DUBAI PUT A CLOCK ON AI.

WHAT HAPPENED

On 4 May 2026, Sheikh Hamdan bin Mohammed, Dubai's Crown Prince, directed the emirate's entire private sector to transition to agentic AI within two years, which puts the deadline in May 2028. In June, the Higher Committee for Future Technology Development and the Digital Economy, which he chairs, set out the plan: roughly 295,000 companies in scope, 100 specialized AI assistants to be delivered over the two-year window, government-backed incubators and funds administered through Dubai Chambers, and a target of seeding 50 agentic-AI firms. The federal government has its own clock: on 23 April 2026 the UAE Cabinet unveiled a framework to deploy agentic AI across 50% of government sectors and operations within two years.

The Dubai directive comes with a deadline, government-backed incubators and funds, and Dubai Chambers to run it, which makes it harder to file away than the usual strategy document.

AGENTS RUN WORKFLOWS

The directive names agentic AI specifically. A chatbot answers questions. An agent runs a workflow: it takes an inbound lead, qualifies it, books the viewing and writes the CRM record without a person driving each step. The difference matters because the economics only work when a system removes work. A chat window added on top of the old process removes none.

THE CONSTRAINT IS READINESS

Three things hold for a Dubai business today:

• The direction is no longer a debate. Whether adoption is eventually enforced, incentivized or simply expected by partners and lenders, "we'll wait and see" now carries its own risk premium.
• The support is real. Dubai Chambers opened agentic AI training tracks in September for the 14,000-plus companies in its business groups and councils, and the program promises incubators and dedicated funds. Early movers use this support, and late movers compete with peers who already did.
• The readiness gap is the actual problem. In Stanford's study of 51 successful enterprise deployments, only about 6% had data that was fully ready for AI, and in 77% of them the hardest problems were process, data and change management. The model was rarely the hard part. A deadline shortens the timeline, and the amount of work stays the same.

THE FIRST 90 DAYS

Buying a platform in a hurry is the common mistake. The approach the evidence supports is boring, and it works: pick two or three workflows, run the numbers on where AI actually pays back, fix the broken process first, then put one agent into production on live data with KPIs attached. Those steps are the first three phases of our loop (audit, design, implement), and operate and optimize take over once the agent is live. Work in that order and you reach the deadline with results you can measure.

The clock is set either way.

As of 21 October 2026: the private-sector deadline is May 2028, and the federal target falls in April 2028. Since this article first ran, Dubai Chambers has opened agentic AI training tracks for the 14,000-plus member companies of its business groups and councils (1 September), and the federal government has said every federal AI agent will connect to a control tower that logs its actions (October). A national committee is also drafting a classification of which tasks an agent may execute on its own and which it may only recommend; see what a machine may decide.