Skip to content

Control 2.10: Insider Risk Detection for Copilot Usage Patterns

Control ID: 2.10 Pillar: Security & Protection Regulatory Reference: GLBA §501(b), Sarbanes-Oxley §§302/404, FINRA 3110 Last Verified: 2026-09-29 Governance Levels: Baseline / Recommended / Regulated


Objective

Configure Microsoft Purview Insider Risk Management (IRM) policies to detect anomalous Copilot usage patterns that may indicate data exfiltration, unauthorized information gathering, or misuse of AI capabilities. Copilot provides a powerful new vector for insider threats — enabling rapid aggregation and summarization of data that would previously require extensive manual effort. This control supports compliance with GLBA safeguard monitoring requirements, SOX internal control obligations, and FINRA supervisory expectations.


Why This Matters for FSI

  • GLBA §501(b) requires monitoring and testing of safeguard effectiveness — insider risk detection for Copilot demonstrates that the firm monitors for internal threats to customer information via AI tools
  • Sarbanes-Oxley §§302/404 requires assessment of internal controls — insider risk policies for Copilot are an internal control designed to detect unauthorized access to financially sensitive information
  • FINRA Rule 3110 (Supervision) requires supervisory systems reasonably designed to prevent and detect violations — detecting anomalous Copilot usage is a supervision measure for AI-assisted activities. Per the FINRA 2026 Annual Regulatory Oversight Report (GenAI Section, December 9, 2025), agentic AI systems require supervisory controls that cover AI workflow engines selecting intermediate actions. IRM policies for Copilot agents directly address this requirement.
  • FINRA Rule 3120 requires designated supervisory personnel to test supervisory procedures — insider risk detection provides testable, evidence-generating supervision of Copilot
  • SEC Enforcement precedent demonstrates that regulators hold firms accountable for insider misuse of information systems — Copilot's ability to aggregate MNPI makes insider risk monitoring essential
  • OCC/Fed Interagency Guidance expects financial institutions to manage AI-related risks, including the risk of misuse by authorized users

Control Description

Microsoft Purview Insider Risk Management uses signals from across Microsoft 365 to identify potentially risky user activity. When integrated with Copilot activity signals, IRM can detect patterns that suggest misuse of AI capabilities. IRM has added several capabilities — a number of them still in preview — that expand coverage for AI and agent governance. Organizations should verify each capability's current availability, preview status, and licensing in their tenant before relying on it.

Copilot Risk Scenarios for IRM Review

Risk Scenario Description Review Method Severity
Bulk data extraction via prompts User submits many prompts designed to extract and aggregate data Risky AI usage indicators plus audit, DLP, or device correlation as available High
Sensitive topic probing Repeated queries about specific restricted topics (MNPI, client data) Risky AI usage indicators, Communication Compliance, or approved audit review High
Cross-segment information gathering Copilot queries that attempt to access data across business segments Information barriers, DLP, and audit correlation Critical
Unusual query volume Copilot interaction volume appears materially different from the user's normal operating pattern Organization-defined audit or reporting review, not a built-in IRM threshold Medium
Off-hours Copilot activity Copilot usage outside normal working hours coincides with other risk factors Audit timestamp review correlated with other signals Medium
Download after Copilot summary User downloads files shortly after Copilot summarizes them Sequence review using audit, file, or exfiltration evidence High
Copilot + USB/print activity Copilot data extraction followed by USB or print activity Multi-signal correlation with device and exfiltration evidence Critical
Pre-resignation data gathering Increased Copilot usage after submitting resignation HR signal plus approved audit or investigation review High
Client data aggregation Copilot used to compile client lists, account data, or portfolio summaries Risky content review plus audit or DLP correlation High
AI usage anomaly Audit or reporting review identifies a meaningful change in Copilot or AI-app activity that warrants investigation Purview audit, DSPM reporting, or organization-defined analytics Medium
Agent interaction spike Audit or reporting review identifies unusual agent activity that merits correlation with Risky Agents alerts or other controls Purview audit, Agent 365 / Defender reporting, or organization-defined analytics High

