UOB deploys A*Star's Meralion AI to support relationship managers in five markets

In a branch in Penang or Bangkok, a client switches from English into Hokkien halfway through a question about an investment, and the relationship manager across the desk nods while quietly reconstructing what was just asked. For decades, banks have handled that moment through hiring, staffing branches with people who share the client's language and trusting them to carry product knowledge and compliance obligations in their heads. UOB is now testing a different answer: an AI system that sits inside the conversation itself.
The bank is deploying a multilingual solution built on Meralion, a model developed by the Agency for Science, Technology and Research (A*Star), to support its relationship managers across Singapore, Malaysia, Thailand, Indonesia and Vietnam. The model handles English, Malay, Thai, Vietnamese and Mandarin, plus Cantonese and Hokkien. It interprets customers' questions across languages and accents, then surfaces prompts with relevant information such as key product details and risk disclosures. The stated purpose is to complement the RM's judgment, not replace it.
What changes for the relationship manager
The interesting shift is in the craft of the job rather than the technology. An RM used to manage one channel: the person in front of them. With a co-pilot listening alongside, there are two. The RM must hear the client, glance at a prompt, decide whether it fits and fold it into a reply without sounding scripted. That is a new skill, closer to an interpreter's split attention than to classic relationship banking.
It also moves where risk sits. A prompt that places a risk disclosure in front of the RM at the right moment could make advice more consistent across five markets. A prompt that is ignored, or followed too readily, creates a different exposure. Leong Yung Chee, chief financial officer of UOB, framed the intent carefully: "We want to apply AI with purpose, and we see potential for this solution to help our RMs better understand customers, navigate language and cultural nuances, and provide clearer, more relevant guidance."
Culture as the measure
The initial phase sits under Project Aura, UOB's programme to scale AI adoption and foster a culture in which employees are empowered to use AI, with RMs named as a priority domain. That framing matters. A model can interpret Thai flawlessly and still fail if the people meant to use it treat it as surveillance or noise. The meaningful test is whether RMs trust the prompts enough to use them and stay sceptical enough to override them.
UOB says the pilot will be assessed on whether the solution can scale across the bank's markets and whether it strengthens governance and oversight. There are no results to report yet. The bank has also signed a memorandum of understanding with the A*Star Institute of Advanced Intelligence and Computing to explore further uses of Meralion. Professor Lim Keng Hui, assistant CEO of the agency's Science and Engineering Research Council, said the parties are also "exploring other emerging technologies that could strengthen decision-making, resilience and innovation in financial services".
What this means for L&D
Putting a co-pilot into live client work is a capability programme before it is a technology programme. Training cannot stop at how the tool works. RMs need practice holding a conversation while a second voice offers suggestions, rehearsal in spotting when a prompt is wrong and clarity on who owns the advice when the machine contributed to it. Role play with the tool running, reviewed against the disclosures it surfaced, will teach more than any demo.
Governance and adoption belong in the same curriculum. If the pilot is judged on oversight, frontline staff need to know what oversight looks like in practice, including when to flag a bad prompt. Organisations building this fluency, as we do through our AI Training, should measure success the way UOB says it will: by whether people use the tool well under pressure, across markets, with accountability intact.
Sources:
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.
Keep reading.

UOB is putting AI skills for 32,000 staff behind a S$900 million target

DBS chatbots move its 1,600 call centre staff onto corporate calls too

More than 60% of Singapore Airlines staff now use its internal AI platform monthly


