Cake Digital Bank builds 'Next GenAI Bank' with Vertex AI and Gemini Flash
Cake Digital Bank, headquartered in Vietnam, aimed to become a leading 'Next GenAI Bank' by integrating AI across its operations to improve personalized customer service, risk management, and regulatory compliance. They built an omni-channel chatbot powered by Google Cloud Vertex AI and Gemini Flash for personalized natural language interactions, and developed an AI biometric eKYC system certified and used by over 15 million users in Vietnam. The bank deployed 50 AI models using Vertex AI and BigQuery for credit scoring, fraud detection, and customer interaction, which accelerated product development cycles from months to days. Their architecture was based on Google Kubernetes Engine providing scalable microservices, enhanced customer engagement, and operational efficiency. Cake Digital Bank partnered with external sectors including entertainment, retail, and insurance to build a digital ecosystem providing customers with tailored financial products.
- Organization
- Cake Digital Bank
- Industry
- Finance
- Location
- Vietnam
- Published
- June 2024
Reported outcomes
Strategic outcomes
Primary read
Use case focus
Showing 3 of 3
- 1Conversational AI
- 2Biometric Authentication
- 3AI Model Development and Deployment
- The bank needed to rapidly implement AI for personalized services while ensuring strict compliance with Vietnam's biometric regulatory requirements for customer authentication.
- They faced challenges in speeding up AI model development, fraud detection, loan approvals, and customer engagement while scaling operations to millions of customers.
- Cake Digital Bank utilized Google Cloud's Vertex AI, Gemini Flash, BigQuery, Cloud Run, and Google Kubernetes Engine (GKE) to build and deploy AI models and digital services.
- They developed a biometric authentication system (FaceAuthen) using Vertex AI to comply with national regulations, which is now widely adopted in the market.
- An omni-channel AI chatbot was created using advanced LLMs (PaLM 2 and Gemini models) in Vertex AI, continuously learning from interactions to personalize user experiences.
- The bank employed a scalable cloud-first architecture enabling fast feature releases and AI-driven product innovation at scale.
- Enabled loan and credit card approvals within two minutes, improving customer experience.
- Achieved significant growth with 4.6 million customers served in under 4 years and recognition as Best AI Bank by The Asian Banker in 2024.
- Reduced product development cycles to days and deployed 50+ AI models across diverse banking functions.
- The FaceAuthen biometric system is used by 15+ million Vietnamese users, supporting regulatory compliance.
- Won Best AI Technology implementation award adding to brand prestige.
Architecture
The architecture leverages Google Kubernetes Engine for scalable microservices handling, Vertex AI for AI model building and deployment, BigQuery for AI-powered data analytics, Cloud Run for serverless execution, and Gemini Flash as the generative AI model powering the chatbot and biometric authentication solution.
Sources & evidence1
AI-generated summary. Verify important details with the linked sources before relying on this case.