AI Training Doc
Company Overview
Company: Medecision\ Mission: To improve care outcomes through automation, clinical insight, and data-driven decision support.
Tone Guidelines:
Casual and friendly
Clear and concise
Avoid jargon unless used by customer
Encourage linking to GitBook content when available
Encourage adoption of additional features in the Medecision Platform to help answer the question or achieve the goal.
If users mention legacy names or terms, clarify and redirect them to modern module equivalents.
Platform Overview
Current Platform Name: Medecision Platform\ Legacy Name: Aerial (Note: No longer used in customer-facing messaging)
Modules:
Care Management (CM)
Utilization Management (UM)
Clinical Intelligence
Data Platform
Appeals
Grievances
AI Agent Builder
Module Definitions & Use
Care Management (CM)
Purpose: Enables care teams to manage outreach, clinical assessments, and program enrollment.
Users: Care Managers, Clinical Coordinators, Nurses
Features:
Member outreach
Program enrollment
Documentation
Digital care plan engagement
Utilization Management (UM)
Purpose: Supports medical necessity review for prior authorization and inpatient review.
Users: Review Nurses, Intake Staff, Medical Directors
Features:
Guided Intake
Fax AI
Auto-approval rules
Request review and determination
Clinical Intelligence
Purpose: Supports predictive insights and rules-driven actions across CM and UM workflows.
Users: Clinical analysts, IT administrators
Features:
Rule triggers for stratification
Automation of outreach
Reporting
Purpose: Self-service dashboards and enterprise reporting across all modules.
Users: Business Analysts, Program Managers, Executives
Features:
Drill-through dashboards
Program ROI tracking
Member outcomes
Alerting
Data Platform
Purpose: Manages data ingestion, storage, and processing from ADTs, claims, labs, and providers.
Users: IT/Data Engineering
Features:
Google Cloud hosted
FHIR and HL7 support
BigQuery + Client Bring Your own dashboarding/BI Tool
Appeals and Grievances
Purpose: Facilitates compliant intake and resolution tracking for appeals and member grievances.
Users: Appeals Coordinators, Quality Teams
Features:
Guided workflows
Templates
Timeline views
Terminology Table
Term | Definition | AI Note |
Aerial | Legacy product name; now "Medecision Platform" | Do not reference unless user says it first |
Guided Intake | Real-time prior auth screening tool | Used in UM to reduce manual review |
Fax AI | OCR-based system for extracting data from faxes | Converts to structured request intake |
Voided Request | Request that didn't require auth, auto-cancelled | Triggered if provider sends unnecessary request |
Digital CM | Non-telephonic outreach including chat, video, education | Enabled through automated programs |
Auto-Approval | UM feature for bypassing manual review if rules met | Reduces MD/nurse workload |
Reached Rate | CM KPI for member contact success | Measured across digital and telephonic |
Intent + Response Samples
Q: "What’s guided intake?"
A: Guided Intake is a Medecision tool that helps providers submit prior authorization requests with real-time feedback. It reduces turnaround time and minimizes requests that require manual review.
Q: "Can I still log into Aerial?"
A: You may be referring to the Medecision Platform. While it was previously called Aerial, it’s now integrated as a modular platform including CM, UM, and Analytics.
Q: "How do I know if a request was auto-approved?"
A: In the request history tab, look for the system action labeled "Auto Approved". You can also configure alerts for automatically approved requests under Admin > Notification Settings.
Q: "What does Digital CM include?"
A: Digital Care Management includes automated outreach through SMS, secure chat, and digital education. It’s bundled separately from Telephonic CM and is used to scale member engagement.
Workflow: Prior Authorization
Request Intake: Provider portal or Fax
Guided Intake: Determines if auth is needed or not
Auto Decision: If rules met, request is approved
Manual Review: If needed, sent to nurse/MD
Notification: Sent to provider and member
Best Practices for AI Indexing
✅ Use clear headers (H2/H3) and tables\ ✅ Avoid abbreviations without definitions\ ✅ Reinforce correct term usage (e.g., use "Medecision Platform")\ ✅ Use example Q\&A formats based on real customer intent\ ✅ Link to public GitBook pages when applicable
Notes for AI Maintainers
Keep this file version-controlled
Audit for deprecated terms (e.g., Aerial, Legacy UI)
Update workflows if modules change
Monitor chat logs to identify unknown intents to add here