Ryt Bank First Malaysian Bank Accepted at EMNLP 2025

Ryt Bank has achieved a historic milestone by becoming the first Malaysian financial institution to have its research paper accepted at the 2025 Conference on Empirical Methods in Natural Language Processing. The paper, titled Banking Done Ryt: Redefining Retail Banking with Language-Centric AI, marks a significant moment for Malaysia in the global artificial intelligence research community.
The acceptance at EMNLP, one of the most prestigious conferences for natural language processing research worldwide, positions Malaysia on the international stage for responsible and research-driven AI innovation in banking. EMNLP 2025 will take place in Suzhou, China from November 4 to November 9, 2025, bringing together leading researchers and practitioners in the field.
The research paper details how Ryt Bank has developed Ryt AI, a conversational banking system built on ILMU, Malaysia’s sovereign large language model. ILMU is a homegrown model with under 10 billion parameters and an 8K context window, developed through collaboration between YTL AI Labs and Universiti Malaya. The system enables customers to perform secure, real-time banking transactions using natural language commands such as transferring money to contacts simply by speaking or typing instructions.
Foong Chee Mun, Chief Product Officer of Ryt Bank, reflected on the achievement by noting that the team built their approach from scratch without an existing handbook for AI-powered banking. He emphasized that the acceptance by the global AI community proves Malaysian talent can define the future of banking. Foong previously co-founded MoneyLion and served as its Chief Technology Officer before joining the YTL-Sea digital bank project.
The Ryt AI system has been designed to meet strict financial regulatory requirements through a multi-layered security architecture. The system incorporates deterministic guardrails for compliance filtering, ensuring that all transactions meet regulatory standards before execution. Additionally, human-in-the-loop verification is required for every transaction, providing an extra layer of oversight and security.
The framework also features a stateless memory architecture that prevents data persistence, enhancing privacy protection for customers. These three protective layers work together to deliver fast, multilingual, and regulator-grade conversational banking while maintaining transparency and explainability. The paper demonstrates that this represents the first global regulator-approved deployment where conversational AI functions as the primary banking interface, rather than being limited to advisory or support roles.
The research emerged from a partnership uniting academic excellence and practical application. Ryt Bank collaborated with YTL AI Labs and Universiti Malaya to push the boundaries of AI implementation in financial services. Dr. Chan Chee Seng, a Full Professor at Universiti Malaya’s Faculty of Computer Science and Information Technology, emphasized that the milestone proves Malaysia can lead in both the science and safety of artificial intelligence.
Dr. Chan noted that by combining research, compliance, and innovation, the team is creating AI models that are globally relevant while remaining locally responsible. His research background includes extensive work in computer vision, machine learning, and image processing, with numerous publications in top-tier conferences. The ILMU project at Universiti Malaya began in early 2023 as a student-led initiative before evolving into a national project under YTL AI Labs.
The Ryt AI system is powered by four specialized agents that work together to orchestrate banking conversations. These include the Guardrails agent, Intent agent, Payment agent, and FAQ agent, each utilizing task-specific LoRA adapters attached to the ILMU model. The architecture replaces traditional multi-screen banking workflows with a single conversational dialogue, hosted entirely within the bank’s infrastructure to ensure consistent behavior and data sovereignty.
Ryt Bank officially launched to the Malaysian public in August 2025 as a joint venture between YTL Digital Capital and Singapore-based Sea Limited. The bank is licensed by Bank Negara Malaysia and covered by deposit insurance up to RM250,000 per depositor through PIDM. The platform offers savings accounts with competitive interest rates, a Visa-powered card with cashback benefits, and a PayLater feature providing instant credit.
The AI assistant currently supports Bahasa Malaysia, English, and Manglish, with plans to expand language support to Mandarin and other regional dialects. Beyond transaction processing, Ryt AI explains financial concepts in simple terms, aiming to improve financial literacy and accessibility for all Malaysians. The system is hosted on local YTL AI Cloud infrastructure, ensuring data remains within Malaysian borders and aligned with national data sovereignty principles.
This recognition at EMNLP 2025 validates Malaysia’s growing capabilities in developing sovereign AI technologies that can compete on the global stage. The research demonstrates that locally developed AI systems can meet international standards for both innovation and regulatory compliance. For a country positioning itself as an AI nation, having a Malaysian financial institution recognized at a premier international AI conference represents tangible progress toward that vision.
The Ryt Bank paper joins a select group of research accepted into the EMNLP 2025 Industry Track, which focuses on real-world applications of natural language processing technologies. The paper has been published on arXiv under identifier 2510.07645 and is available for the research community to review. The work showcases how academic research partnerships with industry can produce practical innovations that advance both technology and financial inclusion.
As digital banking continues to evolve in Malaysia and across Southeast Asia, Ryt Bank’s approach of making AI the core differentiator sets it apart from competitors. While other digital banks focus on mobile-first delivery and ecosystem integration, Ryt Bank has positioned artificial intelligence as its primary value proposition. The EMNLP acceptance provides external validation that this strategy is built on solid scientific foundations and represents a meaningful contribution to the field of AI-powered financial services.



