Gulf companies ‘struggling’ to turn AI investment into returns
AI adoption is accelerating across the region, but poor data, fragmented systems and weak governance are emerging as major barriers to turning investment into business results
27 August, 2026
TT
16
Around 84 per cent of organisations across the Gulf are already using artificial intelligence in some form, but just 11 per cent have scaled deployments far enough to generate measurable returns, according to McKinsey and the GCC Board Directors Institute.
The findings underline a growing challenge for companies investing heavily in AI: the technology is only as reliable as the data and systems behind it.
Accelera Digital Group (ADG), a technology firm operating across Africa, the Middle East and Europe, says many Gulf organisations are still grappling with legacy data silos, inconsistent governance and fragmented cloud environments — creating a hidden cost it describes as the “silent AI tax”.
Meanwhile, Google Cloud’s ‘The ROI of Gen AI’ report found that 74 per cent of organisations experimenting with AI were already seeing some return. But ADG argues there is a significant difference between generating early gains and achieving meaningful returns at scale.
“Most businesses in this region have built AI into their strategy, with strong leadership behind it, but far fewer have asked whether their data can actually support it,” said Mohammed Ashoor, Bahrain country manager at ADG.
“That’s the silent tax. It shows up in stalled pilots, agents sitting idle because nobody trusts the data behind them, and budgets that underperform without anyone quite knowing why.”
The issue is becoming more pressing as Gulf companies increase spending on the technology.
McKinsey and the GCC Board Directors Institute found that 89 per cent of Gulf executives expect AI budgets to rise this year, while just 31 per cent of organisations have moved deployments beyond the pilot stage.
The stakes are also rising as businesses move beyond chatbots and generative AI tools towards AI agents capable of carrying out tasks and taking actions with less direct human supervision.
For regulated sectors including banking, insurance and government, that raises wider questions around data quality, compliance, security and auditability.
ADG argues that the underlying data architecture needs to be governed, traceable and secure before autonomous AI systems are given greater responsibility, rather than attempting to bolt compliance controls on after they have entered production.
Greater focus on AI access
But data quality is only part of the challenge.
Andreas Hassellof, CEO of technology company Ombori, argues that businesses must also set clear limits on what AI agents are allowed to access, change and spend — and establish who ultimately remains responsible for their actions.
“A person still owns the standard, the limits, and the decision to ship,” Hassellof said.
He argues that companies should define those boundaries before AI systems are introduced into real-world workflows, with particular care around payments, customer data and other actions that cannot easily be reversed.
There is also a broader question over the standards embedded within increasingly autonomous AI systems.
Most leading AI models are developed by a relatively small number of companies, principally in the US and China. Hassellof argues that organisations elsewhere should not automatically assume those systems will reflect their own regulatory, operational or institutional priorities.
“The standard should be yours,” Hassellof said.
“A borrowed model can run the loop. It should not, by default, become the source of what good, safe or finished means in a business, a hospital or a government.”
Instead, he argues that companies should retain control over the standards against which AI systems operate, including what they may read or change, what actions they may take and who has the authority to stop them.
That does not remove the need for human accountability as AI becomes more capable.
“A person should remain the responsible party,” Hassellof said. “Someone has to own the objective, the boundary and the result.”
For Gulf businesses, the debate is therefore increasingly moving beyond whether to invest in AI.
As spending rises and AI systems are entrusted with more complex work, the bigger question is whether the data, governance and controls surrounding them are strong enough to turn that investment into measurable returns.
