Skip to content

Dubai banks leave the testing phase: local deployment of models and smart agents to manage operations

Share
Dubai banks leave the testing phase: local deployment of models and smart agents to manage operations

Listen to this article

Read by Anchor

Major banks in Dubai have moved beyond exploratory trials to embed AI models at the core of their daily operations, announcing that the technology has shifted from an add-on interface to a fully operational banking architecture. During Dubai’s Future of Finance 2026 week, which hosted the Dubai FinTech Summit from 2 to 6 November, the disclosed data showed that the largest banks are not only integrating generative tools but also reshaping compliance frameworks, risk management and cyber-defence through an independent local infrastructure.

Emirates NBD, the emirate’s largest bank by assets and majority-owned by the government, leads this path by allocating roughly 50 percent of its technology budget to AI. The bank runs more than 150 AI initiatives, with 65 already in full production. This approach earned the bank the top spot in Evident’s Middle East and Africa bank index released in June 2026 by the London-based Evident firm, which studied over 100 financial institutions and linked the bank’s performance directly to institutional productivity, process modernization and the UAE’s national agenda.

Running advanced models on a locally hosted infrastructure resolves the data-sovereignty and regulatory-compliance dilemma for the banking sector.Emirates NBD has become the first bank in the MENA region to deploy OpenAI’s GPT models locally, allowing it to run large language models at scale while strictly adhering to the UAE’s data-localization and privacy laws. The move aims to improve customer experience, strengthen financial-crime prevention through advanced analytics, and boost employee efficiency.

The most striking shift, however, is in cyber-security. Miguel Rio Tinto, chief digital and information officer at Emirates NBD, says banking will change in the next five years more than it has in the past three decades. That vision now faces a tough field test, as the number of critical security vulnerabilities the bank examines each month jumped from about 1,000 in early 2025 to roughly 15,000 last month. To cope with the pressure, the bank turned to OpenAI’s flagship model as a defensive tool to test its systems and uncover internal weaknesses.

These defensive measures come amid a turbulent regional environment, where daily cyber-attack attempts in the UAE have risen from around 200,000 to as many as 700,000 following the outbreak of the US-Iran war on 28 February 2026. The UAE Cyber Security Council warned against the use of AI by cybercriminals in complex attacks after the UAE successfully thwarted, in July, phishing and malware campaigns targeting its financial sector.

Conversely, Dubai Commercial Bank disclosed a comprehensive technology roadmap focused on independent smart agents, built through strategic partnerships with Microsoft to broaden adoption, Silent IT to enhance compliance and risk management, Komarch for digital-wealth management, Zambe for intelligent automation, and Harvy, Sirion and Flank to modernize legal infrastructure and governance. Thomas Shiryan, chief technology and services officer at the bank, explained that moving to agent-based AI enables re-engineering processes and delivering banking services that are safer and faster, estimating that these agents can handle 55-60 percent of operational tasks while retaining human oversight for final approval, and the bank will retain intellectual-property ownership of the agents it develops.

Shifting from traditional generative models to independent agents resets the skill set and operating cost landscape across regional markets.This development aligns directly with the UAE government’s targets, which launched in April 2026 an initiative to convert 50 percent of federal government operations and tasks to agent-based AI systems by 2028, followed by Dubai’s May action plan to accelerate private-sector adoption of the technology.

For financial institutions and professional talent in the Gulf, Egypt and the Levant, this landscape imposes new engagement rules. Institutions can no longer rely solely on external APIs without a local host that guarantees account confidentiality and compliance with central-bank regulations. The expanding scope of automated cyber-attacks also forces banks to replace periodic manual reviews with proactive defensive audit systems that scan thousands of code vulnerabilities simultaneously. For engineers and developers in regional tech hubs, designing simple command-line scripts is no longer sufficient; building independent-agent architectures and governing their workflows has become the benchmark that determines competitive and partnership opportunities with major banks.

Don't miss the next story

Subscribe for updates