HSBC boss: AI agents may replace traders within four years

AI agents could replace some human financial traders within two to four years, according to Michael Roberts, head of HSBC’s investment bank. He linked the forecast to advances in artificial intelligence, quantum computing and distributed ledger technology.

Roberts described a future in which AI systems use quantum computing to assess complex market movements and trade bonds. Humans would still oversee the systems, he added.

“I think quantum computing will be able to look at very complex types of market movements, and with the advent of AI — which will make that even quicker — you can see agents trading bonds using quantum computing. That’s the future,” Roberts said in an interview.

He added that progress in the three technologies would change the pace, speed and scope of trading over the next two to four years.

HSBC recently announced cuts to human wealth management roles and plans to provide more digitally enabled products and services. The changes could increase the use of AI in customer interactions and wealth-related services.

In July, HSBC announced plans to establish an AI centre of excellence in Singapore and recruit 100 AI specialists. One of the centre’s initial areas of work was expected to be the use of AI in conversations during customers’ wealth management journeys.

ING has also been testing AI agents for work that previously required substantial staff involvement. In June, the bank reported that a pilot for an AI mortgage-processing agent had succeeded and would be extended. The system is operating in the Netherlands, where it reviews applications, explains possible outcomes and outlines next steps before a human employee makes the final assessment and decision.

Marco Li Mandri, ING’s change lead for scalable agentic products, said the bank had no planned job cuts linked to the mortgage project. He indicated that workforce reductions could occur in other areas.

AI is also being used to support investment banking work, including deal preparation. Bahadir Yilmaz, ING’s chief analytics officer, described transactions that can take six to 18 months and involve teams of 10 to 20 people conducting due diligence and analysis.

AI can prepare documents and research for early client meetings, reducing manual work before investment bankers begin discussions, Yilmaz said. Human bankers would continue to manage client relationships and make decisions on transactions.