RAKBANK can retrieve decades of documents in minutes with Azure AI Services, streamlining compliance

RAKBANK provides banking services across the United Arab Emirates, offering personal, wholesale, and business banking through its physical branches and digital platforms. Founded in 1976, the bank continues to evolve to meet customer needs. “At RAKBANK, our vision is to be the best digital bank with a human touch,” says Saket Saith, Group Chief Technology and Data Officer (CTDO) at RAKBANK. “A fully digital approach alone is not enough—we strive to offer exceptional online experiences, supported by our people whenever needed.”

In 2021, RAKBANK embarked on a digital transformation journey focused on four key areas: modernizing platforms, strengthening data capabilities, simplifying technology, and increasing automation. RAKBANK needed to streamline digital experiences for both customers and employees. “We’ve seen too many examples in the financial services industry where banks focus on making the front end look digital, but in the background, processes still rely on manual or paper-based workflows. This creates inefficiencies and breaks the customer journey,” explains Saith.

One of the most significant challenges the bank set out to address is a widespread issue affecting the entire banking industry in the country. Like many financial institutions in the UAE, RAKBANK holds extensive records for both retail and corporate clients, including sensitive documents such as passports, Emirates IDs, know-your-customer (KYC) forms, and other personal identification information (PII). This wealth of data was stored in highly unstructured formats, often as scanned images or PDFs with no standardized naming conventions, scattered across disparate legacy systems. As regulatory requirements in the UAE tightened—particularly around KYC and anti-money-laundering compliance—this fragmented data landscape became a growing liability. The UAE Central Bank enforced more rigorous compliance checks, demanding frequent audits of customer documents, with heavy penalties for non-compliance.

For RAKBANK, this created a major operational bottleneck, with compliance officers spending an average of 80 minutes per case manually locating, verifying, and compiling KYC documents for regulatory review. The sheer volume of data—spanning terabytes of information and covering over 500,000 retail customers and an additional 100,000-plus business and corporate clients—made this process time-consuming and error-prone. RAKBANK needed a smarter approach.

Rethinking the digital transformation approach

At first, RAKBANK explored traditional solutions but realized these could take 18 months or longer to implement. “Instead of relying on highly specialized data scientists or third-party vendors, we decided to use generative AI. It would do all the heavy lifting and allow us to solve the same problem in a fraction of the time,” highlights Saith. “Plus, the flexibility of the technology meant that we could expand its use to a much broader set of applications in the future.”

RAKBANK began its transformation with a successful proof of concept (PoC), leveraging a suite of Azure-powered solutions. At the heart of the system is Azure OpenAI Service, which provides advanced document intelligence capabilities, enabling the system to process, analyze, and extract relevant data from unstructured documents. “We used Azure Form Recognizer on old, hand-filled account opening forms—some dating back 10 or 20 years—and the AI was able to read the handwriting, extract the key information, and structure it into digital records. Manually reviewing this massive data set would have been nearly impossible. We saved thousands of hours of sifting through files,” shares Saith.

Processing documents faster

This setup enabled RAKBANK to digitize and index over 2 million documents, transforming static files into searchable data points and classifying them into 50 document types. The extracted information was then securely stored in Azure Blob Storage. To make this vast amount of information easily accessible, Azure AI Search was integrated, enabling compliance teams to perform keyword-based searches across the entire document repository.

As a result, RAKBANK can now easily run analytics on customer documents to identify missing information or expired records, streamlining compliance checks and ensuring regulatory requirements are met. “For every customer, we have complete visibility into IDs, passports, and essential business documents like articles of association or trade licenses. Azure AI Services enable our compliance team to close cases in 20 minutes compared to 80 minutes or more previously,” shares Saith.

By taking away the tedious task of manually checking KYC documents, employees are freed up for higher-value work. “They can instead focus on compliance strategy and risk mitigation, areas where their expertise really matters,” adds Saith.

With faster case closures and a smoother experience, customers benefit the most. “Clients no longer need to upload endless documents or deal with back-and-forth requests. Thanks to Azure AI Services, everything’s handled digitally and with fewer touchpoints, resulting in a much faster and less frustrating process,” shares Saith. “Their interactions with the bank have become faster, simpler, and more intuitive.”

Offering an intuitive customer experience

Having streamlined its compliance processes, RAKBANK next set its sights on enhancing the user experience through its customer-facing mobile banking application. “Like most typical mobile banking apps, ours is extremely feature-rich, but the features are embedded in different places—menus, tabs, and pages,” Saith says. As the app grew in functionality, customers found it increasingly difficult to quickly access key features.

To address this, RAKBANK introduced a digital assistant powered by Azure OpenAI. “It’s like having a digital assistant inside our app. Customers can simply state what they need, and the AI assistant will guide them directly to the relevant feature,” explains Saith. This approach eliminates the need for customers to scroll through multiple menus or manually search for specific functions. “For example, imagine a customer notices fraudulent transactions on their credit card. In such a stressful moment, they need to block their card quickly,” Saith adds. “Without the AI assistant, they might have to navigate through two or three screens to find the “block card” option, all while feeling anxious and rushed, potentially clicking the wrong buttons.” The AI assistant can process queries in different languages—even ones the app itself doesn’t officially support.

Partnership for AI transformation

RAKBANK’s transformation required not just the right technology but also the right support to implement it effectively. “Our partnership with Microsoft has been critical to the success of these AI programs. One of the biggest advantages was Microsoft’s ability to offer AI capability in the local cloud here in the UAE. Without this, data sovereignty regulations might have prevented us from leveraging AI at all,” shares Saith.

Another key factor was Microsoft’s support in the development of these solutions. “When you’re experimenting with new technology—especially something pioneering in the industry—the upfront investment can be a barrier. But with Microsoft’s support, we were able to approach this in a bite-sized, scalable way,” adds Saith.

RAKBANK’s next phase will fully automate the document renewal process, enabling the system to automatically request updates from customers when documents expire or go missing. This will further streamline compliance workflows, enhance ongoing due diligence, and ensure customer records remain continuously up to date, all while improving the overall customer experience.

Looking ahead, RAKBANK sees AI as a defining force in the industry. “I believe spoken language, like English, will become the programming language of the future. In the future, we will be solving even more complex challenges, driving an unprecedented level of efficiency through generative AI,” concludes Saith.