Skip to content

Platform · Documents

Paper read and filed, letters drafted for review

Every document in the right record.

The documents desk and the doctors get their time back. Scanned paper is read and filed to the patient record, and AI drafts documents from visit data for the doctor to review and sign off.

From the scanner to the signed letter

  • Drafts from the visit

    AI drafts documents from visit data; the doctor reads the draft and changes what is needed.

  • Section by section

    Each section of a report is drafted from its own sources, following the clinic's templates.

  • Incoming mail

    The platform sorts incoming correspondence by kind and sends it to the desk that handles it.

  • PDF and HTML

    Documents come as PDF for the letter that is printed or signed, and HTML for reading on screen.

Documents on screen

  1. The document list of a synthetic patient in a demo clinic: folders, the files and where each came from.The document list of a synthetic patient in a demo clinic: folders, the files and where each came from.

    On the screen

    1. Whole screen
    2. Zoom in
    3. Move across
    Real screenThe document list of a synthetic patient in a demo clinic: folders, the files and where each came from.

    Paper, read and filed

    The documents desk scans incoming paper, and the platform reads it and files it to the patient record.

  1. Long documents

    The platform reads long documents for the team, such as a discharge summary or a specialist's report.

From the tray to the signed letter

Document intake: from a scanned letter to the patient record, with a human check A letter arrives by scan, upload or email; every page is read by text recognition; Condor drafts the findings for review; the persons are matched against the clinic; the document and the confirmed findings are filed to the patient record. Only one strong match files automatically; an ambiguous or missing match goes to staff review, and staff confirm before filing. persons + findings 1 strong match · auto ambiguous / none staff confirms Staff review ambiguous or no match Letter arrives scan · upload · email 1 Every page read text recognition (OCR) 2 Condor Findings drafted AI extraction · for review 3 Patient matched scored against the clinic 4 Filed to record document + confirmed findings 5 lands in the record human checkpoint arrives automatic step needs review
Document intake: from a scanned letter to the patient record, with a human check A letter arrives by scan, upload or email; every page is read by text recognition; Condor drafts the findings for review; the persons are matched against the clinic; the document and the confirmed findings are filed to the patient record. Only one strong match files automatically; an ambiguous or missing match goes to staff review, and staff confirm before filing. persons + findings 1 strong match · auto ambiguous / none staff confirms Staff review ambiguous or no match Letter arrives scan · upload · email 1 Every page read text recognition (OCR) 2 Condor Findings drafted AI extraction · for review 3 Patient matched scored against the clinic 4 Filed to record document + confirmed findings 5 lands in the record human checkpoint arrives automatic step needs review
Document intake: from a scanned letter to the patient record, with a human check A letter arrives by scan, upload or email; every page is read by text recognition; Condor drafts the findings for review; the persons are matched against the clinic; the document and the confirmed findings are filed to the patient record. Only one strong match files automatically; an ambiguous or missing match goes to staff review, and staff confirm before filing. page images text + layout persons + findings ambiguous / none 1 strong match · auto staff confirms Letter arrives scan · upload · email 1 Every page read text recognition (OCR) 2 Condor Findings drafted AI extraction · for review 3 Patient matched scored against the clinic 4 Staff review ambiguous or no match Filed to record document + confirmed findings 5 lands in the record human checkpoint arrives automatic step needs review
Document intake: from a scanned letter to the patient record, with a human check A letter arrives by scan, upload or email; every page is read by text recognition; Condor drafts the findings for review; the persons are matched against the clinic; the document and the confirmed findings are filed to the patient record. Only one strong match files automatically; an ambiguous or missing match goes to staff review, and staff confirm before filing. page images text + layout persons + findings ambiguous / none 1 strong match · auto staff confirms Letter arrives scan · upload · email 1 Every page read text recognition (OCR) 2 Condor Findings drafted AI extraction · for review 3 Patient matched scored against the clinic 4 Staff review ambiguous or no match Filed to record document + confirmed findings 5 lands in the record human checkpoint arrives automatic step needs review

Document intake in five steps: a letter arrives, every page is read, findings are drafted, the patient is matched, and the document is filed to the record. An unclear match goes to staff for review.

