Singapore banks up AI game in apps with expenses tracking for retail customers

SINGAPORE – Tracking your monthly dining, wellness or overseas spending is now as simple as asking a chatbot. With the launch of a new artificial intelligence tool in digital-only lender Trust Bank’s app, these once-manual tasks are now fully automated.

DBS Bank and OCBC Bank have plans for similar roll-outs from later this year. These AI-fuelled upgrades are the latest attempts by local banks to automate and lift customer service.

Announced on Sept 8, Trust Bank’s new AI chatbot feature in its banking app lets customers ask questions related to their card transactions in the way they normally converse. These include spending with specific merchants or overseas while on holiday, as well as spending on categories like transport and dining.

Called Trust AI Ask, the feature seeks to help customers save time from manually navigating menus, search filters or months of transaction history, said the bank, which is backed by Standard Chartered Bank and FairPrice Group.

The bank plans to extend the AI chatbot to other banking activities, including those involving information about customers’ accounts, transfers and rewards.

DBS is adding similar AI capabilities in its banking app from as early as November to allow retail customers to chat with its digibot to check refunds or reward points on their bank cards, request fee waivers, and block or replace cards.

The feature was rolled out to DBS corporate customers in July. The DBS digibot also got an upgrade in August that lets retail customers ask the virtual assistant wealth management-related queries.

Similarly, OCBC announced in July its new AI-powered app, OCBC WoW, which uses the digital likeness of employees as virtual client managers to provide wealth management services to customers.

Using voice or text, customers interact with the app’s virtual managers, which analyse customers’ investments to help them better understand their underlying exposures and sources of risk. The app is being tested among selected clients and OCBC employees.

From the end of 2026, OCBC will test an AI-powered digital companion on its banking app with selected retail banking customers. Customers can use text to interact in a conversational manner with the AI companion to get personalised transaction and banking relationship insights about account balances, transaction history, payments and other account-specific requests.

Tamma Febrian, director of financial institutions ratings at Fitch Ratings, said: “By deploying technology capable of analysing live data sets and offering more natural-language capabilities, banks can make the customer experience more seamless.”

Previous chatbot and investment advisory functionalities were designed around repetitive queries and a narrow set of rules and scenarios, which some customers found less helpful, he added.

For customers, the AI features offer convenience, said Zennon Kapron, director of fintech and payments consulting firm GL Insight. This includes potential time savings, such as requiring fewer steps to get their answers.

Banks benefit, too. Analysts said that lenders enjoy cost savings if AI can, for example, help resolve customer queries that are not routed to customer service or relationship managers. Customer queries also provide data – like what customers care about – that banks can act on, such as by cross-selling services.

Still, the AI features in bank apps are not entirely new, and the banks know it, said Kapron. He pointed out how category breakdowns, spend trackers and investment nudges have been in banking apps since UOB launched its AI-powered Mighty app in 2019.

Phillip Securities research manager Glenn Thum said what has changed is how customers interact with the apps. Instead of a bank deciding which spending categories to show customers, customers ask questions in their own words to get a reply. This sounds incremental but it lowers the barrier to entry for customers to understand their financial data, he said.

The cost of putting a natural-language layer over existing data the banks hold also fell with AI. “That is why three banks can have AI features within a few months of each other,” said Kapron. “The transaction data has been sitting in these apps for a decade and what the banks have built is a new front door on a house they finished years ago.”

Thum added that low interest rates have compressed net interest margins, so AI in bank apps is a cheaper way to help boost lenders’ earnings growth through cost savings, wealth management services and fee income from bank services.

But as adoption of these AI features becomes more widespread across the sector, Febrian said they become “an increasingly expected utility for banks to offer, rather than a structural competitive advantage that could meaningfully boost their franchises”.

Agreeing, Thum added: “What really matters is whether the AI initiatives translate into good cost-to-income ratios, product sales or deposit stickiness.”