Emirates NBD is building Leap as a governed enterprise layer that standardises security, monitoring, model governance and AI development across banking teams.
The framework integrates open-source technologies while reducing fragmentation, helping developers adopt new models without repeatedly rebuilding enterprise applications.
Its broader AI infrastructure already supports 2,300 engineers across 180-plus squads, with 95% daily adoption and over 100 AI use cases.
The banking industry is moving into a new phase of artificial intelligence. Generative AI largely focused on answering questions, summarizing information and assisting employees. Agentic AI goes a step further: it can plan tasks, interact with systems, use tools and execute multi-step workflows. For a bank, however, giving AI that level of autonomy creates a fundamentally different technology and governance challenge.
That is why Emirates NBD is building its own enterprise agentic AI infrastructure rather than allowing individual teams to independently assemble AI agents using public frameworks.
Frameworks such as LangChain and LangGraph have become important building blocks for developing AI agents. Emirates NBD does not reject these technologies. Instead, its strategy is to place them underneath an enterprise-controlled platform.
The distinction is important. If every development team selects its own framework, model, security approach and monitoring tools, a large bank could quickly end up with dozens or hundreds of incompatible AI implementations. That creates duplicated engineering work and, more importantly, inconsistent risk controls.
Leap is intended to function as an abstraction layer. Developers can use underlying open-source technologies while inheriting common capabilities for authentication, authorisation, audit logging, observability, model governance, prompt management, evaluation and enterprise tool integration.
A conventional chatbot may generate an incorrect answer. An agentic system can potentially make a series of decisions based on that incorrect information and then act on them. The consequences in banking can therefore be significantly larger.
An enterprise framework can establish guardrails before an agent reaches production. This also makes compliance scalable. Instead of asking every AI project team to interpret banking requirements independently, the platform embeds many controls into the development environment.
The philosophy is essentially ‘governed by default’ rather than governance being added after an AI application has already been built.
Also Read: UAE Workforce Stays Bullish as 91% Express Confidence in Career Future
There is another reason the infrastructure approach makes sense. The AI ecosystem is changing extraordinarily quickly. Today's preferred foundation model or orchestration framework may not remain dominant next year. If individual banking applications are tightly coupled to particular models, replacing them could require substantial redevelopment.
An abstraction layer provides Emirates NBD with greater flexibility. New models and AI technologies can be incorporated within the platform while applications continue to use a consistent enterprise interface.
The agentic AI strategy also fits into a broader AI transformation already underway at the bank.
Its Developer Hub currently supports 2,300 engineers across more than 180 agile squads, with 95% daily platform adoption, 60% faster time to production and more than 100 active AI use cases across multiple large language models. Reportedly, more than 60% faster account-opening processes and 68% faster KYC review turnaround through its Document Intelligence platform.
These numbers matter because they suggest Emirates NBD is not treating agentic AI as an isolated innovation project. It is attempting to create an industrial-scale operating model for AI.
The most interesting aspect of the strategy is that Emirates NBD is not betting that it can build a better foundational AI model than technology companies. Instead, it is betting that the lasting advantage will come from how effectively a bank can deploy different models inside its own regulated environment.
Emirates NBD's own ranking reinforces the direction: it was placed first in the inaugural Evident AI Index for Banks in the Middle East and Africa in June 2026, with particularly strong performance across talent, innovation and leadership.
Ultimately, building its own agentic AI infrastructure is less about reinventing AI and more about creating the rails on which AI can safely run at banking scale. The strategic calculation is clear: models will change, frameworks will change and AI techniques will change. But a secure, governed platform capable of absorbing those changes could become a much more durable competitive asset.
Also Read: How Will the UAE's AI Strategy in 2027 Transform Foreign Direct Investment?
Leap is an enterprise agentic AI platform designed to help Emirates NBD deploy autonomous AI with embedded security, governance, monitoring and integration capabilities.
The bank wants consistent governance across AI deployments, controlling authentication, data access, auditing, human oversight and model usage without relying on separate team-level implementations.
No. Emirates NBD plans to use technologies such as LangChain and LangGraph underneath its enterprise platform while keeping governance and architecture centrally controlled.
Leap provides an abstraction layer, allowing Emirates NBD to introduce new foundation models and orchestration technologies without requiring widespread changes across existing banking applications.
Its Developer Hub supports 2,300 engineers across more than 180 agile squads, with 95% daily adoption and over 100 active AI use cases.