actions it is authorised to perform and how its activities can be audited. A planned cross-border pilot involving organisations across China, ASEAN, the Middle East and Africa will examine how agent identities and permissions can operate across different jurisdictions.

Zetrix AI and CAICT Introduce Know Your Agent Framework for AI Governance

As artificial intelligence agents move from answering questions to accessing enterprise systems, interacting with digital services and executing transactions, establishing accountability for autonomous actions is becoming an important infrastructure challenge. Zetrix AI Berhad and the China Academy of Information and Communications Technology (CAICT) are addressing this issue through a new Know Your Agent (KYA) service that gives AI agents verifiable digital identities and clearly defined operating permissions.

Announced simultaneously at the World Internet Conference in Wuzhen, China, and the ICT Development Forum 2026 in Dubai, United Arab Emirates, the initiative introduces an identity and authorisation framework intended to help organisations distinguish legitimate AI activity from unauthorised or potentially malicious actions.

The service is designed to verify three fundamental elements: an AI agent’s identity, the individual or organisation it represents, and the authority delegated to it. This could become increasingly important as autonomous agents gain access to corporate applications, financial services, government platforms and other protected digital environments.

KYA Extends Digital Identity to Autonomous AI Agents

Traditional Know Your Customer (KYC) systems focus on establishing the identity of individuals and businesses. KYA applies a similar trust principle to AI agents, creating a mechanism to establish who controls an agent and what it is permitted to do.

Under the proposed framework, each participating AI agent receives a verifiable digital identity linked to an authenticated individual or organisation. Its owner can establish permissions defining which services the agent may access, what information it can use, which parties it can interact with and what transactions it can execute.

For example, a business could authorise an AI agent to process payments to approved suppliers within a specified spending limit. Transactions exceeding that threshold could require additional human approval.

Similarly, an agent attempting to access a government or enterprise service could be required to demonstrate its identity and delegated authority before receiving access to protected information.

This introduces an additional control layer for agentic AI deployments. Rather than treating an AI agent as an unrestricted software process, organisations can establish verifiable boundaries around its activities.

TS Wong, Group Managing Director of Zetrix AI, said the initiative aims to connect agent autonomy with accountability by making identities, permissions and actions traceable to responsible owners.

Verifiable Credentials and Agentic Wallet Technology

The KYA service builds on the W3C Verifiable Credentials framework and incorporates AvatarLabs’ Agentic Wallet technology.

Verifiable credentials provide a standards-based mechanism for issuing and presenting digitally signed claims that can be checked by another party. In the context of KYA, this approach is intended to establish verifiable relationships between an AI agent, its owner and its authorised activities.

The Agentic Wallet component is designed to manage digital credentials, verify delegated permissions and maintain evidence of agent activity.

An important capability involves controlling how authority is transferred when one AI agent delegates work to another. Under the proposed model, a delegated agent should not receive permissions exceeding those granted to the original agent.

Permission revocation is another important component. If an agent’s access is no longer required or its authority needs to be withdrawn, the framework is designed to support revocation of its permissions.

Signed activity records can also provide an auditable trail, helping organisations investigate what an agent did and the authority under which it acted.

These mechanisms address several challenges associated with autonomous enterprise systems, particularly identity management, excessive permissions, agent-to-agent delegation and accountability.

However, verifiable identity alone does not guarantee that an AI agent will behave safely or make correct decisions. Effective deployment will also require security controls, policy enforcement, monitoring and appropriate human oversight.

Cross-Border Pilot Targets Interoperable AI Identity

Zetrix AI and CAICT are planning a cross-border pilot involving organisations from the Middle East, Africa, China and ASEAN.

The initiative will examine whether AI agent identities and delegated permissions can be recognised across different national digital identity ecosystems while preserving local control over identity systems and trust decisions.

Potential applications include cross-border trade, digital payments, government services and interactions between autonomous AI agents.

This presents a significant technical and regulatory challenge. An AI agent authorised to perform a transaction in one jurisdiction may need to demonstrate its identity and permissions to a service provider operating under a different regulatory framework.

The pilot is expected to evaluate identity verification, permission enforcement, revocation mechanisms, privacy safeguards and accountability records in practical cross-border environments.

