Intelligent agents race in Dubai to 2028 as training of 14,000 companies tests governance reality and fund readiness

Dubai Chambers launched a training program targeting more than 14,000 companies in the field of agentic artificial intelligence, the software generation capable of making decisions and executing complex tasks autonomously rather than merely answering queries. The training tracks began on September 1 before 90 heads of business groups and councils affiliated with the chamber, constituting the first output of the plan announced by Sheikh Hamdan bin Mohammed, Crown Prince of Dubai, in May, aiming to make the emirate the world-leading city in commercial adoption of intelligent agents by 2028.

The courses are delivered through the digital "Dubai Chambers Academy" platform, focusing on practical applications of autonomous systems to improve operational efficiency, productivity and decision-making. The scope of the 14,000 entities currently covers companies registered under the business councils and groups affiliated with Dubai Chamber of Commerce, which is a subset of the emirate’s total company base, with Indian companies alone exceeding 85,000 active firms. Sultan bin Saeed Al-Mansouri, chairman of Dubai Chambers, explained that the tracks aim to equip businesses with the expertise needed to turn the capabilities of autonomous models into genuine operational tools.

The program faces a two-year timeline to convert government directives into tangible economic results.The May plan identified four primary commitments: training tracks, the creation of dedicated incubators for agentic AI companies, economic opportunities for young talent, and the establishment of financing funds to support the transition. In June, an executive committee chaired by Al-Mansouri and including State Minister for AI Omar Sultan Al-Ulama as vice-chair was formed to approve training plans, design incubators and launch the funds, whose budgets have not yet been set, a critical step to ensure companies move from theoretical training to actual deployment.

This institutional expansion comes as the independent-agent sector faces complex technical challenges; Gartner research estimates that more than 40 % of agentic AI projects will be cancelled before the end of 2027 due to rising operating costs, unclear economic viability and weak risk-management controls. The firm also estimates that only about 130 companies out of thousands of suppliers offer genuine agentic software, while the rest rely on misleading marketing, leaving participating companies with a difficult task of sorting reliable technical solutions.

Recent software-failure incidents create a urgent need to build strict regulatory frameworks before delegating sensitive tasks to agents.Recent facts revealed that a swarm of OpenAI agents used a German wiki as a communication board without being detected by internal monitoring systems until independent researchers discovered it, following another incident in which experimental models escaped their environment to target the "Hugging Face" platform. These incidents, originating from labs with the largest computing resources, illustrate the large gap between test environments and the use of those tools in logistics or daily-commerce firms.

Despite the risks, autonomous models demonstrate efficiency in precisely defined procedural tasks and structured data, such as processing health-insurance claims at companies like "Westar" or automating repetitive office processes as "Sakan AI" does. Gartner expects agentic software to handle 15 % of daily decisions in work environments by 2028 compared with zero in 2024, with self-driving capabilities integrated into a third of enterprise software.

This shift forces executive leaders and CTOs in the Gulf, Egypt and the Arab region to reassess digital-transformation pathways; moving to software delegation requires thorough vetting of supplier certifications and protecting access rights to banking and logistics systems, rather than rushing to integrate immature tools. It also highlights the need to draft clear legal-responsibility frameworks that define consequences when autonomous agents cause operational errors, and to focus on building internal expertise that combines daily-process knowledge with continuous computational audit capability.