Across the UAE, the conversation about artificial intelligence has shifted. It is no longer about whether to adopt AI, but how to move from experimentation to execution responsibly.
That shift has real momentum behind it. According to an AWS and IDC report, 98% of organizations in the UAE and Saudi Arabia believe AI will fundamentally change how they operate. Currently, 28% are already investing in AI, and a further 50% plan to in the near future. Among those who have invested, 88% report measurable performance improvements.
This is the context behind the UAE government’s own ambition. In April 2026, the UAE announced its intention to transition 50% of federal government sectors, services and operations to agentic AI systems within two years. It’s an ambitious target, and one that reflects a broader truth I see in my conversations with customers across the region: organizations are ready to move past pilots and proofs of concept, and into workflows that matter.
Agentic AI: More Capability, Same Control
I want to be direct about something, because I think it’s worth clarifying: agentic AI is built to work alongside people, not replace their judgment. The most effective automation begins after an organization has mastered the tools it already relies on. The goal isn’t to hand off a task and walk away – it’s to extend what teams can do while keeping humans in control.
Every step of the work is verifiable, and every outcome is checkable. That’s the principle we’ve built into two of our solutions that are helping organizations across the region move from experimentation to governed execution: Amazon Quick and Kiro.
Amazon Quick: Putting Business Context to Work
Amazon Quick is designed to help business teams work with the information and applications they already use, through natural-language interaction, within the access and governance controls an organization has already put in place. It can help employees find information faster, support research tasks and reduce the time spent on repetitive administrative work, but always within a framework where accountability remains with people.
Kiro: Structure Before Code
For engineering teams, Kiro takes a similarly disciplined approach. Rather than generating code line by line from loose instructions, Kiro uses what we call spec-driven development: it turns a plain-language request into a structured specification (covering requirements, design and implementation steps) before any code is written. Developers review and approve that specification first. This gives technical teams an AI collaborator that behaves less like a black box and more like a disciplined partner, one that shows its reasoning before it acts.
Regional Relevance Matters
For the Middle East, relevance isn’t just about language – it’s about context. Organizations can ground their AI experiences in their own internal documents, communications, and business information, giving outputs a stronger connection to the way their business actually operates day to day. And when that grounding is paired with human expertise particularly for customer-facing, Arabic-language, or culturally nuanced communications the results become even more precise and impactful.
But contextual AI only delivers its full value when people have the skills to use it well. That’s why we’re investing directly in the skills base that makes any of this possible. Through our collaboration with e&, including the AI Nation Afaaq program, which aims to train more than 30,000 people across the UAE in cloud computing and AI skills, AWS is helping build regional AI capability. Technology alone doesn’t close a skills gap; people do, and they need the right training to operate these systems responsibly.
What Comes Next
The UAE has set a clear and ambitious direction for agentic AI, both in government and across the private sector. The opportunity lies in scaling autonomy with purpose – where every step forward is grounded in governance and trust. As the country works methodically toward its UAE National Strategy for Artificial Intelligence 2031 ambitions, our focus stays the same: giving organizations the governed, verifiable tools they need to move from experimentation to responsible execution.