IRM Policy Architecture for Copilot

Signal Sources                    IRM Processing                 Outcome
┌──────────────┐
│ Copilot      │──┐
│ Activity Logs│  │
└──────────────┘  │
┌──────────────┐  │
│ Agent        │──┤
│ Interaction  │  │    ┌──────────────────┐    ┌─────────────────┐
│ Logs         │  │    │ Insider Risk     │    │ Alert Generated │
└──────────────┘  │    │ Management       │──→ │                 │
┌──────────────┐  │    │                  │    │ ├─ Low severity  │
│ HR Signals   │──┼──→ │ ├─ Policy match  │    │ ├─ Medium        │
│ (departure,  │  │    │ ├─ Risk scoring  │    │ ├─ High          │
│  PIP, etc.)  │  │    │ └─ Correlation   │    │ └─ Critical      │
└──────────────┘  │    └──────────────────┘    └────────┬────────┘
┌──────────────┐  │                                     │
│ DLP Matches  │──┤                              ┌──────┴──────┐
│              │  │                              │ IRM Triage  │
└──────────────┘  │                              │ Agent +     │
┌──────────────┐  │                              │ Investigation│
│ Device       │──┤                              │ Workflow    │
│ Activity     │  │                              └─────────────┘
└──────────────┘  │
┌──────────────┐  │
│ Cloud App    │──┘
│ Activity     │
└──────────────┘

IRM Policy Templates for Copilot

Template Triggering Event Key Indicators FSI Use Case
Data theft by departing users HR termination/resignation signal Copilot data gathering + file downloads Departing advisors gathering client data
General data leaks No trigger required (continuous) Sensitive-data access plus companion exfiltration or device signals General insider threat monitoring
Security policy violations DLP policy match DLP match in Copilot + continued probing Repeated attempts to access restricted data via AI
Risky browser usage No trigger required Copilot usage + unauthorized cloud storage access Data exfiltration via AI + personal cloud
Patient data misuse (customizable) No trigger required Copilot access to sensitive labeled content Unauthorized access to client financial data
Risky Agents (preview) Agent activity (default policy present) Risky prompts, sensitive responses, access to sensitive data, risky websites, external SharePoint sharing Agent-level misuse of Copilot Studio, Microsoft Foundry, and P4AI SDK agents

Adaptive Protection Integration

Adaptive Protection dynamically adjusts DLP enforcement based on insider risk levels:

IRM Risk Level DLP Policy Adjustment Copilot Impact
Elevated Block without override Copilot blocks all sensitive content processing
Moderate Block with override Copilot blocks sensitive content with manager override
Minor Enhanced monitoring Copilot DLP policy tips always shown

Adaptive Protection also integrates with Conditional Access to dynamically restrict Copilot access for at-risk users. See Control 2.3 for CA integration details.


New Capabilities for Agent and AI Governance

Risky Agents Policy Template

The Risky Agents policy (in preview) provides out-of-the-box coverage for common agent risk scenarios. It is available by default to organizations with supported licenses — when Insider Risk Management is set up, this policy is automatically present and ready to generate alerts based on observed agent activity. Organizations should verify the current preview status, licensing, and tenant availability before relying on it.

Scope: Supports Copilot Studio agents, Microsoft Foundry agents, and agents built using the P4AI SDK. Microsoft prebuilt agents, third-party agents, and SharePoint agents are not listed among the supported agent types; coverage for additional agent types is expected to expand.

By default, the Risky Agents policy detects: - Risky prompts - Agents generating sensitive responses - Agents accessing sensitive data - Access to risky websites (the policy template also covers external SharePoint file sharing and activity above an agent's usual daily activity)

Why this matters for FSI: Per the FINRA 2026 Annual Regulatory Oversight Report (GenAI Section, December 9, 2025), agentic AI systems require supervisory controls that cover AI workflow engines selecting intermediate actions. FINRA Rules 3110 and 3120 — supervisory systems must be reasonably designed to prevent and detect violations, and this requirement now extends explicitly to AI-driven agent activities. The Risky Agents policy provides a foundational supervisory mechanism for Copilot Studio, Microsoft Foundry, and P4AI SDK agent deployment.

