Global Fintech Edge – Innovative Financial Technology Solutions

Skip to content

UOB Explores MERaLiON AI for Multilingual Banking Calls

  • News
  • November 18, 2026

United Overseas Bank (UOB) is exploring a multilingual AI-powered customer call intelligence solution developed by Singapore’s Agency for Science, Technology and Research (A*STAR), using its MERaLiON model to support relationship managers across Singapore and five ASEAN markets. The initiative will examine how AI can interpret multilingual customer conversations, surface relevant product and risk-disclosure prompts, and review calls after they end. The collaboration, supported by a new memorandum of understanding between UOB and A*STAR’s Institute of Advanced Intelligence and Computing, forms part of the bank’s Project AURA programme to deploy AI across its operations.

UOB Explores MERaLiON AI to Support Banking Calls Across ASEAN

Multilingual customer conversations create a practical challenge for banks operating across Southeast Asia. Customers may switch between languages, use regional dialects, or express financial concerns through local terminology. For relationship managers, understanding those conversations accurately is essential not only for customer service but also for explaining financial products, communicating risks, and meeting compliance obligations.

United Overseas Bank (UOB) is exploring an AI-powered solution developed with Singapore’s Agency for Science, Technology and Research (A*STAR) to address this challenge. The system uses MERaLiON, A*STAR’s multilingual AI model, to support customer conversations and post-call compliance reviews. UOB says it is the first bank to explore this particular application.

The initiative will initially be assessed through Project AURA, UOB’s programme to deploy advanced AI across the bank and build AI capabilities among its employees. UOB and A*STAR’s Institute of Advanced Intelligence and Computing (A*STAR IAIC) have also signed a memorandum of understanding to explore how the technology can be enhanced for safe and scalable use in banking.

MERaLiON targets multilingual customer interactions

UOB serves more than eight million customers across ASEAN, including Singapore, Malaysia, Thailand, Indonesia, and Vietnam. Conversations across these markets can involve different languages, accents, local expressions, and code-switching within the same exchange.

MERaLiON, short for Multimodal Empathetic Reasoning and Learning in One Network, is designed to process multilingual speech and language. UOB says the model supports 11 ASEAN languages and language varieties, including English, Mandarin, Malay, Bahasa Indonesia, Thai, Vietnamese, Cantonese, and Singlish.

The proposed application is called the Compliance Intelligence Engine and Evidence-Grounded Post-Call Compliance Review System. It is intended to help relationship managers interpret customer questions and concerns, while providing prompts about relevant product information and risk disclosures during conversations.

This could be particularly useful when customers discuss financial products in a mix of languages or use expressions that are difficult to interpret through conventional speech-recognition systems.

The technology is intended to complement, rather than replace, relationship managers’ professional judgement. Staff remain responsible for the advice they provide and the decisions made during customer interactions.

AI-supported compliance reviews after customer calls

The proposed system also extends beyond live conversation support. After a customer call, it can help review whether important information was communicated clearly and whether fair-dealing standards were followed.

An evidence-grounded review could give compliance teams an additional way to examine customer interactions, identify potentially missed disclosures, and support more consistent oversight.

The system may also reduce manual translation work and automate parts of post-call review. Its analysis could provide managers with information for more targeted relationship-manager coaching.

This creates two related use cases: supporting employees during customer conversations and helping the bank assess those conversations afterwards. The combination is relevant to financial institutions where communication quality and disclosure obligations are closely linked.

However, AI-generated prompts and call assessments will need to be evaluated for transcription accuracy, language coverage, context recognition, and the risk of incorrect compliance flags. A system that misunderstands a dialect or misses a product-related expression could affect the quality of its recommendations.

UOB and A*STAR have therefore framed the collaboration around further testing and enhancement, rather than presenting the solution as a fully deployed banking product.

Singapore’s financial sector is expanding AI research partnerships

The collaboration is part of a wider effort to move AI research into practical financial services applications. A*STAR’s work with UOB builds on its broader partnership with the Future of Finance Institute (FFI), established by the Monetary Authority of Singapore (MAS).

