CHIP

18.4191.80%

UOB explores A*STAR's multilingual AI for client calls

LongbridgeAII'm LongbridgeAI, I can summarize articles.

It can interpret questions and concerns across different languages and accents.

UOB is exploring the use of an artificial intelligence (AI)-powered multilingual customer call solution to enhance customer engagement across ASEAN languages and markets.

The bank said it looks to apply the solution, powered by A*STAR's Multimodal Empathetic Reasoning and Learning in One Network (MERaLiON) model, to help relationship managers (RMs) understand, engage, and advise customers in Singapore, Malaysia, Thailand, Indonesia, and Vietnam.

The first phase will be done through Project AURA, the bank’s strategic programme to deploy advanced AI at scale and build an AI-fluent culture.

The solution, called the Compliance Intelligence Engine and Evidence-Grounded Post-Call Compliance Review System, can support relationship managers during customer conversations by interpreting questions and concerns across different languages, accents, and communication styles.

It can provide relevant prompts on considerations such as key product information and risk disclosures to complement relationship managers’ professional judgement and help them deliver clearer and more relevant customer conversations.

After each conversation, the solution can support an evidence-grounded review to assess whether key information was communicated clearly and fair dealing standards were met, providing an additional layer of governance for the bank.

It could also improve productivity by reducing manual translation and automating parts of the post-call review process, whilst providing insights for more targeted relationship manager coaching.

Relationship managers would remain accountable for the conversations and advice they provide, with AI complementing their professional judgement.

“Banking is built on trust, and every customer conversation matters,” said Leong Yung Chee, group chief financial officer at UOB.

“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,” Leong added.

MERaLiON supports multilingual processing and code-switching across 11 ASEAN languages, including Malay, Bahasa Indonesia, Thai, and Vietnamese, alongside English, Mandarin, Singlish, and traditional dialects such as Cantonese.

The model has achieved the lowest average word-error rate among open-source models across seven languages, according to UOB and A*STAR.

UOB said it will continue to assess how the model's capabilities can be applied safely and responsibly across its regional franchise.

Source: Singapore Business Review The copyright of this article belongs to the original author/organization.

The views expressed herein are solely those of the author and do not reflect the stance of the platform. The content is intended for investment reference purposes only and shall not be considered as investment advice. Please contact us if you have any questions or suggestions regarding the content services provided by the platform.