- Overview
- Medical Record Summarization
- About Medical Record Summarization
- A template's lifecycle
- Create a template
- Edit a template
- Deploy a template
- Send documents for processing
- Review a summary
- Medical Record Summarization (MRS) output data
- Manage deployments
- Customize application appearance
- Template sections reference
- Clinical review reference
- System components
Healthcare Solutions user guide
Medical Record Summarization supports two modes for getting source documents into the processing pipeline. Choose the mode that matches your deployment stage and integration setup.
| Mode | How documents enter MRS | When to use |
|---|---|---|
| Manual batch | A user uploads PDF files directly through the MRS application. | Initial testing, user acceptance testing, or ad-hoc processing. No external integration required. |
| Automated discovery | An external system deposits PDF files to a connected Orchestrator storage bucket. MRS Discovery scans for new files on a configured schedule. | Production deployments where documents are sourced from an EHR, claims platform, or document management system. |
Upload documents manually
Manual upload lets you submit PDF files directly from the MRS application without configuring an external integration. It is the recommended approach for running real reviewer feedback during user acceptance testing, and can also be used between preview and deployment to validate template output quality.
Prerequisites
- You have access to the Medical Record Summarization application.
- The template is in Published status.
Steps
-
Select Templates in the left navigation sidebar.
-
Select the template you want to process documents against.
-
Select Manual upload in the template detail view header.
-
In the upload dialog, select the files to upload. Only PDF files are accepted. Files are stored in the
MRSPacketsOrchestrator storage bucket and queued for processing. -
Select Upload.
Result: The application creates one processing record per file and starts the summarization pipeline. Completed summaries appear in the Clinical Review queue when processing finishes. The run is recorded in the template's Runs tab.
Set up automated discovery
Automated discovery runs on a schedule and picks up new documents without user intervention. When a deployment is active, the MRS Discovery process scans the configured storage bucket at the scheduled interval, identifies files newer than the last processed date, and starts the summarization pipeline automatically.
Prerequisites
- The template is deployed. See Deploy a template.
- Your external system (EHR, claims platform, or document management system) can write PDF files to an Orchestrator storage bucket accessible from your UiPath tenant.
Configure your external system
Your external system must write PDF files to the IncomingMRSPackets Orchestrator storage bucket at the root path configured in the deployment.
For automated discovery to work correctly, organize files using date-based subfolders (for example, YYYY-MM-DDThh-mm-ss). The MRS Discovery process uses the last processed folder date tracked in the IngestionState Data Fabric entity to determine which folders to scan on each run. Files placed outside a date-based folder structure may not be picked up on the next scheduled run.
Only PDF files are supported. Each PDF should represent one complete patient record or document packet. If your source system produces multiple separate documents per patient, consolidate them into a single PDF before depositing to the bucket.
Trigger a run immediately
To process documents currently in the bucket without waiting for the next scheduled run, select Run now from the deployment row in the Admin → Deployments page. For details, see Manage deployments.