Key results
The company
Hong Leong Bank
hlb.com.myCommercial bank providing retail, corporate, and digital financial services.
Result highlights
- Chatbot accuracy increased from 75% to 99%
- 70% of chat volume handled by AI
- 30% efficiency gain in trade operations
The story
A 120-year-old financial institution providing comprehensive banking services across Singapore, Hong Kong, Vietnam, Cambodia, and China with 12,000 employees.
An initial open-source chatbot achieved only 75% accuracy due to static knowledge bases, leading to slow response times and limited transaction capabilities. Additionally, reliance on manual processes for trade finance documents and HR inquiries created significant operational bottlenecks.
The bank implemented an agentic RAG framework on Vertex AI, where Gemini commands specialized agents to execute secure backend transactions and dynamic website searches. Document AI automates the processing of high-volume trade documents, while a private instance deployment ensures data governance across customer and internal tools. The system integrates directly with core banking APIs, allowing the AI to handle complex tasks like credit checks.
Scope & timeline
- 10-25% reduction in development time
- 3x increase in digital chat transactions
Quotes
“We've achieved a new level of scalability and efficiency with Google Cloud and Gemini. By eliminating manual infrastructure management and seamlessly integrating with our core systems, we can dedicate 100% of our focus to AI development. This allows us to rapidly deliver smarter, faster services for our customers and employees.”
Implementation partner
Cloud Space
cloud-space.coProvided expertise in cloud computing and AI automation for HLB's digital transformation.