The fax machine’s data, finally usable
Clinical information still arrives on paper: faxed lab reports, scanned records from other facilities, uploaded PDFs from patients and specialists. In most EMRs these become images in a folder — technically stored, clinically invisible. JamEMR’s document intelligence reads them. Scanned, faxed, and uploaded documents are processed by JamEMR’s clinical knowledge ingestion engine, which extracts the clinically relevant content and proposes it as structured chart data.
A faxed lab report stops being a picture of results and becomes results — dated, structured, and placed on the patient’s Clinical Timeline.
Document intelligence is in active development and pilot validation with practices using JamEMR today.
How it works
- Ingest. A document enters JamEMR by scan, fax, or upload and is attached to the patient’s record.
- Read. The ingestion engine extracts the text and identifies what the document is — a lab report, a discharge summary, a referral letter.
- Structure. Clinically relevant data is extracted into structured form: results with values and dates, diagnoses, medications, key findings.
- Review. The proposed data is presented to a clinician for review before it enters the working record. Every extracted item traces back to its source document, so verification is a glance, not an act of faith.
Why this is the quiet centerpiece
Most of what an EMR “knows” about a patient’s care elsewhere arrives as documents. If those documents stay unread, the chart is systematically incomplete — and every AI feature built on the chart inherits that incompleteness. Document intelligence is how JamEMR closes the gap: it feeds the Clinical Knowledge layer, populates the Clinical Timeline, and makes inbound records searchable and answerable by the AI Assistant.
Processed locally, reviewed by humans
Two commitments govern this capability:
- Local processing. The AI models that read and structure documents run on the practice’s own dedicated local GPU hardware. Protected health information is not sent to third-party consumer AI clouds for extraction.
- Clinician in the loop. Extraction produces proposals, not facts. Nothing becomes part of the working record without clinician review, and every acceptance is recorded in the audit log.
The result is a chart that keeps up with the paper — without the practice hiring someone to retype it, and without the data leaving the building.