The S Curve
Tue, 22 SeptMelbourne · Singapore · New York
Book a call
Customer Experience

Trust Bank shipped AI Ask. DBS and OCBC are racing the same retail front door.

Trust AI Ask is live. DBS digibot from November. OCBC WoW already in test. Three clocks. Not one Singapore-banks blob.

The S-Curve··3 min read

Trust Bank shipped AI Ask. DBS and OCBC are in the same queue.

The transaction data has been sitting in these apps for a decade. You've probably asked a banking chatbot a simple spend question and been walked through a tree of menus that already knew the answer. Trust Bank, backed by Standard Chartered and FairPrice Group, put a new front door on a house it finished years ago. That is the whole story. Everything else is a product name.

Trust announced Trust AI Ask. You ask the app about card spend the way you would ask a person: this merchant, that holiday, dining versus transport. No menu archaeology. They want to extend it to accounts, transfers and rewards.

DBS is putting similar muscle on digibot for retail from as early as November: refunds, reward points, fee waivers, block or replace a card. Corporates got it in July. August added wealth queries for retail. OCBC already has WoW, digital likenesses of staff as virtual client managers for wealth, in test with selected clients and employees. From the end of 2026 it will test a text companion on the banking app for balances, history, payments and the other account chores.

The named people are the analysts. A Trust, DBS or OCBC executive is not quoted. Their reading of the product is more useful than a line about delight.

The front door is new. The house is not.

A new front door on a house they finished years ago.
A new front door on a house they finished years ago.

Tamma Febrian, an analyst at Fitch, said live data plus natural language can make the experience more seamless. Old chatbots were a tree of repetitive queries. Zennon Kapron, founder of GL Insight, said there are fewer steps for the customer, and also this, which is the sentence I would take to a product committee: category breakdowns and spend trackers have been in apps since UOB's Mighty in 2019. The transaction data has been sitting there for a decade. What the banks built is a new front door on a house they finished years ago. The cost of that layer fell. That is why three logos can ship within a few months of each other.

Glenn Thum, an analyst at Phillip Securities, said customers now ask in their own words instead of living inside the bank's chosen categories. Compressed net interest margins make this a cheaper lever for earnings via cost, wealth and fees. Then the point that actually matters: cost-to-income, product sales or deposit stickiness. Febrian agrees the features become an expected utility, not a franchise-defining advantage, once everyone has them.

If you run a bank in Singapore or Hong Kong, this is not an AI training story yet. It is a customer-experience and product story that will become a training story the week a customer asks the bot something the policy cannot answer. Someone has to own the prompt, the refusal and the escalation. That someone is not the vendor. Put the skills work on the exception path, not on a demo of coffee spend.

Kapron's UOB Mighty 2019 point is the one I would put in a board appendix. Spend trackers are old. Natural language is new and cheap. That is why Trust, DBS and OCBC can land in the same season without anyone having invented a new data warehouse. If your transformation deck is still "we need to collect the data," you are solving 2016. The fight is the front door, the policy when the bot is wrong and whether fee waivers and card blocks are actually safe to put in a chat.

Febrian's "expected utility" line is the strategy kill. Do not fund this as a moat. Fund it as table stakes and measure Thum's list: cost-to-income, product sales, deposit stickiness. If you cannot pick one, you are building a demo for the annual report.

Trust AI Ask is live as of 8 September, at least for card questions, with a stated plan to grow into accounts, transfers and rewards. DBS went to corporate customers in July, added wealth queries in August and plans retail from November. OCBC is testing wealth likenesses now and a retail companion from the end of 2026. Three banks, three sequences. A programme plan should keep them separate.

What this means for L&D

Ship the natural-language front door. Measure stickiness, not chatbot satisfaction. Trust is card questions from 8 September. DBS is July corporate, August wealth, retail from November. OCBC is WoW now and a companion late 2026. Three sequences, not one Singapore-banks-go-AI blob. The training job arrives on the exception path, when the policy cannot answer and a human has to take the chat.

Sources: Straits Times, 12 September 2026.

From The S Curve

News and insights for innovation, digital transformation, future of work and L&D leaders.

Stay ahead of learning and development, corporate innovation and digital transformation news. Plus the future of work. For leaders in AU, NZ, HK, SG, the US, the UK and Canada.