We’re almost halfway to 2027, which also marks a decade since 2027 when the UAE launched its National Artificial Intelligence Strategy. At the time, this may have seemed ambitious, but today the direction is clear.
In the last few weeks alone, we have seen Abu Dhabi doubling down on building an AI-native government, while Dubai has launched new initiatives around Agentic AI adoption across the private sector. The signal from the UAE’s leadership is unmissable: AI is and will continue to be foundational to economic growth, competitiveness, and national transformation in the Gulf and elsewhere.
Financial institutions (FIs) in the region are following suit as the UAE has ranked, for the second year in a row, among the world’s leading nations for AI adoption, workforce readiness and AI integration across both government and enterprise environments. Microsoft’s AI Diffusion Report found that the UAE became the first economy globally to surpass 70% AI usage among its working-age population.
When operating in a country leading the world in AI adoption, standing still is not an option, so FIs are under increasing pressure to keep up. But while racing to integrate AI, many businesses are skipping the foundational steps that allow AI to work and scale.
Take off starts at the launchpad
There is no shortage of ambition around AI-driven automation right now. Every institution wants smarter operations, faster decision-making, more personalised customer engagement, lower costs, better compliance, the list goes on. Yet for many, progress stalls even before initiatives can move beyond pilot phases. To put this into perspective, while as many as two-thirds of enterprises worldwide have experimented with AI agents, fewer than 10% have successfully got them to deliver tangible business value.
The reason is not lack of vision; it’s that the launchpad underneath the AI initiative isn’t fit for purpose. A useful analogy here is SpaceX, which did not transform the economics of space travel simply by building more powerful rockets. The breakthrough came from building the precision infrastructure required to launch, recover, and relaunch those rockets. AI adoption works the same way, where the rocket is the model, and the launchpad is accurate and reliable data.
When the foundations can’t support the mission
Banks and FIs today operate across deeply layered systems that have evolved over the years. Most of these systems, however, were added over time to meet business needs and were not designed to work seamlessly together.
What this means is that many institutions today are basically a collection of disparate, disconnected architectures producing inconsistent versions of the same information.
The implications are clear – time wasted and trust erosion. The well-known 80/20 split tells us that 80% of time is spent preparing and reconciling data before analysis can even begin, leaving only 20% for actual analysis and modelling. This significantly slows down decision-making and innovation, as teams spend more time fixing and validating data than actually scaling AI-driven services. And despite the effort going into organising data, institutions operating with data silos trust only 60% of their own data due to fragmentation and inconsistencies, pointing not only to the need for more accurate data, but also better management of it.
In highly regulated financial environments like banking, unreliable and fragmented data also creates greater compliance and operational risk, particularly for multinational financial institutions that choose the UAE as their regional base because of its attractive business environment, while operating across multiple markets, each with different regulatory requirements, local providers and compliance standards.
The institutions building for lift-off
The paradox many FIs face is that they must modernise and digitize to remain competitive, but the very systems they rely on to operate efficiently are slowing down innovation, to the point where 85% of AI projects fail because of poor data quality, and 60% of AI projects are abandoned because they are unsupported by AI-ready data.
This, however, is starting to change as more institutions look for infrastructure that lets them modernise gradually instead of replacing entire systems at once. Modular orchestration and integration layers let FIs improve systems gradually while keeping existing core systems active and production stable. By focusing first on data orchestration, interoperability, and integration layers, FI’s can accelerate innovation cycles quite dramatically while maintaining operational resilience.
The benefits of this approach are readily visible. For example, reporting preparation times can be reduced by up to 70%, significantly speeding up decision-making, and innovation cycles can be reduced from nine months to just nine weeks. At the same time, eliminating integration errors and data inconsistencies can reduce AI retraining cycles by up to 40%.
Who will actually reach orbit?
This matters not only for operational efficiency, but for national competitiveness. Alongside the UAE’s AI ambitions, financial institutions will play a central role in delivering that vision.
Just as reusable rockets transformed space travel through the infrastructure that made repeated launches and landings reliable, AI also needs accurate and reliable infrastructure to scale effectively.
Without strong, interoperable, and scalable data foundations, AI risks remaining stuck on the launchpad, limited to pilot projects and proof-of-concept exercises that never truly lift off. At the end of the day, AI is only as strong as the infrastructure beneath it, so the institutions that will effectively reach orbit will be the ones building those foundations now, creating the launchpads that allow AI innovations to soar.
