From a paper box to an instant answer
CortexEleven runs one continuous pipeline: it ingests your documents, digitises them with OCR, indexes what is inside, and puts a private AI brain on top. Here is exactly what happens at each stage, and what keeps it all under your control.
Five steps from a paper box to an instant answer
Point CortexEleven at a filing cabinet or a folder of scans. It digitises every page with OCR, classifies and indexes the content, builds a knowledge graph of your organisation, and answers questions across all of it with citations.
Five stages, one journey
Every document that enters CortexEleven travels the same path. Nothing is a black box.
Bring in everything, from anywhere
CortexEleven connects to where your documents already live and pulls them in without disruption. There is nothing to re-file and no archive too old to include.
- Network drives, SharePoint, OneDrive, S3 and SMB shares
- Direct scanner feeds and bulk upload of paper archives
- Snap documents with your phone camera in the CortexEleven app, straight into the pipeline
- Every format: PDFs, images, scans, office files and notes
- Runs in the background while your team keeps working
Turn images and paper into machine-readable text
Every page passes through OCR that extracts the text and records a confidence score. Pages the primary model reads poorly are automatically re-run on a more capable fallback, so nothing is quietly misread.
- Per-page confidence scoring on all extracted text
- Automatic fallback for low-confidence or difficult pages
- Handwriting, tables and stamps handled, not skipped
- Low-confidence pages flagged for optional human review
Map the people, matters and facts inside
Extracted text is parsed for the entities that matter to your work and woven into a knowledge graph, so related records link to one another. The archive begins to organise itself as it grows.
- Entities pulled out: people, matters, accounts, dates, obligations
- Documents classified, de-duplicated and categorised automatically
- A knowledge graph connects related records across departments
- Specialised agents keep the index clean and current
A private model that knows your corpus
On top of the index sits a private model: it understands your organisation’s own language and history, running entirely within the boundary you choose.
- Grounded in your documents, not the open internet
- Runs on your hardware or your private tenant
- Understands your sector’s terminology and your house style
- Improves as more of your archive is digitised
Ask anything, in plain language
Staff ask questions the way they would ask a colleague and get answers grounded in the actual records, every claim traceable back to its source page.
- Natural-language and keyword search across everything
- Answers with inline citations to the source document and page
- Analytics on what the organisation asks and where knowledge grows
- A simple search portal for staff, the full platform for admins
The whole pipeline runs inside your boundary
Ingestion, OCR, indexing and the model all run where you decide, cloud tenant, on-premise appliance, or fully self-hosted. Your documents are never sent to a third-party model, and every sensitive action is captured in an audit trail you can inspect.
Drives, cloud stores, scanners and phone photos flow into one pipeline.
Every response is traceable to the source document and page.
Data and models stay inside the boundary you choose, always auditable.
Turn your archive into a searchable knowledge hub
See CortexEleven digitise and search a sample of your own records - on your hardware, in your environment.