Case Study: a leading European energy company boosts search efficiency and accuracy with Eviden and AWS GenAI

GenAl lights up search efficiency and accuracy in Confluence and GitLab repositories

A leading European energy company faced significant challenges in efficiently retrieving accurate information from its vast Confluence documentation and GitLab code repositories. The time-consuming and often inaccurate search process was draining productivity. The company partnered with Eviden, an AWS Premier Tier Services Partner, to implement a generative AI solution to address this internal knowledge management problem.

Eviden designed and implemented a retrieval-augmented generation (RAG) solution built on AWS. The system uses AWS Glue to ingest data from Confluence and GitLab, Amazon OpenSearch Serverless for vector search, and Amazon Bedrock's large language models for analysis and a chatbot interface. This advanced AI solution, which included techniques like reranking and summarization, significantly improved the efficiency and accuracy of information retrieval for the client, accelerating search times and boosting overall productivity.