The 16% problem: AI readiness in Gulf companies is a data question

Gulf leaders are among the most confident in the world about AI. Their data is among the least ready for it. Closing that gap is a data-first decision, not a technology purchase.

Ask a room of Gulf executives whether their company is ready for AI, and most hands go up. Ask whether their AI can read the company’s own contracts, price lists and board papers, and the room goes quiet.

That quiet now has a number. In PwC’s 29th Global CEO Survey, published in January, 82% of Middle East chief executives said their culture enables AI adoption, and 70% said they had a clearly defined AI roadmap, against 51% globally. Yet only 16% of CEOs in the GCC agreed that their most-used AI tools can access all the relevant documents and data. The global figure, modest as it is, stands at 29%.

High confidence, low data access
Middle East CEOs: culture enables AI 82%
Middle East CEOs: clear AI roadmap 70% Global 51%
GCC CEOs: AI tools reach all relevant data 16% Global 29%

Source: PwC, 29th Global CEO Survey: Middle East findings (January 2026)

Put those numbers side by side and the region’s AI problem turns upside down. It isn’t a shortage of will, budget or appetite. The AI simply can’t see the business it was bought to serve.

Readiness has been measured in the wrong place

For three years, “AI readiness” has mostly meant three questions. Do we have the tools? Do we have the people? Do we have the budget? Gulf companies answer yes faster than most: licences get bought, teams get trained, pilots get announced.

None of those questions asks whether the company’s own information is fit to be read by a machine. That is the question that decides whether anything useful comes out the other end.

Gartner put the cost plainly in February 2025: through 2026, organisations will abandon 60% of AI projects that aren’t supported by AI-ready data. In the same research, 63% of organisations said they either lacked the right data management practices for AI or didn’t know whether they had them. The model is rarely what breaks. The ground under it is.

What “not ready” looks like on an ordinary Tuesday

Data problems sound abstract until one of them costs you a week. Three patterns come up again and again in mid-sized and large companies in the UAE and Saudi Arabia.

The number with three definitions. Finance reports revenue net of discounts. Sales reports it at list price. The board pack uses a third figure that includes a one-off contract. Each is defensible on its own. An AI that reads all three will answer with total confidence, and give a different answer each time someone asks.

The file nobody owns. The approved price list, the current organisation chart, the latest version of the standard contract: each lives in four places, and nobody can say which one is final. People work around it by asking a colleague. An AI can’t ask a colleague.

The access that doesn’t match the organisation chart. Shared drives built up over years give some people too much and others too little. Put AI on top and it inherits the problem: it can surface a salary file to the wrong person, or miss the document the right person needs.

None of this is new. What is new is that AI turns a quiet inefficiency into a visible, repeated and confident error.

Six questions to ask before the next AI licence

Defined properly, AI readiness is a data question, and it breaks into six checks a leadership team can run in a week without buying anything.

  1. Governance. Is there one approved home for each type of document, and does everyone know where it is?
  2. Lineage. Can any figure in the board pack be traced back to a document and a page?
  3. Ownership. Does every critical file have a named person who keeps it current?
  4. Quality. Are outdated and duplicate versions removed, or simply left to pile up?
  5. Metrics. Is there one written definition for each figure the company runs on?
  6. Trust. Would an AI see exactly what each employee is allowed to see, and nothing more?

A company that answers yes to most of these is ready for AI that does real work. A company that can’t will get AI that sounds right and isn’t.

Why this matters more in 2027 than it did in 2024

Here is the shift worth planning for. AI capability is becoming something every company can rent on roughly the same terms. Next year, your competitor’s AI will be as capable as yours, priced about the same and running in the same cloud regions.

What no one can rent is your company’s own information, organised so a machine can use it. That asset compounds. Every document that gets an owner and every figure that gets a single definition makes the next AI project cheaper and faster to deliver. Companies that skip the work pay the full cost again with every new pilot.

In the Gulf, the gap between those two groups is likely to be wide, precisely because confidence is so high. Leaders who believe they are ready will buy faster. Leaders who check their data first will spend less and get more out of it.

A data-first decision

The practical move is unglamorous. Before the next licence, pilot or agent, spend two weeks finding out where your data actually stands: what exists, who owns it, where the definitions disagree, and what an AI would see if you switched it on tomorrow.

That is the principle Unknot Technology Labs Ltd is built on: govern the data first, then put AI to work on it. We don’t install AI on data that cannot support it. Once the data is ready, the engine takes on the work that slows a company down and delivers it complete, every source shown, ready for approval.

The 16% figure won’t stay where it is. The question for every leadership team in the region is whether they will be among the companies that move it, or among those still wondering why the pilots stalled.

Ahmed A. H. Hamid is the founder of Unknot Technology Labs Ltd at the DIFC Innovation Hub in Dubai. He spent more than eight years in business development and strategy consulting with medium and large enterprises and professional-services firms in the UAE and Saudi Arabia.

Sources: PwC, 29th Global CEO Survey: Middle East findings (January 2026); Gartner, Lack of AI-Ready Data Puts AI Projects at Risk (February 2025).

See it on ten of your own documents.

A 45-minute session at your office. Bring contracts, policies or price lists; leave with answers, finished work and their sources.

We reply within one business day.