DBS bets on AI to break advice bottleneck

Faster preparation could widen client coverage, but portfolio suitability remains the industry’s hard limit.

DBS Bank is using agentic artificial intelligence to attack one of wealth management’s biggest constraints: the time advisers spend preparing market views instead of speaking with clients.

The bank says the technology can compress hours of preparation into 15 minutes, allowing relationship managers to cover more clients. But the test is whether that speed produces personalised advice rather than faster distribution of generic trade ideas.

“AI is the best assistant that you can ever have,” said Wilfred Quek, Managing Director, Group Head of Wealth Management Solutions Global Financial Markets at DBS Bank.

DBS is applying AI to market summaries, views and trade idea generation. The aim is to free advisers from repetitive work and redirect time toward portfolio decisions.

“The misconception is that once there's AI, there's less need for human interaction,” Quek said.

That distinction matters because a trade idea that suits one client may be inappropriate for another. Wealth managers must still assess holdings, risk exposure and suitability before acting.

“The front line must look at the client's portfolio to decide whether it's suitable for this client,” Quek said.

DBS wants AI to perform that first layer of analysis by matching investment ideas with portfolios. The system could flag which clients should receive an equity, rates or foreign-exchange recommendation and which should not.

The model promises scale, but it also raises the standard for accuracy and oversight. Any error repeated across portfolios could magnify suitability risks.

Quek said the objective is to use AI to create capacity while preserving adviser judgment. “That should change the nature of how business is being done, and how your client can be better served.”

For DBS, the edge will not come from producing more ideas. It will come from proving that AI can make advice faster without making it less personal.