Inside Abu Dhabi’s race to build an AI-native government

The UAE has spent over a decade digitizing public services: online forms, government apps instead of service-center visits, and paperwork replaced by digital records. 

Now, Abu Dhabi aims to take things even further.

By 2027, the emirate wants to become the world’s first fully AI-native government, where artificial intelligence is not just used to answer questions or speed up individual services, but is built into how the government operates. Building AI systems that can anticipate what they need and, in some cases, act before they ask.

Backed by $3.5 billion under the Abu Dhabi Government Digital Strategy 2025–2027, Abu Dhabi is aiming for a 100% end-to-end digital transformation, and more than 200 AI use cases deployed across the government by 2027.

But building an AI-native government is much more than just launching a few successful AI tools.

It means connecting government data, moving services to sovereign cloud infrastructure, redesigning decades-old processes, and deciding how much responsibility to hand over to AI systems.

And perhaps the biggest question is: Can Abu Dhabi do all of that by 2027?

The shift is already happening, says His Excellency Wesam Lootah, Director General of GovDigital at the Department of Government Enablement, Abu Dhabi (DGE).

“We are not simply running pilots on the side; these are fully integrated, system-level transformations serving citizens and residents daily,” Lootah adds.

Through TAMM, Abu Dhabi’s unified government services platform, more than 1,150 public and private services are now available in one place. According to Lootah, AI resolves more than 95% of service requests autonomously.

AutoGov, which Lootah describes as the world’s first AI public servant, is already working in the background to renew trade licenses, manage disability card renewals, and process routine utility payments without users having to fill out a form or make a request.

“By treating AI as institutional infrastructure, we are actively saving 54 million users and business hours annually,” Lootah says.

According to research from INSEAD, the UAE has already moved beyond the pilot stage. The country has spent years building many of the foundations needed to deploy AI at scale, from dedicated AI leadership and sovereign cloud infrastructure to government-backed technology companies and governance systems.

Abu Dhabi, in particular, has focused heavily on infrastructure.

“In Abu Dhabi, we are seeing an infrastructure-first approach built around shared platforms, sovereign cloud, and reusable AI capabilities,” says Islam Abdul Karim, Regional Head at Yango Group.

“What remains a vision is not the technology itself, but the transformation of government operating models,” Abdul Karim says.

He adds: “The next phase is making AI the foundation of how government processes are designed and delivered, not simply an assistant layered onto existing workflows.”

WHAT DOES AI-NATIVE ACTUALLY MEAN?

The term “AI-native” is increasingly being used by companies, startups, and governments.

But what does it actually mean for a government?

For Lootah, it is about changing the relationship between residents and public services.

“Being ‘AI-native’ means redesigning government around intelligence, anticipation and human outcomes,” Lootah says.

He adds: “It means moving away from a government you have to visit, to an agentic government that anticipates your needs and works quietly on your behalf.”

Abu Dhabi has also set specific targets.

The bigger goal is to change how people interact with the government. An AI-native government could work differently.

The system could recognize when a license is about to expire, when someone becomes eligible for a service, or when a payment is due. Instead of waiting for someone to ask, the government could act first.

Unlike traditional AI assistants that primarily answer questions or generate content, AI agents can plan tasks, integrate with various systems, and carry out entire processes with minimal human involvement.

Ahmed Galal, Director of Public Sector at Dell Technologies, says this creates new challenges for governments.

“Agentic AI doesn’t just process data; it participates in managing processes,” Galal says.

Every action requires access to data and computing infrastructure. And the more AI systems are allowed to do, the more complex the technology behind them becomes.

“This creates complex workloads, which legacy infrastructure struggles to support,” Galal says. “To deploy agentic AI safely, governments must adopt a controlled approach where performance, data sovereignty, and security are paramount.”

BUILDING THE INFRASTRUCTURE

AI strategies usually aim for smarter services, automation, and better predictions of what people need. However, these goals depend on having the right infrastructure, such as data centers, computing power, storage, and cybersecurity systems.

Galal says data availability and quality remain among the biggest challenges for organizations trying to move AI projects from testing to production.

“Even the most promising use cases cannot scale without a trusted, AI-ready data foundation,” he says.