FFI began operations in November 2026 with an initial focus on AI in finance. Its programmes are intended to help financial institutions share implementation experience, develop practical AI capabilities, and strengthen risk management practices. Monetary Authority of Singapore+1

On November 18, MAS and FFI also announced initiatives under Singapore’s National AI Missions for Financial Services. These include work on financial crime prevention, safe AI-agent transactions, and the development of AI capabilities for financial institutions. The speech-to-text work involving A*STAR IAIC is among the areas being advanced toward practical financial-sector applications.

This policy environment gives the UOB collaboration broader significance. Rather than focusing exclusively on general-purpose chatbots, financial institutions and research organisations are examining how AI can support specific operational needs where accuracy, explainability, and governance matter.

Human oversight remains central to AI-enabled banking

The potential value of MERaLiON lies in applying multilingual AI to a specific banking workflow. Relationship managers need to understand customers accurately, explain financial products clearly, and communicate relevant risks. Compliance teams need evidence that important information was conveyed appropriately.

Combining conversational support with post-call review could help connect these responsibilities, while reducing some of the administrative work involved in reviewing multilingual interactions.

There are still implementation questions. The bank will need to assess how the system handles regional accents, code-switching, banking terminology, sensitive customer information, and differences in regulatory expectations across markets. It will also need to establish how AI prompts are presented to staff and how post-call assessments are validated.

Reporting by The Straits Times on November 18 indicated that UOB had demonstrated an early version of the tool using a simulated multilingual call. The bank had not set a firm deployment date, with further development and adaptation to its terminology still required.

For UOB, the project is part of a wider effort to make AI useful in day-to-day banking operations while keeping employees central to customer relationships. For A*STAR, it offers an opportunity to test locally developed multilingual AI against the linguistic and regulatory demands of a major regional bank.

The outcome will depend on whether the technology can reliably interpret real customer conversations and produce useful, evidence-backed compliance insights across different ASEAN markets. Its significance is not simply that AI can listen to banking calls, but that it may help institutions make multilingual customer engagement more consistent without removing human accountability.

Market Landscape

AI adoption in banking is expanding beyond customer-service chatbots into employee assistance, compliance monitoring, risk management, and operational decision support.

UOB’s MERaLiON initiative reflects several developments in financial technology:

  • Multilingual banking AI: Speech and language models are being adapted to understand regional languages, accents, dialects, and code-switching in customer interactions.
  • Compliance intelligence: AI can help identify whether required product information and risk disclosures were communicated during customer conversations.
  • Relationship manager enablement: AI-supported prompts and post-call insights can assist employees without replacing their responsibility for customer advice.
  • Evidence-based monitoring: Automated call reviews can provide additional information for compliance teams and employee coaching.
  • Locally developed AI: Singapore’s financial sector is exploring homegrown AI capabilities through partnerships between banks, research institutions, and MAS-backed initiatives.

The commercial opportunity will depend on model accuracy, explainability, data protection, and the ability to adapt AI systems to different regulatory and linguistic environments.

Top Insights

  • UOB is exploring A*STAR’s MERaLiON model to support multilingual customer conversations and relationship manager workflows across five ASEAN markets.
  • The proposed system can provide prompts on product information and risk disclosures while relationship managers retain responsibility for customer advice.
  • Post-call reviews could help assess communication clarity, fair-dealing standards, and potential compliance gaps while supporting employee coaching.
  • The collaboration between UOB and A*STAR IAIC aims to improve the model’s suitability for safe and scalable financial-sector deployment.
  • The initiative forms part of Singapore’s broader effort to develop and test AI technologies against practical financial services requirements.

Get in touch with our fintech expert

Related Posts

  • News
  • November 18, 2026
  • 25 views
Visa Expands Agentic Ready Program to Merchant Payments

Visa is expanding its Agentic Ready program beyond card issuers to include merchants, payment service providers (PSPs), acquirers, and business-to-business (B2B) commerce…

  • News
  • November 18, 2026
  • 17 views
Republic Bank Launches Earned Wage Access Service

Republic Bank has launched Republic Bank EarlyFunds, a digital banking service that allows eligible customers to access up to $1,000 of wages…

Leave a Reply

Your email address will not be published. Required fields are marked *

Guest Column

Loading articles...

SUBSCRIBE

Get the latest insights and updates

delivered to your inbox.

Newsletter Signup

You have successfully subscribed to the newsletter

There was an error while trying to send your request. Please try again.

Global FinTech Edge will use the information you provide on this form to be in touch with you and to provide updates and marketing.