Document Intelligence
Every drawing, specification and transmittal, indexed and searchable.
Ten thousand files across a shared drive, an email chain and three transmittal portals. Nobody is certain which revision is current, and the scanned ones cannot be searched at all.
Upload the project. Citebuilt extracts the text, reads the scanned pages with OCR, understands the drawings, and indexes the lot so it can be retrieved by meaning rather than by filename.
Scanned drawings and photographed markups are handled the same as native PDFs. A drawing that exists only as an image stops being a dead end and becomes part of the searchable record.
The document register tracks revisions, so a query returns what is current rather than whatever happened to be indexed first. Superseded revisions stay available — you can ask what changed — but they do not masquerade as live information.
This is the layer everything else sits on. The copilot, compliance checking and standards retrieval all read from the same index, so there is one version of the project rather than one per feature.
What’s included
6 capabilities in Document Intelligence.
RAG library
Browse the embedded documents powering retrieval.
Embedded document library
Every ingested document, chunked and embedded for search.
OCR & drawing intelligence
Scanned PDFs and drawings read with OCR and discipline detection.
Document ingestion & vision
Visual understanding fed straight into review.
Document register
Every drawing and document, with revisions and review state.
Document library
What has been embedded and is searchable by the AI.
Who it’s for
- Document controllers
- Design managers
- Project engineers