Gulf firms face widening AI returns gap

Gulf companies are pouring money into artificial intelligence, but many are finding that rapid adoption is not translating into measurable financial returns as weak data foundations, fragmented technology systems and governance gaps constrain deployment at scale.

Around 84% of organisations across the Gulf are using AI in some form, yet only about 11% have reached a level where scaled deployments are producing clearly measurable value. The gap is sharpening scrutiny of spending decisions as executives move beyond experimentation and demand evidence that AI can improve revenue, margins, productivity or customer outcomes.

The problem is becoming more pronounced as investment accelerates. Nearly four in five Gulf organisations have incorporated AI into their strategic plans, while 85% expect their AI budgets to increase during 2026. Almost 40% anticipate significant spending increases. Generative AI remains the leading priority, accounting for the main technology focus for about 35% of organisations surveyed across the region.

Yet only 34% of Gulf organisations have the enterprise-wide data foundations needed to scale AI effectively. Many are attempting to deploy advanced models across legacy technology stacks built over several decades, with customer information, operational records and financial data often stored in separate systems that cannot communicate easily.

Access to usable corporate information has emerged as one of the clearest constraints. Only 16% of GCC chief executives say their most frequently used AI tools can access all relevant documents and data, compared with 29% globally. Limited access reduces the ability of AI systems to generate accurate insights, automate complex processes or operate across departments rather than within isolated applications.

The bottleneck is particularly significant because companies are shifting from simple AI applications towards systems capable of performing multi-stage tasks. Early deployments concentrated heavily on email drafting, document summarisation and customer-service assistance. Attention is now moving towards research, analysis, decision support and autonomous AI agents that can execute workflows with limited human intervention.

That shift raises the cost of poor data. An AI assistant producing a draft document can be corrected by an employee, but autonomous systems interacting with customers, approving transactions or triggering operational processes require reliable information and clear accountability. Incorrect or incomplete datasets can introduce errors across automated workflows at much greater speed and scale.

Governance is consequently emerging as another dividing line between AI adopters and companies able to derive value from the technology. Only 13% of organisations in the UAE apply comprehensive AI governance frameworks across all their initiatives, even as about 20% are implementing dedicated AI governance platforms. Nearly half of executives also report difficulty managing complexity created by growing numbers of AI assets.

The challenge is not uniform across the Gulf. Some companies are already reporting strong results where AI programmes are closely linked to business objectives. A survey of technology decision-makers in Saudi Arabia found that 91% said AI initiatives were meeting or exceeding expectations, with half reporting performance above expectations. The results suggest that disciplined investment, workforce preparation and clearly defined use cases can produce returns even while broader regional adoption remains uneven.

Financial services provide another indication that AI can improve commercial performance when properly integrated. Research examining GCC banks found links between AI adoption, profitability and market valuation, although performance varied considerably between institutions and depended on implementation maturity.

Executives are also confronting an organisational issue. AI programmes launched separately by technology, finance, marketing and operations teams can create duplicate systems, incompatible datasets and overlapping licences. Consolidating those projects around shared infrastructure and common governance is becoming essential as companies move from dozens of pilots towards enterprise deployment.

Middle East chief executives nevertheless remain more aggressive adopters than many global peers. About 43% of GCC CEOs report extensive AI use in sales, marketing and customer service, almost double the 22% global level. Close to 59% of Middle East organisations have also formalised responsible-AI and risk processes, indicating that governance is moving higher on corporate agendas.

The emerging focus is therefore shifting from how many AI projects companies launch to how effectively those projects generate business value. Organisations are increasingly being pushed to narrow portfolios, define measurable outcomes before deployment, improve interoperability between systems and treat high-quality enterprise data as infrastructure rather than a by-product of technology operations.