# Kognitos + Reveal HealthTech: Automating Medical Document Processing

> Kognitos and Reveal HealthTech demo a medical-intake automation: extract patient data, validate identifiers across documents, resolve a DOB discrepancy, and write to SharePoint.

**Page**: https://www.kognitos.com/videos/kognitos-and-reveal-healthtech-processing-varied-medical-documents/
**Watch on YouTube**: https://www.youtube.com/watch?v=aZThnvOA3tQ
**Length**: 3m 30s

## What's in this video

A joint demo with Reveal HealthTech showing how Kognitos automates the front end of medical intake: extracting patient data from varied forms, validating it across documents, resolving exceptions with a clinician in the loop, and writing the records to a structured destination (SharePoint and an Excel report).

## The intake pipeline

- **Email trigger**: new patient and records-request documents arrive by email; Kognitos can also pull from other sources.
- **Extraction**: for each document, Kognitos extracts the patient identifiers — name, MRN, date of birth — plus age, related doctors, and any relevant notes.
- **Cross-document validation**: the same patient identifiers are checked across the document set to catch inconsistencies. In the demo, patient *Charles*'s date of birth is inconsistent across forms.
- **Consolidation**: validated data is rolled up into an Excel *new patient intake* report attached to a notification email.
- **SharePoint filing**: each patient's documents are placed into a unique SharePoint folder. (Other file-storage systems work too.)

## Human-in-the-loop exception handling

When the date-of-birth mismatch is detected, the automation pauses and emails a designated team. The email links to an interface where a clinician can review the documents, type the correct DOB into a dialog, and — importantly — teach Kognitos how to handle similar mismatches in the future. Once the DOB is supplied, the automation resumes from where it paused.

## Why this approach fits healthcare ops

Hospital intake teams deal with handwritten forms, faxed referrals, and mixed PDF layouts that break template-based IDP. Kognitos extracts from those varied formats and — crucially — keeps a clinician in the loop on identity discrepancies, which keeps the automation safe in a clinical context.

## FAQs

**Q: What kinds of medical documents does this automation process?**

New patient intake forms and records-request documents (e.g., request for information release). Documents arrive by email in the demo, but Kognitos can pull from any supported source.


**Q: Which patient identifiers does Kognitos validate?**

Name, medical record number (MRN), and date of birth — checked across every document in the intake to flag inconsistencies before downstream entry.


**Q: What happens when there's a date-of-birth mismatch?**

The automation pauses and emails a designated team a link to a human-in-the-loop interface. A clinician reviews the documents, types the correct date of birth, and can teach Kognitos how to handle similar mismatches going forward. The automation then resumes.


**Q: Where does Kognitos store the processed documents?**

In the demo, each patient gets a unique SharePoint folder. Kognitos works with other file-storage systems too — SharePoint is shown as one example.