Control 2.9: Defender for Cloud Apps — Copilot Session Controls — Verification & Testing
Test cases and evidence collection for validating session controls.
Test Cases
Test 1: Session Monitoring Activation
- Objective: Confirm Copilot sessions are monitored by Defender for Cloud Apps
- Steps:
- As a test user, perform several Copilot interactions in Word and Teams
- Navigate to Defender > Investigate > Activity Log
- Filter for the test user and verify Copilot activities appear
- Confirm activity details include interaction context
- Expected Result: Copilot activities appear in the Defender activity log
- Evidence: Activity log entries for test user
Test 2: Content Inspection Detection
- Objective: Verify content inspection detects sensitive data in sessions
- Steps:
- Use Copilot to interact with a document containing test sensitive data
- Verify the session policy content inspection triggers
- Confirm the alert is generated with the correct severity
- Verify the detection details include the sensitive information type
- Expected Result: Content inspection detects and alerts on sensitive data
- Evidence: Alert record with detection details
Test 3: Alert Generation and Delivery
- Objective: Confirm alerts are generated and delivered to the security team
- Steps:
- Trigger a session policy condition (e.g., sensitive data detection)
- Verify an alert appears in Defender portal > Alerts
- Confirm email notification is delivered to configured recipients
- Verify alert severity matches the policy configuration
- Expected Result: Alerts generated and delivered within expected timeframe
- Evidence: Alert notification and email confirmation
Test 4: Generative AI App Catalog Coverage
- Objective: Verify the organization has reviewed and governed generative AI app usage via the MDCA catalog
- Steps:
- Navigate to Defender portal > Cloud Apps > Cloud app catalog > filter by Generative AI category
- Confirm the catalog loads and displays generative AI apps
- Check the "Discovered apps" view to identify any generative AI apps used in the organization that are not Microsoft 365 Copilot
- Verify that high-risk discovered generative AI apps have governance policies applied (blocked, unsanctioned, or explicitly approved)
- Confirm Microsoft 365 Copilot is marked as "Sanctioned" in the catalog
- Expected Result: Generative AI catalog has been reviewed; high-risk apps are governed; sanctioned apps are documented
- Evidence: App catalog screenshot with discovered apps; governance policy configuration for high-risk apps
Test 5: Agent Threat Detection Verification
- Objective: Confirm agent threat detection is operational for Copilot agent deployments
- Steps:
- Navigate to Defender portal > Incidents & alerts and filter for agent-related alerts
- Confirm that agent monitoring is active for organizational Copilot agent deployments
- Review any existing agent-related incidents to verify the detection is surfacing actionable intelligence
- Verify at least one custom agent anomaly detection rule is configured
- Confirm agent threat alert routing: alerts should reach the SOC or security team within the defined SLA
- Expected Result: Agent threat detection is active, alerts are routing correctly, and custom rules are configured
- Evidence: Agent alert configuration; sample agent incident records; alert routing confirmation
Evidence Collection
| Evidence Item |
Format |
Storage Location |
Retention |
| Session policy configuration |
Screenshot/PDF |
Compliance evidence repository |
7 years |
| Activity log samples |
CSV |
Compliance evidence repository |
7 years |
| Alert records |
CSV |
Compliance evidence repository |
7 years |
| Content inspection test results |
PDF |
Compliance evidence repository |
7 years |
Compliance Mapping
| Regulation |
Requirement |
How This Control Supports It |
| FINRA Rule 3110 |
Supervisory system monitoring |
Session controls support compliance with AI interaction monitoring requirements |
| SEC Rule 17a-4 |
Electronic communication monitoring |
Session logging helps meet communication monitoring obligations |
| FFIEC Handbook |
Security monitoring |
Real-time session controls support compliance with security monitoring requirements |
| NIST CSF |
DE.CM-1 Network monitoring |
Session controls provide monitoring for AI workloads |