Other problems include fragmented infrastructure and a lack of in-house AI skills.

For Abu Dhabi, another major part of the strategy is data sovereignty.

“All critical data and architecture for our TAMM platform sit securely on sovereign cloud infrastructure within the UAE’s borders, placing them entirely under our legal jurisdiction,” Lootah says.

The Department of Government Enablement is also working with global technology partners to develop sovereign hybrid cloud infrastructure that can support national AI workloads.

Keeping data inside the country, however, is only one part of the issue.

“In a world where cyber threats grow more sophisticated by the day, sovereignty isn’t just about where data lives but who controls it, how it’s protected, and how quickly an organization can respond when something goes wrong,” Galal says.

This becomes even more important as governments move toward AI agents.

The more tasks AI systems can carry out, the more government systems and data they need access to.

And that raises another difficult question: What happens when something goes wrong?

WHEN AI MAKES A MISTAKE

An incorrect decision about a government benefit, permit, or public service can have much more serious consequences.

As governments give AI systems more responsibility, the question of accountability becomes harder to ignore.

“This is one of the defining governance questions of the AI era,” Abdul Karim says.

Today, most governments use a human-in-the-loop approach. AI systems may make recommendations or carry out certain tasks, but humans remain involved in important decisions.

In Abu Dhabi, Lootah says humans will continue to have the final responsibility.

“In our ecosystem, accountability must follow autonomy, which is why there is always a ‘human in the loop’ at every stage, from telephone and video support to in-person home visits for those who need them, ensuring no one is left behind.’

He adds: “We are building an agentic government where AI systems initiate actions within clearly defined mandates, but human beings and institutions always retain ultimate accountability.”

“If an escalation is needed, a human steps in immediately.”

SHOULD EVERYTHING BE AUTOMATED?

A fully AI-native government may sound like a government where everything is automated. But that is not Abu Dhabi’s goal.

“Regarding automation, it is less about which services are ‘hardest’ to automate and more about which services intentionally require a human touch,” Lootah says.

He describes this approach as “co-cognition.”

Lootah says, it is where AI handles high-volume pattern recognition and routine tasks, “freeing up our human civil servants to apply judgment, empathy and community connection where it matters most.” 

THE BIGGEST ADVANTAGE

While many countries have published AI strategies, Abdul Karim says far fewer have aligned leadership, procurement, infrastructure, talent development, and government delivery around a common national objective, as the UAE has.

Still, the next stage will be harder. 

“This includes redesigning institutions around autonomous AI, developing mature governance for AI agents, measuring public value beyond efficiency gains, and ensuring trust and accountability at scale,” adds Abdul Karim.

Lootah does not view Abu Dhabi’s AI strategy as a competition with other countries.

“We view our digital evolution not as a race against specific nations, but as a shared global journey,” he says. “It is a partnership, not a competition.”

“However, our most important benchmark is human impact,” Lootah adds.

CAN ABU DHABI MEET THE 2027 DEADLINE?

The technology is moving quickly. But the institutions needed to govern it are still catching up. The systems available in 2027 could be far more capable than those being used today. They could also create new risks that governments have not yet prepared for.

Abdul Karim believes one of the biggest challenges ahead will be governing increasingly autonomous AI agents.

“As AI shifts from assisting people to independently orchestrating workflows and making operational decisions, governments will need entirely new models for oversight, accountability, auditing, and continuous evaluation,” he says.

“The technology is progressing faster than the institutional frameworks required to govern it.”

So, what happens if not every service is ready?

For Lootah, meeting the target does not mean automating everything.

“Our north star is people-first design, and the ultimate measure of our success is public trust,” he says.

The goal for 2027, he says, is to build the foundation for a government that can anticipate residents’ needs and remove friction from their daily lives.

“We will never import the tech mantra of ‘move fast and break things’ into public service, because breaking things is simply not an option in high-stakes domains like health and public safety,” Lootah says.

And that is likely to be the real test of Abu Dhabi’s AI ambition.

The emirate has invested in building its infrastructure. It has already taken AI projects past the pilot phase and set a clear deadline for its next steps.

But the success of the world’s first AI-native government will not only depend on how many services are automated by 2027.

“Success will be defined by the quality of the lives we improve,” adds Lootah.