MENA’s AI Opportunity Is a Productivity Story, Not a Jobs Apocalypse

The loudest argument about artificial intelligence asks how many jobs it will destroy. For the Middle East and North Africa, that is probably the wrong starting point. The more immediate question is whether firms can turn increasingly accessible AI tools into better work, better services and higher productivity.

The World Bank’s October 2026 regional economic update estimates that fewer than 10% of jobs across MENA, Afghanistan and Pakistan face near-term automation risk. By contrast, between 13% and 20% of jobs could carry significant potential for augmentation: technology helping people perform existing work faster, more accurately or at a higher level.

That distinction matters. It moves the conversation away from a distant contest between humans and machines and toward a practical management problem. The region does not merely need models, data centres and announcements. It needs organisations capable of changing how work is done.

The opportunity sits inside existing jobs

AI adoption is often presented as the arrival of a new industry. Its larger economic effect may come through ordinary sectors: banks processing compliance work, logistics companies planning routes, hospitals organising records, governments handling requests, retailers forecasting demand and professional-services firms researching or drafting more efficiently.

In these cases, the value is not created by replacing an entire occupation. It comes from reorganising a sequence of tasks. Repetitive steps can be accelerated while human attention shifts toward judgment, client interaction, exception handling and accountability.

That means productivity gains will be uneven. A company that adds an AI subscription without changing processes may produce little more than faster first drafts. A company that maps workflows, improves data quality, assigns responsibility and measures outcomes can create a durable operating advantage.

Three gaps will determine who benefits

The World Bank identifies three structural constraints: localisation, usage and foundational capital.

The localisation gap is particularly important for Arabic. Regional languages and dialects remain underrepresented in the digital material used to train and evaluate leading systems. A tool may appear fluent while still missing legal context, local terminology, cultural references or the difference between dialects. For companies, that turns localisation into a product and governance requirement rather than a cosmetic translation exercise.

The usage gap separates access from productive adoption. Executives may experiment with consumer chatbots while core teams continue using fragmented data and manual approval chains. The relevant measure is not how many employees have tried AI. It is how many high-value workflows have been redesigned with controls, ownership and measurable outcomes.

The foundational-capital gap includes compute, connectivity, skills, education and management capacity. Gulf economies are investing heavily in infrastructure, while middle-income economies often possess strong technical talent but less capital. More fragile markets may gain most from smaller, purpose-built systems that work with limited connectivity rather than frontier-scale models.

Management quality becomes the differentiator

AI will reward firms that already understand their operations. Before automating a process, leaders must know where decisions are made, which data can be trusted, what error rates are acceptable and who remains accountable when a system is wrong.

This creates a new leadership agenda. Boards should ask where AI can improve revenue, cost, risk or customer experience—not merely whether the company has an AI strategy. Managers should separate reversible experiments from high-stakes deployments. Employees need permission to challenge outputs, while audit trails must make important decisions explainable.

The strongest regional companies may not be those building the largest models. They may be those learning fastest where AI works, where it fails and how to combine it with sector expertise.

A regional division of capabilities

No single country needs to build every layer of the AI stack. The Gulf can contribute computing capacity, capital and ambitious public-sector demand. Egypt, Jordan, Lebanon, Morocco and Tunisia can contribute engineering, language expertise and service-delivery talent. Universities, publishers, governments and cultural institutions can help develop representative datasets and evaluation standards.

Regional collaboration will be difficult. Data rules, procurement systems, commercial incentives and language standards differ. But the alternative—every market duplicating infrastructure while sharing too little knowledge—would be costly and slow.

What should be measured next

The next phase should be judged through outcomes: revenue per employee, processing time, error rates, customer satisfaction, adoption beyond pilot teams and the number of Arabic-language use cases performing reliably in real conditions.

The central risk is not that AI immediately eliminates work across MENA. It is that investment concentrates in infrastructure and demonstrations while most firms fail to translate access into productivity. The opportunity is real, but it will be captured through operating discipline rather than technological spectacle.

Sources

• World Bank: Middle East, North Africa, Afghanistan and Pakistan Economic Update, October 2026
• World Bank press release: AI, jobs and regional growth

Featured image: Annie Spratt / Unsplash.