Home Case Studies Boosting RAG Retrieval Recall by 56.4% for Telecom Support Tickets

Boosting RAG Retrieval Recall by 56.4% for Telecom Support Tickets

One of the fastest-growing network AI company

Our enhancements achieved a 56.4% improvement in RAG retrieval recall, enabling the client to confidently showcase the system’s capabilities to the telecom company.

Meet our client

Client:

One of the fastest-growing network AI company

Industry:

Telecoms & Media

Market:

US

Technology:

LLM

Client’s Challenge

One of the fastest-growing network AI company had implemented a RAG system but struggled with the retrieval component. The data, consisting of technical and structured telecom support tickets, was in French, adding complexity to the system’s functionality and accuracy.

Our Solution

We improved the system’s retrieval capabilities by introducing advanced chunking, rephrasing techniques, and deploying better embedding models to handle the specific and technical nature of the data more effectively.

Client’s Benefits

Our enhancements achieved a 56.4% improvement in RAG retrieval recall, enabling the client to confidently showcase the system’s capabilities to the telecom company. The solution promises increased customer service efficiency through faster resolution times and improved support quality, enhancing overall customer satisfaction.

Posted in

See more projects