GCC firms are racing to adopt AI, but only 11% are seeing returns: Some 84% of GCC organizations now use AI in at least one business function — just four percentage points behind the global average, according to McKinsey’s State of AI in the GCC 2025 report. But only 11% qualify as “value realizers,” meaning they have scaled AI and can attribute at least 5% of earnings to it.

Most firms are still stuck in pilots: Only 31% have scaled AI across the business, while 54% remain in the experimenting or piloting stages and 15% have not adopted AI at all. Cybersecurity (68%), output inaccuracy (53%), and data-quality issues are the top barriers to wider deployment. Lack of available capital to effectively adopt and scale usage also poses a challenge, one executive said.

Where AI is being used: Marketing and service operations remain the most common functions, but the fastest growth since 2023 is in product and service development: some 60% of organizations say they are piloting AI agents — which can execute workflows, coordinate tasks, and make decisions — marking a shift toward more action-oriented applications that can cut selling, general, and administrative expenses, and shorten development cycles.

Sectoral adoption is broad across the region. While energy and infrastructure respondents were heavily represented in the sample, this should not be read as sectoral leadership, Karan Soni, associate partner at McKinsey&Co, told EnterpriseAM, noting that uptake spans both public and private sectors.

Roles are shifting: Rather than replacing jobs, AI adoption is prompting companies to redesign workflows so staff move from routine tasks to higher-value work. The shift is also creating demand for new roles — including context engineers and AI product owners — though firms will need to upskill existing employees to fill them, Soni said.

The good news? Value can show up quickly: Well-designed transformations can deliver measurable results within 6-12 months, Soni told us. Companies seeing returns “start smart and scale fast,” using quick Ws to build momentum. Many are also “using AI to build AI,” accelerating delivery and reducing the talent required to bring new use cases online.