How DBS, Southeast Asia’s largest bank, is capturing the full value of AI and Machine Learning in Singapore

Estimates by The McKinsey Global Institute in 2024 suggest that AI could add up to US$340 billion (S$450.7b) in annual value every year to the global banking industry¹.

To this end, Singapore-headquartered DBS bank had the foresight to seize the advantages afforded to pioneers and early adopters of digital technologies, like AI, a full decade before. And the bank accomplished this from an established tech node in Southeast Asia — Singapore.

DBS’ Chief Data and Transformation Officer, Nimish Panchmatia explains: “Singapore has advantages [such as] a world-class education system with leading institutions of higher learning, strong investments in R&D, and a supportive ecosystem enabled by progressive regulations that enable innovation to thrive.”

Incidentally, DBS’ journey in AI adoption has already begun yielding positive results. The bank has identified over 350 AI use cases and 800 models implemented to date. 

It had also reported a significant economic impact of S$370 million in cost savings and value-add derived from AI in 2023 alone. The bank's success story underscores the potential of AI to drive innovation, enhance customer experiences, and deliver tangible business value.
 

BUILDING AI WITH ‘PURE’ INTENTIONS

But these successes did not happen overnight, nor did the bank adopt a broad-based approach to investing resources in all identified use cases at once. Instead, the bank ensures its AI initiatives are closely tied to its overall business objectives and this is further complemented by a framework for approving use cases.

As AI solutions rely heavily on vast amounts of data, the bank needed a solution to help its groups and divisions navigate vital decisions concerning data usage and use case development. This culminated in the bank’s "PURE" framework in 2018, which ensures that DBS only embarks on formulating AI use cases that are:

• Purposeful in its use of data.
• Unsurprising in its use of data from the customer’s (organisation’s and/or individual’s) point of view.
• Respectful in its use of data with due consideration given to social norms.
• Explainable in its rationale for the use of data; the how and why of the AI model’s data use should be explainable to the bank’s customers.

Nimish notes that since its implementation, all bank employees must undertake a mandatory e-learning curriculum to familiarise themselves with the PURE framework.
 

THE IMPORTANCE OF A DATA CHAPTER

In tandem with its vision of becoming an AI-powered bank, DBS launched a Data Chapter in 2023 —bringing together 700 data professionals across its various divisions to foster a culture of continuous improvement and innovation to scale AI.

The Data Chapter consolidates both resourcing and decision-making in AI and ML adoption which enhances efficiency in adoption. DBS’ data professionals are embedded across various units to form cross-functional, inter-disciplinary “squads” to enable sharing of knowledge.

Nimish explains: “The Data Chapter model—which embeds data professional in the units whilst keeping a central view for skilling and resourcing—enables the diffusion of knowledge and facilitates AI industrialisation across all aspects of the bank’s operations.”

Aside from forming squads, DBS has also implemented a comprehensive training curriculum that caters to all employee skill levels from novices to data experts looking to sharpen their skills.

At the recent Fortune Brainstorm AI Singapore conference in July, Tan Su Shan, Deputy CEO and Group Head of Institutional Banking at DBS, noted that the priority in adopting new technologies is to secure buy-ins from employees and giving them tools to act responsibly in implementation. She said: “Let them own the model. Let them own the feedback loop. Let them own the outcomes.”

And the bank is walking the talk, giving employees a variety of options to learn at their own pace and select what they need to improve on from a range of online courses, workshops and community programmes. Since 2021, over 9,000 employees have taken upskilling courses in data and AI.
 

DBS’ AI USE CASES AND ITS BENEFITS

Going forward, DBS remains focused on scaling up its employees’ data and AI capabilities and streamlining its development and deployment processes through its unique AI framework. 

Nimish explains: “Our experience in AI and our robust governance structure will help us balance reaping the benefits of Gen AI while managing the risks of this still-emerging field.”

All in all, among Singapore’s local enterprises, the bank offers a clear vision for what companies can also achieve in their AI journeys.

Help and support is also readily available for companies — no matter the size — in Singapore to turn their AI adoption journeys into a value-added service or product offering. Actively engaging in public-private collaborations can be a good starting point.

Programmes like Google Cloud’s AI Trailblazers Innovation Sandbox and tools like the AI Verify Toolkit can enable business to test, validate and progress their AI use cases in accordance with industry best practices and standards.
 

Footnotes:

¹ McKinsey & Company. "Scaling Gen AI in Banking: Choosing the Best Operating Model". McKinsey & Company, 26 Mar 2024,
https://www.mckinsey.com/industries/financial-services/our-insights/scaling-gen-ai-in-banking-choosing-the-best-operating-model. Accessed 6 Aug 2024.