AI products move faster when product, infrastructure, and delivery live in one team.MVPs live in 90 days, full IP ownership. Start a project
← Back to blogs

05 Nov 2026

AI & Automation

RTA Upgrades Mahboub with Agentic AI: What Private UAE Companies Should Copy

Dubai RTA is preparing Mahboub to complete linked vehicle services (fines then registration) in one conversation using agentic AI, UAE PASS and Dubai Sadad. The same pattern works for private clinics, distributors and operators: one bounded multi-step workflow on the channel customers already use.

Solinify

· 7 min read

RTA Upgrades Mahboub with Agentic AI: What Private UAE Companies Should Copy

On 22 October 2026 Dubai’s Roads and Transport Authority announced that Mahboub, its virtual assistant, has been upgraded with agentic AI so customers can complete linked services inside a single conversation. The initial scope covers vehicle licensing: paying fines then renewing registration, issuing a clearance certificate, and renewing number-plate reservations. Identity is verified through UAE PASS and payment goes through Dubai Sadad. The system supports Arabic and English with models fine-tuned for local context. The upgraded version is being prepared for public rollout as part of RTA’s Artificial Intelligence Strategy 2030; no exact launch date has been given.

Later updates tied the work to a broader strategy that aims to move 50 percent of RTA operations to an agentic AI ecosystem within roughly two and a half years and generate more than AED 300 million in annual added value. Mahboub is expected to handle a rising share of services (targets of 10 percent by end-2026 and 50 percent by end-2027 in some reporting).

What “agentic” means in this case

Mahboub is moving from answering questions or triggering single services to coordinating multi-step procedures. The customer states the goal (“renew my registration and clear the fine”); the system activates specialised agents that execute the necessary steps in sequence, using the same identity and payment rails. The user does not switch pages or re-authenticate for each action.

That is the practical definition of an agent for most UAE private companies: not a general-purpose chatbot that can discuss anything, but a bounded system that can read live data, take a limited set of actions, and stop when it should hand off.

The private-sector version of the same pattern

Most clinics, trading companies, restaurants and distributors already have the equivalent of Mahboub’s channel: WhatsApp. Customers message the business number they already know. The equivalent of “linked services” is usually one high-volume job that currently requires staff to open two or three systems:

  • A clinic: check the next free slot for a specialty, confirm the patient is registered, book, and send the reminder — all from the inbound WhatsApp message.
  • A distributor: look up stock for the SKU the customer named, confirm the price for that customer’s tier, create the order draft, and hand to the salesperson if the credit limit is close.
  • A restaurant group: take a reservation or delivery order, check capacity or menu availability, and pass payment or special-request cases to a human.

The agent does not need to handle every request. It needs to finish the common path and stop cleanly on the exceptions. After 1 October 2026 every free-form reply inside the 24-hour window is a service message; the first 1,000 delivered service messages per business number per month remain free, then the UAE rate (approximately USD 0.0157 / AED 0.0576) applies. Designing for short, accurate replies is therefore part of the business case.

What private companies should copy, and what they should not

Copy the multi-step coordination and the use of existing identity/payment rails where they exist (UAE PASS is public; private companies use their own CRM or ERP as the source of truth). Do not wait for a public marketplace agent or a training course to do the integration. Chamber AI Hub and the various agentic AI training tracks make experimentation cheaper, but they do not automatically connect to a company’s live stock file or appointment calendar.

The sequence that has worked for operators we work with is short:

  1. Pick one job that already arrives on WhatsApp and already has a trusted record in the CRM or ERP.
  2. Define the fields the agent may read and the hard stop conditions (money above a threshold, clinical advice, complaint, low confidence, new customer).
  3. Measure first useful reply time, handoff rate and error rate for 30 days on one number or one branch.
  4. Expand only if the rejects are boring and the free-tier message volume is predictable.

Related reading on Solinify:

RTA is showing what multi-step agentic service looks like on a public channel with national identity and payment systems. Private companies can run a narrower version of the same idea on the WhatsApp number and the record they already own. The engineering is mostly the integration and the handoff rules, not another general chatbot.

Related Articles

Ready to build your next big thing?

Solinify Tech Studio specializes in building, launching, and scaling AI-powered products and startups. Let's turn your vision into a future-ready business.

Learn About Venture Studio→
← Back to all blogs