The Middle East has stopped debating whether to go digital and started asking whether it is working. That is progress, and it is also where the hard part begins.
The region has won the platform phase of digital transformation: the money, the infrastructure and the political will are in place. My argument is that the next phase is not a technology problem. It is an operating model problem, a skills problem and, increasingly, a trust problem.
I have spent 17 years marketing technology across Pakistan, the Gulf, the UK and the US, including leading a market entry into the UAE and Saudi Arabia. The pattern I would flag from that work is simple: a region that buys the platform first and designs the job second pays for it later.
The platform phase is won
Start with what the Gulf has got right. Saudi Arabia invested more than SAR 31.9bn ($8.5bn) in 2025 in communications and IT services to advance digital government, according to MIT Sloan Management Review Middle East. Citizens are ready for it. In BCG’s 2026 Digital Government Citizen Survey, 79% of respondents in Qatar, Saudi Arabia and the UAE expect government services to be AI-centred, against 66% globally. Only about 4% want human-only service.
The infrastructure has also been tested. IDC’s July 2026 analysis found that firms with structured multiregional failover recovered quickly during this year’s regional conflict, while those on single-region designs fared worst. The Gulf did not just build digital capacity. It showed that the capacity holds under pressure.
AI is being bought faster than it is being scaled
Here is the gap that deserves more attention than the spending headlines. Capgemini’s October 2026 research found that only 13% of UAE public sector organisations and 33% of Saudi ones report successfully scaling AI initiatives. The same report says success depends on operating models, workforce readiness and governance, not on technology spending alone.
Business tells the same story. Deloitte’s 2026 State of AI in the Enterprise report found that 84% of Middle East firms have not yet redesigned jobs or workflows to incorporate AI. Two-thirds report better efficiency, but only 20% say AI is contributing to revenue growth, and just 21% have mature governance for autonomous AI systems. Aditi Nitin, Deloitte Middle East’s AI and data leader, said “AI transformation is ultimately a human transformation”.
I agree, and I would put it more bluntly. A tool bought without a redesigned workflow is a licence fee, not a transformation. When a pilot stalls, the fault is usually in the meeting where nobody owns the process the tool was meant to change.
Trust is the asset being spent
Citizen appetite is high, but it is not unconditional. BCG found that confidence that AI’s benefits outweigh its risks fell by 21 points since 2024, although the GCC remains more confident than the global average. The top worries are job loss, at about a third of respondents, and the accuracy of results, at about 28%. People also say they want technical safeguards such as independent certification, testing and documentation, along with public reporting.
That is a clear brief for governments and vendors. If a service is automated, show how it was tested and who checks it. Trust is cheap to keep and expensive to rebuild.
The skills gap is the binding constraint
IDC found that 56% of tech buyers across the Middle East, Turkiye and Africa cite the skills gap as the region’s third-largest technology risk. The usual response is to train specialists: more data scientists, more cloud architects. That matters, but I think the larger gap sits one layer up. The department head, the operations lead and the marketing director decide whether an AI tool changes how work is done. If they do not understand what it can and cannot do, the platform sits idle.
AI literacy for ordinary managers, not only engineers, is the cheapest lever the region has not yet pulled at scale.
The strongest objection: this is just what an early build looks like
The fair counter is that no region scales AI in two years, and that the Gulf is moving faster than most. Capgemini itself describes regional adoption as “accelerating ahead of the global standards”, and nearly two in five Middle East public sector organisations are data front-runners, above the global benchmark. Deloitte finds 54% of firms expect at least 40% of their AI experiments to reach production within three to six months. Buying the platform first is how the Gulf got to move at this speed.
I accept all of that. My answer is that a 13% scaling rate in one of the region’s leading markets means most of the spending is waiting on process change that nobody has scheduled, and trust is slipping while it waits. Hoping the gap closes by itself is the one plan I would not fund.
What I would do
- Governments and large firms: report scaling, not spend. Publish the share of services where AI is in production and the measurable outcome it delivered. Budgets announced are not value delivered.
- Chief executives: redesign the job before buying the tool. Tie every AI budget line to a named workflow and a named owner, with the old process retired on a set date.
- Regulators and vendors: publish the assurance. Independent testing, documented limits and public reporting are what citizens say they want. Offer them before they are demanded.
- Marketing and HR leaders: teach managers, not only specialists. Make AI literacy a requirement for people who run teams, and stop selling the model: sell the redesigned workflow, the owner and the measured result.
The platform was the easy part, and the region has built it well. The people are now the project.








