This week, much of the region’s technology community has gathered in Abu Dhabi for AI Everything, and the question on the show floor has moved on. A year ago, banks and insurers asked whether to deploy AI agents; today they want to know how to make those agents deliver results across an entire customer journey.

Government is setting the pace. On 5 October, the Dubai Future Foundation launched an agentic AI accelerator covering 33 government entities and more than 300 public-facing use cases, building on the federal target to run half of government services and operations on agentic AI within two years. Few countries have put a number and a deadline on agentic AI. The UAE has done both.

Finance is keeping up. The Dubai Financial Services Authority’s survey of 661 DIFC firms found that 52% now use AI, up from 33% a year earlier, while generative AI use rose by 166%; the regulator also noted that, for many firms, adoption is still at an early stage. UAE finance has AI in many places. The value comes when those pieces work together.

Work in financial services is fragmented

Follow a corporate account opening at a typical UAE bank. The application lands in a portal or a relationship manager’s inbox; documents and identity are checked, sanctions screening runs in a separate tool, a risk team reviews the file, a committee signs it off and operations sets up the account in core banking. That’s six steps, six systems and one waiting customer.

AI agents speed up individual steps: an agent can extract a trade licence in seconds or draft a risk summary in minutes. Then the file sits in a queue for a reviewer who has no idea it is ready. The task gets faster, while the journey stays as slow as its slowest handoff.

That gap is the coordination tax, made up of the hours lost chasing approvals, re-keying data and fielding “where is my application?” calls. In banking, insurance and payments, where one journey can cross half a dozen teams, it is often the largest cost nobody actually measures.

Process is the foundation

Institutions that turn AI into outcomes start by understanding how work actually runs, and regulated firms have a head start here because they already document their processes for supervisors. The task is to map those processes as they really operate (workarounds and exceptions included) and give each step to the right worker.

Steps that must run identically every time, such as sanctions checks and payment releases, suit rules-based workflow. Reading documents and reasoning over unusual cases suit AI agents. Judgement belongs to people. With the Central Bank’s AI guidance for the sector placing clear emphasis on human oversight, deciding where people sit in a process is now part of designing it.

Orchestration makes AI operational

Once the process is clear, orchestration connects it. It coordinates people, systems, data and agents across a whole journey, deciding who or what acts next and routing exceptions to the right person before they become delays.

Think of an airport, where more aircraft add capacity only when air traffic control sequences them. AI agents work in the same way.

For a bank or insurer, orchestration brings one view of every application, claim or payment. Core banking and policy platforms keep their role while work flows between them, and each orchestrated journey becomes a template for the next, so the tenth deployment goes live faster than the first.

A 90-day starting point

Finance can borrow the government’s 90-day rhythm. Pick one journey that crosses at least three systems and two departments, such as SME lending or a motor claim, and map it as it runs today. Count the handoffs and measure the time lost between them; that figure is your coordination tax. Then assign each step to a rule, an agent or a person, orchestrate the whole journey and track time to decision. Reuse what works.

The orchestrators will lead

At the accelerator launch, Digital Dubai’s Director General said the future of digital government depends on an integrated ecosystem that brings expertise, data and capabilities together. The same holds inside every bank.

Over the next year, the number of AI agent use cases across UAE banking and government will grow rapidly. The institutions that pull ahead will be those with a platform to orchestrate and govern those agents, taking them from pilot to production at scale, while keeping people in control. In the AI era, operational advantage belongs to the orchestrators.