UAE companies and organisations have some of the highest rates of agentic artificial intelligence adoption in the world, a survey has shown.
The report – 2026 Global AI Confessions Report, CIO Edition – was commissioned by data science and AI platform company Dataiku and is based on the responses from 685 chief information officers in the UAE, US, UK, France, Germany, Japan, South Korea and Singapore.
It found that 62 per cent of UAE chief information officers say their companies have more than 50 AI agents working in the organisation, while 15 per cent say they operate more than 500 agents.
"That's the highest proportion across the eight markets surveyed, and substantially higher than the 9 per cent global average," the report said.
In recent years, AI agents – digital personal assistants powered by artificial intelligence to carry out repetitive tasks – have become increasingly prevalent across various sectors.
Technology companies have claimed that AI agents can match and often outperform humans in many functions, freeing employees to do other tasks.
Dataiku's analysis also showed that businesses and organisations in the UAE were continuing to embrace agentic AI.
"Among CIOs whose organisations are creating or planning embedded roles for AI experts or forward engineers, 22 per cent expect to recruit more than 10 people over the coming year, the highest compared to other countries in the study," it said.
Room for improvement
The analysis also highlighted areas in which UAE businesses and organisations can improve how AI agents are monitored.
This comes after incidents in which AI models escaped containment and compromised the digital infrastructure of outside entities.
"Only 5 per cent of UAE CIOs say their organisation could reliably identify and contain a problematic agent across environments within one to two hours, among the lowest results in the study and half the worldwide figure of 10 per cent," the survey read.
The poll also indicated that 27 per cent of the UAE chief information officers surveyed said that they could intervene later in the day to address a hypothetical issue with a rogue AI agent, meaning that two thirds of those surveyed felt they "could not contain the agent on the same day".
Kurt Muehmel, head of AI strategy at Dataiku, said that perspective is needed in looking at the results from the survey.
"UAE organisations already do one important part of this well, with 80 per cent of UAE CIOs having seen an agent act against business intent or policy while still operating within its technical parameters, but 62 per cent say they caught the problem before it caused harm, the best result globally," Mr Muehmel said.
The results of the survey can provide a road map for making the UAE's approach to AI more robust.
"Keep one inventory of every agent, whoever built it and whatever platform it runs on," Mr Muehmel said, when asked for steps that can be taken by businesses and organisations in the UAE.
He said an owner should be designated for every agent and that a decision must be made in advance about who would have the authority to stop it.
"Standardise how agents are tested, changed and retired ... and practise containment the way you practise a cyber security incident response."
Sid Bhatia, Dataiku's vice president for the Middle East, Turkey and Africa, said that overall, the latest analysis bodes well for the UAE's push to be a leader in AI and making use of AI agents.
"The UAE has established a strong foundation for agentic AI, with enterprises deploying at scale, investing in specialist expertise, and preventing a high proportion of agent incidents from causing real-world harm," Mr Bhatia said.
