Think about the last time a patient complained about waiting. Not for a bed or for a scan result, but for something administrative such as a prior authorisation, a referral, or a follow-up appointment that fell through the cracks. These aren’t clinical failures – they’re coordination breakdowns. And for the most part, they’re entirely preventable.
This is precisely the problem that agentic AI – a new generation of autonomous systems that can observe, plan, and act across multiple tasks without constant human instruction – is beginning to solve. Not sometime in the future but right now, in hospitals and health systems around the world – and, increasingly, right here in the UAE.
From tool to teammate
Most people in healthcare have already encountered AI in some form, either a diagnostic algorithm, an ambient documentation tool, or a chatbot handling appointment reminders. These are useful but they are, fundamentally, tools. You prompt them, they respond, but the work still flows through a human coordinator.
Agentic AI is different – in ways that are set to astound us all. An agentic system can take a clinical or administrative objective, such as“schedule this patient for post-surgical follow-up, coordinate with the pharmacy, and update the EHR when done”, and execute it end-to-end, adapting as it goes. McKinsey’s 2025 State of AI report describes these systems as “virtual coworkers” capable of orchestrating multi-step workflows that previously required several human handoffs.
This represents a significant shift. We’re moving from AI as a tool you pick up to AI as a colleague you delegate to. And for healthcare, a sector defined by complexity, coordination, and chronic workforce pressure, the implications are enormous.
What’s happening on the ground
It’s tempting to frame agentic AI as a future story. But it isn’t. The deployments are live, and the evidence is accumulating.
Here at home, the clearest signal came earlier this year when Emirates Health Services partnered with Boston Health AI to launch Amal, the UAE’s first AI physician assistant, across its healthcare network. Amal conducts comprehensive pre-consultation interviews with patients, capturing structured medical histories before the doctor walks in. The system is culturally adapted for the UAE population, with dialect and appearance tailored for the communities it serves. EHS is now the first public health system in the UAE to implement a multi-agentic AI clinical intelligence platform.
It did not stop there. EHS also unveiled Hamda, an intelligent digital agent that is overhauling contact centre operations, and Maitha, described as the world’s first AI-powered nursing workforce management system aimed at handling recruitment, training co-ordination, and career development pathways. These are not pilots in a single department – they are systemic deployments across an entire public health network.
The appetite for this kind of innovation in the region and across all industries is unmistakable. IDC projects AI spending across the META region to grow at a CAGR of 34% through to 2028, reaching USD 14.6 billion, up from USD 4.5 billion in 2024.
Where the real value lives
Three areas stand out as particularly transformative for facilities operating in Dubai and the wider GCC.
The first is the administrative burden. Clinical staff across the region spend a disproportionate share of their time on documentation, authorisations, and scheduling, time that is not spent with patients. Deloitte’s analysis of healthcare transformation in the Middle East identifies digitalisation of these workflows as one of the most significant levers available to regional providers. Agentic AI can close this loop by not just flagging tasks but completing them, routing information across systems, and surfacing only the exceptions that genuinely need a human decision.
The second is patient access and experience. The UAE’s healthcare system serves a highly diverse, multilingual population with sophisticated expectations. Agentic systems that can engage patients in their preferred language, gather clinical history conversationally, and reduce waiting, as Amal is designed to do, address a genuinely local challenge in a way that generic digital tools have not.
The third is proactive care. Agentic AI can monitor patient data continuously, from wearables, from lab trends, from HER patterns, and flag deterioration before it becomes a crisis. For a health system managing growing rates of chronic disease across an ageing resident population, this shift from reactive to anticipatory care is a gamechanger in healthcare.
The policy tailwind
Healthcare facilities in Dubai do not need to make this case to their governments. The case has already been made. The UAE’s National AI Strategy 2031, overseen since its launch by the world’s first Minister of State for Artificial Intelligence, identifies healthcare as a priority sector for AI deployment and commits to proof-of-concept projects in advanced diagnostics and digital health infrastructure.
PwC estimates that AI will contribute USD 320 billion to the Middle East economy by 2030, with the UAE seeing the largest proportional impact of close to 14% of GDP. Healthcare is explicitly part of that story, and facilities that move early will be better positioned to benefit from the infrastructure, talent pipelines, and regulatory frameworks the government is actively building around them.
This means the risk has shifted. For much of the past decade, early AI adoption in healthcare felt like moving ahead of the regulatory curve. In the UAE today, cautious non-adoption is the riskier position.
The human-in-the-loop imperative
None of this means throwing caution aside. The most important lesson from every successful agentic AI deployment in the UAE, and globally, is that human oversight is not a brake on progress. It is necessary to make progress sustainable.
PwC’s 2026 AI predictions state that agentic workflows are spreading faster than governance models, and organisations that build accountability frameworks before they scale deployments consistently outperform those that bolt governance on afterwards. In healthcare, where the stakes of a misaligned decision are not a delayed invoice but a patient outcome, this principle is non-negotiable.
There is also the question of errors. Agentic AI systems can get things wrong. They can misinterpret context, generate plausible-sounding but incorrect outputs, or act on incomplete data. In healthcare, that is not something that can be brushed aside. The response, though, is not to stand back, but to be deliberate about where autonomy is granted.
Administrative co-ordination, scheduling, patient intake and workforce management are high-volume, high-friction tasks where AI errors are catchable and correctable. Autonomous clinical decisions about diagnosis or treatment are a different matter entirely, and no responsible deployment conflates the two. The goal is not a system that never makes mistakes. It is a system designed so that mistakes are visible, accountable, and far less frequent than the human errors they replace.
Adopting agentic AI within your healthcare framework is inevitable. The questions that should be asked at every level are “where do we start, who is accountable, and how do we connect individual deployments into something coherent over time?”
Start before you’re ready
One of the most common refrains in conversations about agentic AI is, “We’re waiting to see how it matures”. This is an understandable instinct, but it carries a cost that tends to be invisible until it isn’t. The facilities deploying these systems today are not just gaining efficiency – they are building institutional fluency and learning how to govern AI workflows, retrain staff for new roles and integrate these systems with existing infrastructure. That learning takes time.
Healthcare has been through this before. The facilities that engaged early with electronic health records, even the clunky first-generation versions, were dramatically better positioned than those who waited for the technology to stabilise. There was never a moment when it was fully stable – there was just earlier and later.
Agentic AI is following the same curve. A single well-chosen pilot – one workflow, one department, with proper oversight and clear success criteria – builds more genuine readiness than any amount of monitoring from the sidelines.
Lead, follow, or catch up
Healthcare has always been about coordination – between clinicians, between departments, between patients and systems. Agentic AI doesn’t offer a replacement for that coordination. It’s infrastructure that makes it dramatically less dependent on humans.
The facilities that will define the next decade of healthcare in this region are not waiting to see how this plays out. They are making deliberate choices now about where to begin, how to govern it, and how to bring their teams with them through the transition. The window for deliberate, strategic adoption is open. It won’t stay that way indefinitely.
