UAE moves agentic AI from answering questions to closing transactions

The UAE's government services have entered a phase where the software does more than answer. Systems are beginning to execute transactions, coordinate procedures across entities and anticipate what a customer needs. Agentic AI is the label attached to that shift, and it is now showing up in live services rather than demonstrations.

In practice the model lets a government system carry several steps inside a single interaction. Paperwork shrinks. Repetitive requests stop being repeated. Customers spend less time and effort. Some recurring services now complete on their own, with no application filed by the customer at all.

A number of agentic initiatives are already running. Others remain under development or are being readied for public launch. Across them, the design emphasis is on reducing what a customer has to do, while keeping safeguards in place for data protection, approvals, customer consent and human oversight where a decision needs one.

Abu Dhabi: renewal without a request

Abu Dhabi's TAMM platform carries the clearest example. Its AutoGov feature renews certain licences and medical appointments on its own, without waiting for an expiry date to pass or for the customer to start the process. The platform serves 3.8 million users and offers more than 1,150 government and private-sector services. It processes 95 percent of requests using AI and has closed more than 1.9 million service cases through conversational AI.

The Department of Government Enablement launched GovGPT in October, a bilingual generative assistant that acts as a knowledge agent for government employees. A month earlier, the Abu Dhabi Executive Council used an AI-powered meeting platform for the first time. It analysed proposals and documents before sessions and produced executive and real-time summaries.

Ajman: two departments, one transaction

Ajman announced in July that it had completed the UAE's first government transaction to renew a trade licence using agentic AI. The process notifies a customer before the licence expires and points them to an AI assistant inside the Ajman One app. Where the commercial tenancy contract also needs renewal, the assistant coordinates with the Department of Municipality and Planning, then finishes the licence renewal once the contract is renewed.

In October the emirate showed a new generation of integrated services built to finish procedures in one experience, without switching platforms or hunting for the next required step. Commercial tenancy contract renewal is integrated with trade licence renewal. An AI financial adviser gives employees and specialists access to financial information, legislation and documents. The Ajman Free Zones Authority launched OneClick, unveiled at AI Everything Abu Dhabi 2026, which walks an investor through business set-up in a single conversation: defining the business concept, recommending activities and packages, collecting information, verifying identity and compliance, processing documents, supporting electronic signing, taking payment, and submitting and tracking the application through to licence issuance.

DEWA and Dubai's roads authority

Dubai Electricity and Water Authority converted 16 of its 21 services into AI-enhanced autonomous services in under 90 days. That is 76 percent of its total. It said the remaining five would follow within two months. The converted set includes electricity and water activation; two services respond to technical reports, and thirteen are supported by a payment reconciliation framework.

Dubai's Roads and Transport Authority rebuilt its Mahboub virtual assistant on agentic AI. Mahboub coordinates and executes connected services inside one conversation, and can settle vehicle fines before renewing a registration. Its first phase covers vehicle registration renewal, fine payments, clearance certificate issuance and renewal of vehicle plate number reservations. Identity checks run through UAE PASS; payments through DubaiPay.

The national layer

The UAE Cabinet approved a framework in May 2026 to widen agentic AI use, with a target of moving 50 percent of government sectors, services and operations onto agentic models within two years. The national programme includes training 80,000 federal employees in agentic tools, and it announced specialised assistants for procurement, tax auditing, customer happiness and technical support.

On October 22, 2026, the government launched the Agentic AI Application Lab for Government Services. The lab helps ministries and entities pick priority services, then develop, test and approve solutions with customers, monitor experience and feedback, and judge readiness and service quality before launch and wider deployment. The National Committee for Agentic AI issued a unified guide in August 2026 for designing these experiences. Its principles: understand the customer's need, use government data so the same information is not requested twice, explain the steps in a transaction, obtain consent before acting, track the request to completion, and tell the customer what was done and what still needs them.

The pattern has reached meteorology. The National Centre of Meteorology announced in June the launch of the UAE's first agentic AI assistants in the field, and approved a roadmap to expand their use.

Interest is not confined to government. Among energy leaders, 62 percent rank AI among their top two priorities.

What limits the model now is less the capability of the agents than the wiring between institutions. A licence renewal that depends on a tenancy contract forces two departments into a single workflow, and that is a governance and data-sharing problem before it is a machine-learning one. The consent and human-oversight principles in the national guide mark where automation stops and a signature is still required. The next visible tests are already scheduled: DEWA's remaining five services, the wider rollout of Ajman's integrated services, and the lab's readiness assessments ahead of each launch.