Enabling Remote Learning During the Pandemic

Mashreq Bank, headquartered in Dubai, has helped shape the region’s banking and financial industry over the last five decades. Their global presence spans Europe, Asia, Africa, and the United States. Since 2018, Mashreq has worked towards becoming a fully digital and AI-driven bank. “Digital transformation, advanced analytics, data analytics, and big data are all linchpins in our journey to create a true single source of truth,” says Amith Rajan, Executive Vice President and Head of Advanced Analytics & Data, Mashreq Bank. “This allows us to continue being one of the region’s most progressive banks.”

In 2020, Mashreq launched its Digital Retail Banking Platform, converting over 70 percent of its customer journeys to digital. The next step—to optimize these digital experiences via a three-pillar approach. The first pillar is creating a rich data source that enables easy access and governance. The second is leveraging machine learning (ML) and ML Operations (ML Ops) to analyze and monetize the data. The final pillar is innovating their offers through the use of Azure Cognitive Services.

Creating a golden data source

“We’re a legacy bank, so we have lots of data silos and a unique footprint with multiple international offices and different types of technology,” Rajan explains. The bank addressed this hurdle by bringing everything into one space, accessible by all employees. Its hybrid solution meets regional banking regulations and enables easy data governance. Mashreq can now fully leverage its existing on-premise data storage with powerful data processing from the Azure cloud.

“This solution brings everything together—breaks down the data silos, works with different regulators, and navigates multiple technology landscapes. We are still on the journey and in 2023 we aim to use all of our data to understand and improve customer experience,” says Rajan.

Pinpointing what the customer needs

Pillar two leverages ML and AI to better understand customers’ needs and wants. “We want to understand our customers in all respects so we can better anticipate their needs,” explains Rajan. “We look at customer journeys from cradle to grave and find ways to generate more positive engagements for them.”

This started with the next best offer tool, a model that helps customer relationship managers anticipate what their clients will be interested in buying. Currently, the bank has 15-20 models that focus on wholesale banking products and many more in the retail space. Data is mined from the cross-functional databases and is used to suggest the next best product to recommend.

“This year, we managed to get about 20 percent to 30 percent acceptance rate on this model output,” Rajan shares. “We also successfully converted those opportunities to generate almost twice the amount that we originally invested. It’s worked well and we will build on this success significantly in the next year.”

Towards complex digital automation  

Cognitive services help the bank increase efficiency even further. With it, Mashreq hopes to automate some back-office operations and enable digital onboarding. Work is also underway to use voice-to-text services, voice enabled conversational intelligence, text analytics, and natural voice recognition to rate customer satisfaction. All this would be powered by Microsoft’s AI algorithm. “Most AI models require mind-boggling levels of training to see genuine results,” explains Rajan. “We are piggybacking off Microsoft’s pre-training so it’s a much smaller leap.”

Future projects for the bank include utilizing AI and ML to help with fraud detection and internal credit scoring for customers. Mashreq is also looking internally, focusing on employee satisfaction and retention, and talent acquisition using HR analytics.

Modernizing their home office is just the first step in Mashreq’s digital journey. Their next goal is to expand their digital footprint.

"We want to leverage the partnership both from a data point of view and an analytics point of view to make it available to the rest of our global network,” concludes Rajan. "This way, we can expand the infrastructure outside of our headquarters while still maintaining good governance.”