Enterprise Knowledge Search
TE Connectivity needed to help product-development teams find and reuse engineering knowledge trapped in dozens of incompatible systems without spending hours searching documents or depending on subject matter experts.
Reported impact
TE Connectivity gave engineers a secure way to search and summarize proprietary knowledge much faster, reduced time spent navigating siloed systems, and created an enterprise platform that can scale as more content is ingested.
TE Connectivity, Capgemini, and AWS built TELme as a generative-AI research environment that turns scattered engineering knowledge into a searchable, conversational workspace. The program ingested 2.5 million documents in just over three months and was designed to scale toward 75 million documents held across 66 databases. AWS supplied the cloud infrastructure plus managed services including Amazon Bedrock and Amazon OpenSearch Service, while Capgemini used a retrieval-augmented generation architecture to combine those services with Claude 3.5 and TE's proprietary data. The result is a secure interface where engineers can search, summarize, and retrieve technical information without trawling through siloed systems or relying on subject-matter experts to answer basic questions. TELme is organized under a single API, continues to ingest more content, and is refined daily through joint feedback and governance so it can serve product-development teams at enterprise scale.
Engineering and operational documents are ingested from dozens of databases, indexed for search and retrieval, combined with proprietary context through a RAG architecture, and surfaced through TELme so engineers can ask questions and receive summarized answers in a secure UI.
"2.5 million engineering documents"
"AWS supplied the cloud infrastructure and fully managed services, such as Amazon Bedrock ... and Amazon OpenSearch Service"
"Capgemini used retrieval-augmented generation (RAG) ... to integrate these services into an enterprise-scale solution."
"Product development teams needed to sift through countless documents scattered across dozens of incompatible systems to conduct background research."
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Capgemini Sample Jul 21, 2025 Navantia Navantia wanted to modernize knowledge work across the organization by adopting generative AI in a controlled way that improved productivity, collaboration, and day-to-day information handling. 91.7% of allocated licenses activated Microsoft 365 Copilot · SAP · SharePoint 
Microsoft Sample Sep 24, 2026 Rockwell Automation Rockwell needed to reduce the time and inconsistency involved in diagnosing machine faults, preserve retiring workers' tacit shop-floor knowledge, and shorten training time for new technicians supporting hundreds of… costs are down by about 25 percent Azure AI Search · Azure OpenAI Service · Azure security, identity and governance services 
Atos Sample Sep 9, 2026 Flender China Flender China needed to interpret engineering standards, drawings, spreadsheets, and other technical documentation accurately in a highly regulated environment, but knowledge was fragmented across formats and… 30-50% faster knowledge discovery and document retrieval 