- Overview
- Medical Record Summarization
Healthcare Solutions user guide
About Medical Record Summarization
The solution's UI is only available in English.
Medical Record Summarization (MRS) is a UiPath vertical solution that uses agentic AI to automatically generate structured clinical summaries from patient medical records. It is built on the UiPath platform and powered by DeepRAG - UiPath's proprietary agentic search and synthesis technology. MRS enables healthcare organizations to reduce the time clinicians spend reading and synthesizing patient documents before making care decisions.
For more information about DeepRAG, refer to DeepRAG in the Agents user guide.
The application targets high-volume clinical workflows where manual document review creates bottlenecks: prior authorization, utilization management, discharge planning, appeals, and care coordination. It is delivered as a UiPath App hosted on UiPath Automation Cloud and uses UiPath Orchestrator for back-end processing.
How it works
MRS is built for high-volume, asynchronous summarization. Medical records or patient packets are deposited to a connected storage location, where they are automatically picked up and queued for processing. Each queued record is summarized in turn: the configured template is applied, and DeepRAG generates a structured clinical summary with inline citations.
Once generated, each summary enters the Clinical Review queue, where a human reviewer can approve it, reject it, or edit it before approval, or submit feedback to be taken into consideration for future summaries. Every claim in the AI-generated output is linked to a specific passage in the source document, so reviewers can verify the AI's interpretation against the original record without leaving the application.
DeepRAG
DeepRAG is the AI engine that powers summary generation. It autonomously navigates large, unstructured medical documents to find the relevant passages for each configured section. It processes multiple content types found in medical records - including text, images, scanned documents, tables, charts, and handwriting. For each section, it synthesizes a narrative and attaches inline citation markers that link back to the exact source text.
DeepRAG's output is fully controlled by the template configuration. Different templates produce different summary structures, section depths, and levels of clinical detail, making the AI output adaptable to the specific workflow and reviewer type. For more information about DeepRAG, refer to DeepRAG in the Agents user guide.
Key capabilities
| Capability | Description |
|---|---|
| AI-generated clinical summaries | Structured narratives generated from source documents, with inline citations |
| multimodal document processing | Processes text, images, scanned documents, tables, charts, and handwriting within source medical records |
| Configurable summary templates | Section-level control over content, output type, and reviewer persona |
| Human-in-the-loop review | Approve, Reject, or Override actions in the Clinical Review workflow |
| Version-controlled templates | Every edit creates a new version; full audit history is preserved |
| Deployment scheduling on the Deployments tab | Automated runs via UiPath Orchestrator triggers on a configurable schedule |
| Manual batch processing | Upload PDF files directly for on-demand summarization. Recommended during user acceptance testing to collect real reviewer feedback before deploying. |
| Usage analytics | Summaries generated, pages summarized, estimated time saved, and review time |
Use cases
| Use Case | Description |
|---|---|
| Prior authorization | Generate structured summaries to support insurance approval requests and reduce manual preparation |
| Utilization management | Summarize patient history and current admission details for UM nurses and physicians |
| Denial management | Synthesize clinical evidence to support denial appeals |
| Discharge planning | Generate discharge summaries and care transition documents |
| Care coordination | Provide cross-team summaries to support coordinated care across providers |
| Referrals | Summarize patient history for referring and receiving providers |
Application sections
MRS has three sections accessible from the left navigation sidebar.
| Section | Purpose |
|---|---|
| Templates | Create and manage clinical summary templates; view aggregate usage metrics |
| Clinical review | Review AI-generated summaries and take approval actions |
| Admin | Configure deployments, customize application appearance, and manage experimental features |