Tier recommendations:

Level Requirement
Baseline Review the default Risky Agents (preview) policy; confirm it covers deployed Copilot Studio, Microsoft Foundry, and P4AI SDK agents; configure alert routing to compliance team
Recommended If the default Risky Agents policy doesn't meet the firm's needs, create a supported custom policy or companion monitoring workflow for the additional FSI scenarios; add agent-specific alert routing to compliance and risk teams; document agent risk management as part of supervisory procedures
Regulated Mandatory compliance review of all agent risk alerts within 24 hours; quarterly agent risk assessment reporting to senior management; document Risky Agents oversight in firm supervisory procedures per FINRA Rules 3110/3120

Generative AI Apps and Risky AI Usage Indicators

IRM now separates the app-selection layer from the risky-behavior layer for AI monitoring:

  • Generative AI apps indicators let you decide which Microsoft Copilot experiences, enterprise AI apps, and other AI applications are monitored
  • Risky AI usage indicators (preview) score risky prompts and sensitive AI responses from those monitored apps
  • These indicators feed into existing IRM risk scoring alongside traditional indicators (file downloads, USB usage, HR signals)
  • For Microsoft 365 Copilot data, Microsoft states there are no pay-as-you-go billing requirements or charges for detecting inappropriate or risky interaction activity; non-Microsoft 365 AI data can require pay-as-you-go billing

Integration: Configure these settings at Insider Risk Management > Settings > Policy indicators. Expand Generative AI apps indicators to choose the apps to monitor, then enable the Risky AI usage indicators (preview) and any companion indicators appropriate for your deployment scale.

Organization-defined review heuristic Example evidence source Use
Increased Copilot interaction volume Purview Audit CopilotInteraction trend analysis Use only as a local triage heuristic; this isn't a built-in IRM indicator or Microsoft-documented threshold
Increased agent activity Agent 365 / Defender reporting plus Purview audit context Use as supplemental investigation context; pair with Risky Agents alerts rather than treating it as a built-in IRM threshold
Broader AI app adoption DSPM / Purview activity or audit reporting Use to prioritize monitoring coverage and review whether additional AI apps should be in scope
After-hours AI usage Purview audit timestamps correlated with other risk signals Use as contextual evidence in an investigation, not as a Microsoft-documented built-in indicator threshold

Data Risk Graphs

Data risk graphs provide a visual investigation experience within IRM — powered by Microsoft Sentinel integration — that combines asset and activity data into a single view, summarizing a user's alert-related exfiltration activity over time to reveal contextual connections. Setup requires onboarding a Microsoft Sentinel data lake (which uses pay-as-you-go billing).

Retirement notice: Microsoft now states that the data risk graph experience will be retired on November 24, 2026. Organizations that still depend on it should use it while available and document a replacement investigation workflow before that date.

Data risk graphs currently support SharePoint and OneDrive exfiltration activities, such as: - Anonymous or company sharing links created or used in SharePoint or OneDrive - File downloads from SharePoint and OneDrive - Files renamed in SharePoint or OneDrive - Clusters of users involved in related activity that individually appears low-risk but collectively may suggest coordinated behavior

FSI application: Financial institutions with complex organizational structures (trading, advisory, operations, compliance) can use data risk graphs to visualize the file assets, users, and exfiltration activities connected to an alert. A graph showing multiple operations staff downloading or sharing investment-banking documents may warrant investigation under Chinese wall supervisory procedures.

Access path: For tenants that onboarded before September 24, 2026, set up remained available via Microsoft Purview > Insider Risk Management > Recommended actions > Set up data lake and data risk graph. Once configured, open a graph from an alert: Insider Risk Management > Alerts (preview) > (select alert) > Data risk graph tab. Microsoft documents that the graph starts with the most recent seven days of data and grows toward the full 30-day window over time. Current Learn doesn't name a direct successor to the data risk graph experience, so organizations should plan a replacement investigation workflow before retirement.

IRM Triage Agent (Security Copilot)

The IRM Triage Agent, powered by Security Copilot, is in preview. It automates the initial triage phase of the IRM alert queue:

  • Categorizes alerts by severity based on the activity pattern and available context
  • Recommends investigation priorities, helping investigators focus on the highest-risk alerts first
  • Provides context summaries: what activity was detected, what data was involved, what the user's recent behavior pattern shows
  • Reduces alert fatigue in high-volume IRM environments by pre-processing the alert queue before human investigators review it

Microsoft now documents the Triage Agent as a Microsoft Purview agent deployment, not an Insider Risk Settings toggle. Deployment runs through Microsoft Purview > Agents > Explore agents, and analysts review triaged items from Insider Risk Management > Alerts (preview). The feature requires both Insider Risk Management licensing and Security Copilot capacity prerequisites: pay-as-you-go billing, provisioned security compute units (SCUs), Microsoft 365 data sharing in Security Copilot, and the Microsoft Purview plug-in.

