Reducing workload and making information more accessible to boost productivity and increase customer satisfaction.
Erste is a top 3 player in the Hungarian banking scene, with nearly 100 branches all across the country. They were looking to boost information flow within the company to improve employee access to product details. Using an agentic RAG, we created an enterprise-ready system that allows clerks to find specifics of banking products with ease.
documents processed
higher accuracy
speed to results
In the past, clerks turned to specialists with product-related questions, such as financing details or interest rates. This often caused an immense workload for specialists. In turn, the slowdown in the processing time of questions from customers resulted in lower satisfaction rates. In addition, the bank’s internal database of product details contains a multitude of file formats from text files to PDFs and tables, resulting in even slower processes.
Clerks can find answers to their product-related queries via a chat-based agentic RAG solution, with references to original documents through Advanced Document Processing. For further reliability, the system asks further questions instead of hallucinating. The solution integrates seamlessly with the internal authorization systems, handling different levels of access and making sure that users only have access to data they have the right to see. In addition, Azure AI Search allows changes to the original knowledge base to be reflected instantly.
AI
Digital Products
Banking
Docker
Angular
FastAPI
LangGraph
OpenAI
MLFlow
Python
Azure
Docker
Angular
FastAPI
LangGraph
OpenAI
MLFlow
Python
Azure
DECREASE IN CORE PROCESSING TIME
FASTER R&D AND TIME TO MARKET
INCREASE IN CUSTOMER PORTAL ORDERS
COST SAVINGS VIA CLOUD MIGRATION
REDUCTION IN PRODUCTION DOWNTIME
MONTHS FOR FULL PLATFORM INTEGRATION
Ready for takeoff?