Industry / AI
NextGroup, years of documents you can query in plain English
7 TB of documents on a NAS made searchable in natural language, with the exact source cited. Smart Archive was born here.
The challenge
NextGroup kept seven terabytes of documents on a company NAS: quotes, drawings, contracts, archived email. The answer to any question already existed, but finding it cost more than redoing the work. No internal search tool coped with that volume, and the data could not leave the company.
The solution
We made the whole archive searchable with a RAG system that respects existing permissions and answers in natural language, citing the exact source the answer comes from. The model queries the data through MCP while staying inside the company perimeter. The project became the foundation of Smart Archive, now available as a product.
The result
of documents searchable with one question, source cited, data always inside the company.
The product born from this project
Smart Archive
Next case