Content Preview in IRM Alerts (preview)

IRM supports content preview (in preview), allowing investigators to review flagged content directly within Activity explorer to validate risk without opening a full case. Content preview is supported for activities that involve accessing or transmitting content, including relevant files from SharePoint, Exchange, and OneDrive for Business; it is not supported for state-change, permission-change, deletion, or signal-only events.

  • Workflow improvement: Investigators can assess alert severity and context faster by previewing the relevant flagged content in Activity explorer — reducing mean time to triage
  • Access controls: Content preview access is governed by the same role-based permissions that control IRM investigation access. Organizations should verify that preview permissions are restricted to authorized investigators and that preview activity is logged for auditability
  • FSI application: Faster triage is particularly valuable for firms with high Copilot adoption generating significant IRM alert volume, and for supervisory teams operating under FINRA Rule 3110 SLA expectations for alert review

Enhanced Risky AI Usage Indicators for Copilot (preview)

IRM's risky AI usage indicators (in preview) provide expanded coverage for Copilot-specific risk scenarios. Organizations should verify current availability and behavior in their tenant:

  • Copilot prompt risk scoring: IRM now evaluates the sensitivity of Copilot prompts in context with other user activity, providing more nuanced risk scoring that distinguishes between routine Copilot usage and potentially risky information-gathering patterns
  • Web-grounding risk indicators: When Copilot uses web search to supplement responses, IRM can flag instances where web-grounding occurs in the context of sensitive data interactions — supporting detection of data leakage risk via mixed internal/external content generation
  • Agent interaction risk refinement: Microsoft continues to evolve how preview risk indicators evaluate Copilot Studio agent activity. Organizations should verify current tenant behavior before relying on any specific agent-volume interpretation or tuning assumption.

FSI application: Large financial institutions with active Copilot deployments generate significant IRM alert volume. The Triage Agent reduces the manual overhead of initial alert processing while maintaining supervisory rigor — investigators review pre-categorized, contextualized alerts rather than raw signals.

Per OCC Bulletin 2025-26 on proportionate model risk management: The IRM Triage Agent represents AI-assisted governance that reduces manual overhead while maintaining appropriate oversight. The Triage Agent itself should be documented as a model per OCC Bulletin 2011-12 (SR 11-7) model risk management guidance — with appropriate validation, ongoing monitoring, and escalation procedures if the Triage Agent's recommendations are found to be systematically incorrect.

Tier recommendations:

Level Requirement
Baseline Deploy the Triage Agent in manual, one-alert-at-a-time mode; review categorizations weekly; document as a model in the firm's model inventory
Recommended Configure scheduled automatic runs after validation; Triage Agent recommendations guide investigator queue prioritization; quarterly review of categorization accuracy
Regulated Enable with human-in-the-loop validation before alert dismissal; Triage Agent model documented with performance monitoring; any systematic mis-categorization triggers model remediation process per SR 11-7

Copilot Surface Coverage

M365 Application IRM Signal Collection Anomaly Detection Adaptive Protection Notes
Microsoft Copilot Chat Yes Yes Yes Primary monitoring surface
Word Yes Yes Yes Document creation/editing signals
Excel Yes Yes Yes Data analysis and export signals
PowerPoint Yes Yes Yes Presentation creation signals
Outlook Yes Yes Yes Email drafting and sending signals
Teams Yes Yes Yes Chat and meeting activity signals
OneNote Yes Yes Yes Note creation signals
Loop Yes Yes Yes Collaboration signals
Copilot Pages Yes Yes Yes Page creation and sharing signals
Microsoft 365 agents (including declarative/SharePoint scenarios surfaced as Microsoft 365 agents) Partial Partial Partial Risky Agents doesn't explicitly list SharePoint agents; validate tenant behavior and use CopilotInteraction audit plus companion controls where direct Risky Agents coverage isn't present
Copilot Studio Agents Yes Yes Yes Covered by Risky Agents (preview) policy
Microsoft Foundry Agents Yes Yes Yes Covered by Risky Agents (preview) policy
P4AI SDK Agents Yes Yes Yes Covered by Risky Agents (preview) policy

