KYA Debuts in the Middle East: AI Agent Identity Becomes Trust Infrastructure

From Dubai to Wuzhen, agent identity verification is moving from concept to commercial deployment, even as regulator data exposes the AI trust gap inside the Gulf’s financial hub.

On November 5, Malaysian-listed Zetrix AI and the China Academy of Information and Communications Technology (CAICT) simultaneously launched the Know Your Agent (KYA) service at the World Internet Conference in Wuzhen and the ICT Development Forum 2026 in Dubai, a commercial framework that gives AI agents verifiable digital identities and defined permission boundaries. This is the first KYA product with a deployment footprint spanning four regions, following the concept’s introduction into the HKMA’s agentic finance discussion during Hong Kong FinTech Week.

The same day, the Dubai Financial Services Authority (DFSA) released its 2026 AI survey, adding a reality check to the launch: AI adoption among DIFC-licensed firms has reached 71%, yet only 14% have deployed AI agents. The speed of capability adoption and the gap in trust infrastructure arrived in the same week.

1. How the KYA Service Works

The KYA service is built on the W3C Verifiable Credentials standard and integrates AvatarLabs’ Agentic Wallet technology. Its core mechanism is not a static ID card for agents, but a binding of three dimensions, identity, authorization, and behavior. Businesses or individuals can set which services an agent may access, which information it may use, which counterparties it may transact with, and a spending ceiling; operations above the threshold must route to human approval. Permissions can be revoked at any time, and an agent delegating subtasks cannot grant authority beyond what was originally assigned.

The design has a concrete use case. A business can authorize an AI agent to pay approved suppliers within a set limit, while transactions above the limit require human confirmation. This echoes the KYA guardrail that the HKMA embedded in its Project SAGE proposal last week: identity that can be confirmed, performance that can be verified, and permissions that can be constrained are becoming the shared precondition for agents moving from technical demos to commercial delegation.

2. The Trust Gap, in Regulator Data

The DFSA’s third annual AI survey, released November 2, shows DIFC AI adoption has doubled to 71% in two years, with generative AI usage rising from 22% in 2024 to 67%. But adoption breadth does not equal deployment depth: 53% of respondents are exploring or piloting AI agents, only 14% have reached deployment, and 74% of adopters run AI applications in only a small part of their business.

Governance is equally behind. Only 61% of adopters have an AI governance framework and 41% keep a register of AI use cases. A joint paper published October 30 by the DFSA and the Cambridge Digital Innovation and Regulation Initiative (C:>DIR) breaks agentic AI cyber-resilience into four dimensions, speed, reach, authority, and trust, with the authority dimension pointing directly at “whether the outcomes of actions by AI-enabled systems remain within their intended mandates.”

The paper does not name KYA, but it describes exactly the problem KYA solves. Regulators put permissions and accountability at the top of the risk list; identity vendors turn permission checks into a product. The two lines converge in the same month.

3. The Two Faces of Business and Risk

Zetrix AI’s timing for launching KYA is delicate. The company has spent RM3.7 billion (about $850 million) on development for its blockchain and AI products, more than half within the past 18 months, including RM900 million in the first half of this year alone. Its share price collapse triggered forced liquidation of pledged shares from nearly 30% in August to 1.84% by October 21, breaching a covenant under its RM2 billion sukuk program; MARC Ratings downgraded the program to C on October 30.

Putting the KYA launch next to the financial distress makes one question hard to avoid: when a company itself mired in debt crisis sells “trusted identity,” there is tension between the product’s value and the issuer’s credit. But that does not invalidate the direction of the KYA product itself. The trust gap it targets is real. The DIFC data, Hong Kong’s SAGE project, and the HKU KYA research group’s report released November 3 (76% of surveyed practitioners encountered agent-trust-related concerns in the past year) all point to the same conclusion.

4. Takeaways for Practitioners

The substantive signal of this launch is that agent identity certification now has a deployable commercial form, embedded directly in the Middle East’s most active financial sandbox. Louvre Abu Dhabi will host the first Gulf edition of Proof of Talk on December 3-4, with CZ, Franklin Templeton CEO Jenny Johnson, and BlackRock’s Middle East digital assets head Tony Ashraf in attendance; the StableDay track is dedicated to stablecoins and tokenized money. Identity infrastructure like KYA is likely to be a default premise in the next round of conversations.

5. Conclusion

The KYA launch and the DFSA’s trust data, arriving in the same week, pushed “agent identity” from concept to the intersection of product and regulation. The Gulf financial hub is proving with its 71% adoption rate that AI has become the default tool, and reminding everyone with its 14% deployment rate that before agents can truly take over transactions, identity, authorization, and accountability must first be institutionalized. Whoever completes this infrastructure first holds the first batch of entry tickets to the machine-native economy.