According to CAICT’s Xiaoyu YOU, Head of International Department, ASTRON network, the initiative aims to support a cross-border trusted network that enables interoperable Agentic AI.

The participating organisations will need to demonstrate that interoperability can work without requiring every jurisdiction to surrender control over its own identity infrastructure or regulatory decisions.

Initiative Targets International AI Standards Development

Beyond the initial service launch, the partnership is intended to contribute to international discussions around AI identity, trust and accountability.

Zetrix AI and CAICT plan to submit findings from the cross-border pilot to the International Telecommunication Union (ITU), particularly its Telecommunication Standardization Sector Study Group 17 (ITU-T SG17), through the Focus Group on Trust and Identity for Humans and Agentic AI (FG-TIDA).

The proposed submissions are expected to cover implementation findings, cross-border use cases and recommendations involving agent identity verification, delegated authority and accountability mechanisms.

The partners also intend to contribute to discussions involving IEEE and other standards development organisations. These submissions are intended to inform ongoing technical discussions and pre-standardisation activities rather than establish that a globally adopted AI identity standard already exists.

This distinction matters because the industry has yet to converge on a universal framework for identifying, authorising and auditing autonomous agents across different enterprise platforms and jurisdictions.

AI Agent Identity Could Become Core Enterprise Infrastructure

The KYA initiative reflects a broader shift in enterprise AI architecture. As organisations introduce agents capable of interacting with applications, accessing sensitive information and initiating transactions, conventional user authentication may not be sufficient to govern every automated interaction.

Enterprises will increasingly need to establish which agents are operating, who owns them, what permissions they possess and whether their actions remain within approved boundaries.

Technology providers including Microsoft, Google, Salesforce and other enterprise AI platform developers are expanding agent capabilities across business applications. This creates an opportunity for identity and governance infrastructure that can operate across different platforms rather than remain restricted to individual AI ecosystems.

Zetrix AI and CAICT are positioning KYA as a potential foundation for this emerging requirement, particularly where agents interact across organisational and national boundaries.

The commercial significance will depend on whether the framework can demonstrate reliable credential verification, secure delegation, effective revocation and interoperability in real-world deployments.

For now, the cross-border pilot represents an important next step toward testing these requirements. If successful, it could contribute to a more consistent approach to establishing trust in autonomous digital interactions.

The underlying challenge is no longer simply enabling AI agents to perform tasks. It is ensuring that every agent operates with verifiable authority, accountable ownership and traceable permissions.

Market Landscape

The KYA launch comes as the enterprise AI industry moves toward autonomous systems capable of accessing applications, exchanging information and executing business transactions.

Three technology categories are becoming increasingly interconnected:

• AI identity and authentication: Organisations need mechanisms to distinguish legitimate AI agents from unauthorised systems and establish ownership.
• Agentic AI governance: Delegated permissions, access controls, activity monitoring and revocation mechanisms are becoming important requirements for autonomous enterprise workflows.
• Interoperable digital identity: Cross-border AI interactions require trusted credential verification that can operate across different platforms, organisations and regulatory environments.

The competitive landscape includes several adjacent technology ecosystems:

These platforms address different aspects of identity, authentication and AI governance rather than offering identical KYA services.

Zetrix AI and CAICT’s initiative introduces an additional focus: portable, verifiable identities and delegated authority for AI agents operating across organisational and national boundaries.

The main commercial opportunity lies in establishing trust infrastructure that allows autonomous agents to interact without requiring organisations to relinquish control over identity, permissions and accountability.

Top Insights

• Zetrix AI and CAICT launched a Know Your Agent service that assigns verifiable digital identities and delegated permissions to autonomous AI agents.
• The framework uses W3C Verifiable Credentials and AvatarLabs’ Agentic Wallet technology for credential management and permission verification.
• Organisations can define agent access boundaries covering services, information, transaction counterparties and spending limits.
• A planned cross-border pilot will evaluate interoperability across China, ASEAN, the Middle East and Africa.
• The initiative aims to contribute implementation findings and recommendations to ITU, IEEE and other international standards discussions.

Power Tomorrow’s Intelligence — Build It with TechEdgeAI