Governance Levels

Level Requirement Rationale
Baseline Enable IRM with Risky AI usage as the primary Copilot template; add Data leaks or Data leaks by priority users where the business case needs traditional exfiltration triggers; enable Generative AI apps indicators and Risky AI usage indicators (preview) at Insider Risk Management > Settings > Policy indicators; review the default Risky Agents (preview) policy; configure alerts to notify the compliance team; review alerts weekly; PAYG note: non-Microsoft 365 AI data in generative-AI indicators can require pay-as-you-go billing, while Microsoft 365 Copilot data does not Provides foundational insider risk detection and agent monitoring without assuming that optional Security Copilot agent features are already licensed or approved
Recommended Add Data theft by departing users with HR connector; configure Adaptive Protection to dynamically adjust Copilot DLP; create custom indicators for FSI-specific patterns (MNPI probing, client data aggregation); enable risk score boosters such as Activity above user's typical day, Priority user group member, and Potential high-impact user; enable Communication Compliance indicators where appropriate; integrate DSPM for AI / DSPM findings into investigation workflows; if Security Copilot SCUs and pay-as-you-go are approved, deploy the IRM Triage Agent in manual mode first and move to scheduled runs after validation; use the data risk graph while available if Sentinel data lake onboarding is complete Comprehensive insider risk program incorporating AI-specific indicators, optional triage automation, and agent governance — suitable for most FSI firms
Regulated All Recommended requirements plus: Risky Agents alerts reviewed within 24 hours with mandatory investigation tracking; custom policy with the required Generative AI apps and Risky AI usage indicators; Adaptive Protection enforced at all risk levels; DSPM prompt/response monitoring integrated with IRM investigations for full Copilot interaction visibility; if deployed, IRM Triage Agent with human-in-the-loop validation and documented model governance per SR 11-7; legal hold integration; quarterly IRM effectiveness audit including agent risk assessment and DSPM signal coverage review; annual red team exercise testing Copilot-based insider threat scenarios; documented replacement plan for data risk graph before the November 24, 2026 retirement Full insider risk governance with continuous improvement, explicit preview-feature governance, and board-ready evidence for firms with active insider threat programs

Setup & Configuration

Step 1: Enable Insider Risk Management

Portal: Microsoft Purview > Insider Risk Management > Settings

  1. Complete the Insider Risk Management setup wizard
  2. Configure privacy settings (anonymization for initial alerts)
  3. Enable required data connectors (HR, DLP, device)

Portal: Microsoft Purview > Settings > Data connectors > HR connector

  1. Set up the HR data connector to import termination and resignation dates
  2. Map HR system fields to IRM schema
  3. Configure automated import schedule (daily recommended)
  4. This enables the "Data theft by departing users" template

Step 3: Enable Generative AI Apps and Risky AI Usage Indicators

Portal: Microsoft Purview > Insider Risk Management > Settings > Policy indicators

  1. Navigate to Policy indicators settings
  2. Expand Generative AI apps indicators and choose which Microsoft Copilot experiences and any connected enterprise AI apps you want to monitor
  3. Enable the Risky AI usage indicators (preview) and any companion indicators needed for your use case
  4. Configure only the documented indicator settings and thresholds that the policy workflow exposes for the selected indicators
  5. If Insider Risk analytics is enabled, use the documented real-time analytics recommendations where available instead of inventing unsupported fixed tuning windows

Step 4: Review and Configure Risky Agents Policy

Portal: Microsoft Purview > Insider Risk Management > Policies

  1. Locate the default Risky Agents (preview) policy in the policy list
  2. Review the policy scope — confirm it covers Copilot Studio agents, Microsoft Foundry agents, and agents built using the P4AI SDK deployed in the tenant
  3. Review the default Risky Agents coverage. If it doesn't meet your needs, create a custom Insider Risk Management policy instead of assuming the default policy is directly editable.
  4. Configure alert routing: route agent risk alerts to both the compliance team and the agent deployment owner
  5. Note agents not yet covered (prebuilt Microsoft agents, third-party, SharePoint agents) and apply compensating monitoring via DSPM for AI or Defender for Cloud Apps

Step 5: Create IRM Policy for Copilot

Portal: Microsoft Purview > Insider Risk Management > Policies > Create policy

Policy 1: Risky AI Usage for Copilot

Setting Value
Template Risky AI usage
Name FSI-Copilot-Risky-AI-Usage
Users All Copilot-licensed users
Indicators Generative AI apps indicators + Risky AI usage indicators + any companion office or device indicators you approve
Indicator settings Configure supported indicator thresholds in the policy workflow, or use real-time analytics recommendations where Learn documents them as available
Alert volume Medium (balance detection with noise)

Policy 2: Departing User Copilot Monitoring

Use the Data theft by departing users template when HR or account-deletion triggers are required for a named risk population.

