Chat with Your Documents (AI Knowledge Base)
Upload documents and get cited, grounded answers — retrieval-augmented AI across the Reports Studio, Qualitative Studio, Quick Start, and the dashboard assistant.
4 min readOpen in app
Upload your own documents (PDF, DOCX, TXT, CSV) and ask questions of them in plain language. Nexmetry retrieves the most relevant passages and generates an answer that is grounded in — and cited back to — your sources, not the model’s general knowledge.
Privacy-first by design
Documents are split and embedded locally; only the short, relevant excerpts needed to answer your question are sent to the AI for generation. Sensitive- and PHI-classified content is blocked from that step by default.
Where you can use it
- Reports → Studio (Insight Notebook) — chat with uploaded sources while you write; this is also where org-scoped Evidence Synthesis lives.
- Qualitative Studio → Semantic Search — query transcripts by meaning, with "Find passages" (verbatim excerpts) and "Ask" (cited synthesis) modes.
- Quick Start → Documents — upload supporting files and ask questions with citations.
- Dashboard AI Assistant — answers also draw on your organisation knowledge base.
How it works
- 1UploadAdd documents to the surface you are working in (PDF, DOCX, TXT, MD, CSV). They are indexed for retrieval and marked accordingly.
- 2AskType your question in plain language. Nexmetry finds the most relevant passages across your documents.
- 3Get a cited answerThe AI answers grounded in those passages, with citations to your sources — and tells you when something is not in your documents rather than inventing it.
Organisation knowledge base
A platform administrator can curate a shared organisation knowledge base (managed in Admin → Knowledge Base) so every member’s dashboard AI Assistant can answer from your company’s documents.