Open the full diagram (opens in a new tab)
Letter generation: every section written in parallel, checked, and approved by the doctor The doctor picks a letter template for a visit or patient, or asks Condor; the patient data is gathered once per letter; every template section is written at the same time by one section workflow (patient data, tools, write, check data, layout, language and style, save), with repair of failed checks up to three passes. The doctor reviews, edits or regenerates a section with an instruction; a late lab result reopens the review. On approval the branded PDF is filed and a follow-up to-do, “Send report to referrer”, is created for pack referral letters; the doctor sends the letter by secure email. draft ready approve on approval doctor sends Sent to the referrer doctor clicks send secure email · audited Follow-up to-do “Send report to referrer” pack rule · referral letters or ask Condor Doctor picks a letter template · visit or patient 1 Patient data gathered once per letter · current values 2 Sections written all in parallel · then checked 3 Doctor reviews edits · regenerates a section a late lab result reopens it 4 Letter approved branded PDF · in the record 5 regenerate a section + instruction 3 ONE SECTION WORKFLOW · RUNS ONCE PER SECTION, ALL AT THE SAME TIME patient data tools write check data check layout check language check style save A failed check repairs only the faulty span and checks again (up to 3 passes). A section that still fails is marked; it never blocks the rest of the letter. target 60 s per letter · hard stop 120 s · older templates still on the previous engine the approved letter doctor decides starts here automatic step doctor action
Letter generation: every section written in parallel, checked, and approved by the doctor The doctor picks a letter template for a visit or patient, or asks Condor; the patient data is gathered once per letter; every template section is written at the same time by one section workflow (patient data, tools, write, check data, layout, language and style, save), with repair of failed checks up to three passes. The doctor reviews, edits or regenerates a section with an instruction; a late lab result reopens the review. On approval the branded PDF is filed and a follow-up to-do, “Send report to referrer”, is created for pack referral letters; the doctor sends the letter by secure email. draft ready approve on approval doctor sends Sent to the referrer doctor clicks send secure email · audited Follow-up to-do “Send report to referrer” pack rule · referral letters or ask Condor Doctor picks a letter template · visit or patient 1 Patient data gathered once per letter · current values 2 Sections written all in parallel · then checked 3 Doctor reviews edits · regenerates a section a late lab result reopens it 4 Letter approved branded PDF · in the record 5 regenerate a section + instruction 3 ONE SECTION WORKFLOW · RUNS ONCE PER SECTION, ALL AT THE SAME TIME patient data tools write check data check layout check language check style save A failed check repairs only the faulty span and checks again (up to 3 passes). A section that still fails is marked; it never blocks the rest of the letter. target 60 s per letter · hard stop 120 s · older templates still on the previous engine the approved letter doctor decides starts here automatic step doctor action
Letter generation: every section written in parallel, checked, and approved by the doctor The doctor picks a letter template for a visit or patient, or asks Condor; the patient data is gathered once per letter; every template section is written at the same time by one section workflow (patient data, tools, write, check data, layout, language and style, save), with repair of failed checks up to three passes. The doctor reviews, edits or regenerates a section with an instruction; a late lab result reopens the review. On approval the branded PDF is filed and a follow-up to-do, “Send report to referrer”, is created for pack referral letters; the doctor sends the letter by secure email. template + anchor one data snapshot draft ready approve on approval regenerate a section + instruction doctor sends or ask Condor Doctor picks a letter template · visit or patient 1 Patient data gathered once per letter · current values 2 Sections written all in parallel · then checked 3 Doctor reviews edits · regenerates a section a late lab result reopens it 4 Sent to the referrer doctor clicks send secure email · audited Letter approved branded PDF · in the record 5 Follow-up to-do “Send report to referrer” pack rule · referral letters 3 ONE SECTION WORKFLOW · RUNS ONCE PER SECTION, ALL AT THE SAME TIME patient data tools write check data check layout check language check style save A failed check repairs only the faulty span and checks again (up to 3 passes). A section that still fails is marked; it never blocks the rest of the letter. target 60 s per letter · hard stop 120 s · older templates still on the previous engine the approved letter doctor decides starts here automatic step doctor action
Letter generation: every section written in parallel, checked, and approved by the doctor The doctor picks a letter template for a visit or patient, or asks Condor; the patient data is gathered once per letter; every template section is written at the same time by one section workflow (patient data, tools, write, check data, layout, language and style, save), with repair of failed checks up to three passes. The doctor reviews, edits or regenerates a section with an instruction; a late lab result reopens the review. On approval the branded PDF is filed and a follow-up to-do, “Send report to referrer”, is created for pack referral letters; the doctor sends the letter by secure email. template + anchor one data snapshot draft ready approve on approval regenerate a section + instruction doctor sends or ask Condor Doctor picks a letter template · visit or patient 1 Patient data gathered once per letter · current values 2 Sections written all in parallel · then checked 3 Doctor reviews edits · regenerates a section a late lab result reopens it 4 Sent to the referrer doctor clicks send secure email · audited Letter approved branded PDF · in the record 5 Follow-up to-do “Send report to referrer” pack rule · referral letters 3 ONE SECTION WORKFLOW · RUNS ONCE PER SECTION, ALL AT THE SAME TIME patient data tools write check data check layout check language check style save A failed check repairs only the faulty span and checks again (up to 3 passes). A section that still fails is marked; it never blocks the rest of the letter. target 60 s per letter · hard stop 120 s · older templates still on the previous engine the approved letter doctor decides starts here automatic step doctor action

A letter or report in five steps: the doctor picks a letter, patient data is gathered, the sections are written in parallel and checked, the doctor reviews, and the letter is approved.

Open the full diagram (opens in a new tab)

Questions about documents

See a letter go from tray to signature

Book a demo and we walk through a scanned referral and a drafted letter, from arrival to sign-off.