DBS Bank discovers big data challenges and solutions with AI.

DBS Bank discovers big data challenges and solutions with AI.

DBS Bank’s Journey to Adopting Artificial Intelligence

DBS Bank, one of Singapore’s leading banks, has embraced the challenge of integrating artificial intelligence (AI) into its operations, realizing that success goes beyond simply figuring out the training models. Sameer Gupta, the bank’s Chief Analytics Officer, revealed that data proved to be a major barrier during the AI adoption process. In 2018, DBS Bank embarked on a journey to leverage AI in four primary areas: developing analytics capabilities, fostering a data culture and curriculum, upskilling employees in data-related fields, and enabling effective data utilization.

Gupta explained in an interview with ENBLE that the bank’s vision was to enhance organizational benefits by using data to drive AI-driven initiatives. For this to happen, DBS Bank recognized the need to make AI accessible throughout the entire company, deliver economic value, and constantly reduce the cost of AI solutions.

To achieve their goals, DBS Bank focused on developing the right use cases and talent, including machine learning engineers, and cultivating a data culture that encouraged all employees to embrace data and AI in their work. This included providing training programs to guide staff on when and how to use data effectively.

Establishing the necessary infrastructure was a crucial step in DBS Bank’s AI adoption process. This involved setting up a robust data platform, a data management structure, and data governance mechanisms. DBS Bank implemented a framework, called PURE, which stands for purposeful, unsurprising, respectful, and explainable. This framework guides the bank in using data responsibly and ensures that all data use cases are assessed accordingly.

DBS Bank’s data platform, ADA, acts as a central repository for all relevant data. It enables better data governance, quality assurance, discoverability, and security. Today, more than 95% of the data required for DBS Bank’s AI-powered operations is discoverable on the ADA platform. The platform currently holds over 5.3 petabytes of data, encompassing 32,000 datasets consisting of videos and structured data.

However, Gupta revealed that getting to this point was a tremendous challenge. Organizing and making the data discoverable required significant manual effort and expertise, as tools for automating these tasks were lacking. Moreover, DBS Bank used multiple applications, each holding data essential for their AI initiatives. This necessitated considerable effort to consolidate datasets onto a single platform while ensuring secure data extraction and usage.

Despite these challenges, DBS Bank successfully runs over 300 AI and machine learning projects, contributing to a revenue uplift of SG$150 million ($112.53 million) and savings of SG$30 million ($22.51 million) in risk avoidance in 2022. These projects span various functions, such as human resources, legal services, and fraud detection.

Looking ahead, DBS Bank anticipates generating even greater economic value and cost avoidance benefits through AI, with a target of SG$350 million ($262.56 million) this year and SG$1 billion ($750.17 million) within the next three years. With around 1,000 data engineers, data scientists, and data engineers, DBS Bank is continuously expanding its AI capabilities to better serve its five million customers across the region.

When asked about DBS Bank’s exploration of generative AI, Gupta confirmed they were already running more than 10 pilots. However, he emphasized that it was still early days for these projects, with further conversations needed to assess the benefits of generative AI across various teams, including marketing, sales, and IT. Adherence to DBS Bank’s PURE principles and Singapore’s FEAT principles, which guide the responsible use of AI, is of utmost importance.

Gupta stressed that the number of AI models in use is immaterial; what matters is achieving optimal efficiency and accuracy. He highlighted the misconception that the model itself is all-important, emphasizing the need for a holistic approach. This entails working through technical elements, constantly monitoring AI usage, and gathering feedback to ensure continuous improvement.

Addressing recent service outages experienced by DBS Bank, Gupta acknowledged the need to identify areas for improvement in effectively anticipating and managing such disruptions through AI and data analytics. A special committee, comprising four of the bank’s board members and external experts, is conducting a full review of the company’s technology resiliency. Gupta expressed DBS Bank’s commitment to providing uninterrupted digital banking services and surpassing customer expectations.

Singapore is actively supporting the financial sector’s innovative efforts through technology adoption. The Financial Sector Technology and Innovation Scheme (FSTI 3.0) sets aside SG$150 million ($112.53 million) over three years to drive capability development and adoption in AI, data analytics, and regulation technology. DBS Bank aims to industrialize its AI development and deployment processes, ensuring scalability and accessibility. Measuring the positive outcomes of AI and its benefits to employees and customers remains a top priority.

Gupta concluded by emphasizing that there is no magic bullet for successful AI adoption. Perseverance, continuous learning, and agility are key to unlocking the full benefits. As DBS Bank continues its AI journey, it is committed to meeting the evolving needs of its customers and exceeding their expectations with innovative solutions.