Setting Value
Template Data theft by departing users
Name FSI-Copilot-Departing-Users
Triggering event HR connector — resignation or termination date
Indicators All Microsoft 365 indicators + Copilot + file download + USB/print + AI usage
Thresholds Lower thresholds (more sensitive) for departing users
Policy timeframe 90 days before departure date to 30 days after

Step 6: Configure Custom Indicators

Create custom indicators for FSI-specific Copilot risks:

Indicator Name Signal Example organization-defined heuristic
Copilot MNPI probing Copilot prompts or responses involving MNPI keywords Example only: repeated MNPI-related audit hits within a local review window
Copilot bulk extraction Copilot audit activity plus correlated file access or export evidence Example only: organization-defined outlier review based on audit or reporting data
Copilot client data aggregation Copilot prompts or responses referencing client names/accounts Example only: repeated client-reference patterns identified in approved evidence sources
Copilot off-hours usage Copilot interactions outside business hours Example only: after-hours audit activity correlated with other risk factors
Agent data volume anomaly Agent access or tool activity from approved evidence sources Example only: organization-defined outlier review paired with Risky Agents or Defender evidence

These are illustrative organization-defined tuning examples, not Microsoft-documented built-in IRM indicators or published default thresholds.

Step 7: Enable Adaptive Protection

Portal: Microsoft Purview > Insider Risk Management > Adaptive Protection

  1. Enable Adaptive Protection
  2. Configure risk levels mapping to DLP enforcement:
  3. Elevated risk → Block Copilot from processing sensitive content
  4. Moderate risk → Enhanced Copilot DLP monitoring
  5. Set automatic risk level assignment based on IRM alerts
  6. Coordinate with the Conditional Access configuration in Control 2.3 to enable CA-level Copilot access restriction for high-risk users

Step 8: Set Up Data Risk Graphs

Portal: Microsoft Purview > Insider Risk Management > Recommended actions

  1. For tenants that onboarded before September 24, 2026, select the Set up data lake and data risk graph recommended action and complete the Microsoft Sentinel data lake onboarding (the data lake uses pay-as-you-go billing). Current Learn states new data risk graph onboarding closed on September 24, 2026 for Insider Risk Management.
  2. After setup, open an alert at Insider Risk Management > Alerts (preview) and select the Data risk graph tab to review the connected assets, users, and exfiltration activity
  3. Use the graph's growing 7-day-to-30-day summary window of alert-related SharePoint and OneDrive exfiltration activity (sharing links, downloads, renames) to add investigative context while the feature remains available through November 24, 2026
  4. Incorporate data risk graph review into investigation procedures for alerts involving potential cross-department data movement

Step 9: Enable IRM Triage Agent

Portal: Microsoft Purview > Agents > Explore agents, then Insider Risk Management > Alerts (preview)

  1. Enable the IRM Triage Agent only after verifying that Security Copilot prerequisites are met, including SCUs, pay-as-you-go billing, Microsoft 365 data sharing, and the Microsoft Purview plug-in
  2. Deploy the agent from Agents > Explore agents and configure it to run manually on one alert at a time initially (Baseline)
  3. Review Triage Agent categorizations for the first 30 days to validate accuracy
  4. After validation, configure scheduled automatic runs so Triage Agent recommendations can drive investigator queue order (Recommended)
  5. For Regulated: document a human-in-the-loop workflow so investigators review Triage Agent recommendations before dismissing alerts
  6. Document the Triage Agent in the firm's model inventory per OCC Bulletin 2011-12 (SR 11-7)

Step 10: Configure Alert and Investigation Workflow

  1. Define alert triage responsibilities (compliance team, security team)
  2. Configure alert notification routing (including agent-specific alert routing)
  3. Establish investigation procedures that include:
  4. Alert review and initial assessment (with Triage Agent context summary)
  5. Activity timeline reconstruction (Copilot interactions + agent interactions + correlated signals)
  6. Data risk graph review for cross-department access patterns
  7. Escalation criteria (when to involve legal, HR, management)
  8. Documentation requirements for regulatory evidence

