Artificial intelligence spending is accelerating across the Gulf, with UAE AI spending representing a significant portion of this growth. The harder part is turning that money into companies that actually operate differently.
Organizations in the UAE increased AI spending by 105% year over year, while Saudi Arabian organizations pushed spending up by 124%, according to ServiceNow’s 2026 Enterprise AI Maturity Index.
Those are huge increases.
The maturity scores tell a less comfortable story.
UAE organizations scored just 48 out of 100 on ServiceNow’s AI maturity scale. Saudi Arabia came in slightly higher at 50 out of 100. In other words, investment is moving much faster than execution.
UAE AI Spending Is Growing Faster Than Its AI Infrastructure
There is no shortage of ambition in the UAE.
Government-backed AI programs, new infrastructure projects, investment funds and aggressive enterprise adoption have made the country one of the most closely watched AI markets outside the United States and China.
Inside individual companies, though, things are messier.
Only 14% of UAE organizations surveyed had replaced legacy systems with platforms capable of integrating AI effectively. Saudi Arabia was barely different at 13%.
That matters because adding increasingly capable AI models on top of disconnected databases and aging enterprise software doesn’t magically create an AI-native business.
Sometimes it just creates a faster version of the same fragmented workflow.
ServiceNow’s broader global research reached a similar conclusion. Only 16% of organizations surveyed globally had replaced fragmented legacy systems with an integrated IT foundation, despite AI spending increasing dramatically.
Agentic AI Is Arriving. Autonomous Workflows Aren’t.
The numbers around agentic AI are particularly revealing.
Around 57% of UAE organizations and 48% of Saudi organizations surveyed said they had implemented agentic AI. Sounds advanced. Look one layer deeper and adoption becomes much less impressive.
Only 7% of UAE organizations and 10% of Saudi organizations were using agentic AI to create autonomous workflows. Most deployments are still closer to AI-assisted work than genuinely autonomous business processes.
An employee might use AI to summarize a document, generate code, respond to customers or search internal information faster. That’s useful. It can save hours. It isn’t the same thing as an AI agent independently coordinating several systems, making decisions and completing an entire workflow.
That gap is becoming important as “agentic AI” turns into one of the biggest enterprise technology pitches of 2026. Companies may technically have agents. Far fewer have rebuilt their operations around them.
Data Is Becoming the Uncomfortable Part of the AI Story
Buying another AI platform is relatively straightforward.
Fixing ten years of inconsistent corporate data is not. In the UAE, 77% of executives surveyed identified problems with data accuracy, accessibility and management as a major obstacle to expanding AI. The figure was 67% in Saudi Arabia.
ServiceNow’s global findings show the problem isn’t unique to the Gulf. Across its wider survey, 71% of organizations reported struggling with data accuracy, access and management. Generative AI made it possible for companies to experiment without rebuilding everything underneath. Enterprise-scale AI is starting to expose the consequences of that shortcut.
An AI agent that can take action across a company needs dependable information. It needs to know which customer record is correct, which system owns a transaction, who can approve an action and what data it’s allowed to access. Without that foundation, autonomy starts looking less exciting.
AI Governance Is Lagging Too
Another weak point is governance. Only 16% of UAE organizations and 18% of Saudi organizations surveyed had implemented AI testing, auditing and risk-management processes. That’s a surprisingly small number considering where enterprise AI is heading.
AI tools are no longer limited to generating text inside isolated chat windows. Agents are increasingly being designed to interact with databases, business applications, customer records and financial processes.
The more authority companies give those systems, the harder governance becomes to postpone. That issue is already visible in regulated sectors, where deployments such as ruya Bank’s agentic AI systems are being built around human oversight, data controls and common governance layers.
Globally, ServiceNow found only 26% of organizations had systems in place to manage AI governance and compliance. The technology is becoming autonomous faster than many companies are becoming comfortable supervising it.
The UAE’s AI Push Has Created a Head Start — Not a Finished Transformation
None of this means the UAE’s AI strategy is failing. Quite the opposite. Its government-led AI push, investment environment and regulatory activity have given companies an unusually strong starting position. UAE AI maturity also increased 13% year over year, while Saudi Arabia improved by 17%.
The country is also investing heavily in the infrastructure required to support enterprise deployment, including local AI inference and data-residency options for organizations with sovereignty and compliance requirements. But national AI ambition and enterprise AI maturity aren’t the same thing.
A country can build data centers, attract AI companies and fund enormous technology projects while businesses underneath that ecosystem are still figuring out how to connect an AI agent to an old ERP system. That’s roughly where the market appears to be now.
AI Could Soon Consume Nearly One-Fifth of IT Budgets
The spending surge probably isn’t finished. Executives surveyed in both Saudi Arabia and the UAE expect AI to represent almost 20% of total IT budgets by 2027. That creates an interesting question.
What happens if AI budgets keep expanding but the infrastructure underneath them doesn’t? There is a limit to how much value companies can extract by adding another model, copilot or agent to disconnected processes.
ServiceNow’s highest-performing global organizations — which it describes as AI “Pacesetters” — offer a glimpse of the other possibility. These organizations are generating an average 160% return on AI investment, with that figure projected to reach 194%. Their advantage isn’t simply spending more.
They have done more of the boring work. Integrated systems. Better data. Cross-functional workflows. Governance. Clear responsibility for what AI is allowed to do. None of those make for particularly dramatic AI announcements. They may end up deciding who actually wins.
The Next Gulf AI Race May Be About Execution
For the past few years, the AI race in the Middle East has largely been described through investment announcements, semiconductor access, data centers, government strategies and partnerships with major technology companies. That phase isn’t over.
But another race is beginning inside companies. Which businesses can move from isolated AI tools to connected workflows? Data cleanup will also become a major test. Leaders must decide how much authority AI agents should have without losing control over them. Finally, companies still need to replace the legacy systems they have been promising to fix for years.
The UAE and Saudi Arabia have already demonstrated that they’re prepared to spend aggressively on artificial intelligence. The UAE is also pushing AI deeper into the infrastructure businesses already use, including agentic AI bundled directly into business connectivity services. Now comes the less glamorous test. Making all of it work.