Financial Sector Considerations

  • Departing Financial Advisors: When registered representatives or financial advisors depart, they may use Copilot to rapidly compile client lists, portfolio summaries, and contact information. IRM policies should trigger enhanced monitoring immediately upon resignation notice. This addresses a long-standing industry concern about "book theft." Organizations can use departure-related policies together with local audit review heuristics to watch for suspicious late-stage Copilot usage without assuming a built-in IRM volume threshold.
  • MNPI Abuse Detection: Employees with MNPI access who use Copilot to gather related market information may be engaging in insider trading preparation. Custom indicators should detect patterns where users query Copilot about companies they have MNPI about and then conduct related searches. If the resulting alert also involves SharePoint or OneDrive activity (sharing, downloads, renames of related files), the data risk graph on that alert can help visualize the connected file-exfiltration context.
  • Agent Supervisory Obligations: Per FINRA 2026 Oversight Report, the deployment of Copilot Studio, Microsoft Foundry, and P4AI SDK agents creates supervisory obligations that parallel traditional employee supervision. Firms should treat Risky Agents alerts with the same investigation rigor applied to human user alerts.
  • IRM Triage Agent Model Risk: The Triage Agent is an AI model that makes recommendations about insider risk alert severity. Per OCC Bulletin 2011-12 (SR 11-7), the firm should document the Triage Agent as a model, validate its performance against human investigator judgments, and establish escalation procedures for systematic categorization errors.
  • Privacy Considerations: IRM involves monitoring employee activity, which raises privacy concerns. Financial firms should document the business justification for IRM monitoring (regulatory requirement), communicate monitoring policies to employees, and use anonymization features for initial alert triage.
  • Union and Employment Law: Some jurisdictions have laws governing employee monitoring. Consult with employment counsel to confirm IRM monitoring practices comply with applicable labor and privacy laws.
  • Investigation Documentation: IRM investigations should follow documented procedures that preserve evidence for potential regulatory referrals, employment actions, or litigation. Integrate IRM investigation workflows with the firm's existing internal investigation procedures.
  • Regulatory Reporting: If IRM investigations reveal potential securities law violations (insider trading, market manipulation), the firm may have reporting obligations to FINRA or the SEC. Establish escalation procedures that include legal counsel and compliance leadership.
  • Board Reporting: For Regulated-level implementations, include IRM metrics in quarterly board risk reports. Metrics should include: number of Copilot-related alerts, agent risk alerts, investigations opened, investigations resulting in action, and risk trend analysis.

Verification Criteria

  1. IRM Policy Active: Verify at least one IRM policy is active that includes Copilot activity signals
  2. Risky Agents Policy: Confirm the default Risky Agents (preview) policy is present and active for Copilot Studio, Microsoft Foundry, and P4AI SDK agents; verify alert routing is configured
  3. Generative AI Indicators: Confirm the required Generative AI apps indicators and any Risky AI usage indicators (preview) are enabled and tuned appropriately for the organization's Copilot deployment scale
  4. IRM Triage Agent: If the organization deploys the Triage Agent, verify it is deployed through Purview Agents, producing context summaries for IRM alerts, and documented in the model inventory
  5. Data Risk Graphs: For tenants that previously onboarded, confirm the data lake and data risk graph recommended action has already been completed, that the Data risk graph tab is accessible from IRM Alerts (preview), and that the organization has a replacement plan before the November 24, 2026 retirement date
  6. Signal Collection: Confirm that Copilot interaction data appears in the IRM signal timeline for test users
  7. HR Connector (if applicable): Verify the HR data connector is importing termination/resignation data successfully
  8. Alert Generation: Trigger a test scenario based on a configured risky-AI or companion signal and confirm an IRM alert is generated
  9. Alert Routing: Verify that IRM alerts are delivered to the designated compliance team within the defined SLA; agent risk alerts route to both compliance and agent deployment owners
  10. Adaptive Protection: Confirm that Adaptive Protection is enabled and that risk level changes result in corresponding DLP enforcement adjustments for Copilot
  11. Custom Indicators: Verify that any FSI-specific custom indicators the organization chose to deploy (for example, MNPI probing or client-data aggregation) are configured and functional
  12. Investigation Workflow: Confirm documented procedures exist for investigating IRM alerts related to Copilot and agent usage, incorporating Triage Agent recommendations
  13. Privacy Controls: Verify that anonymization is enabled for initial alert triage and that access to IRM investigation data is restricted to authorized investigators
  14. Periodic Review: Confirm that IRM policies are reviewed quarterly with threshold adjustments based on false positive rates and detection effectiveness

Additional Resources