Manage risks, alerts, and activities
Detecting a risk is not blocking it
If an insider risk alert flags a departing employee, what stops the next file from leaving? Most of this domain detects, records and investigates rather than blocks: Microsoft Purview Insider Risk Management uses different types of events and indicators to collect signals and create alerts, the unified audit log (the tenant-wide store of audited activity) captures, records, and retains thousands of user and admin operations, and collection policies filter which events are ingested into Purview and can't impact the behavior of DLP policies. An enforcing control such as a data loss prevention (DLP) policy stops the action, and Adaptive Protection bridges detection and enforcement: it dynamically assigns appropriate data loss prevention, data lifecycle management, and Conditional Access policies to users based on the insider risk levels (Elevated, Moderate, or Minor) that Insider Risk Management assigns, so a DLP rule can block users with the Elevated risk level from saving or sharing sensitive data. Each level comes from admin-defined conditions and is neither a policy's risk score nor an alert severity (Low, Medium, or High). The classic trap is answering a requirement to stop an action with a control that only detects or records, such as an insider risk policy, an audit log retention policy (which sets how long audit records are kept) or a collection policy.
The domain unfolds in three steps: score risky users, investigate activity and content, protect data used by AI
All three SC-401 domains carry the same published weight, 30–35% of the exam; read this domain's three steps in order: investigation responds to alerts, including DLP alerts, Purview alerts in Microsoft Defender XDR and the alerts that insider risk policies generate, and protecting data used by AI applies scoring and investigation to Copilot and other AI apps alongside its own access, processing and compliance controls. To score risky users, Implement and manage Microsoft Purview Insider Risk Management separates collection from scoring: global indicators decide which activity is collected, a triggering event (such as a resignation date from the HR connector) determines if a user is active in an Insider Risk Management policy, and the policy indicators selected in that policy determine a risk score for an in-scope user. Its alerts become cases, and each case focuses on one user; reach for this topic when a requirement concerns one person's pattern of activity over time. To investigate activity and content, Manage information security alerts and activities reads activity records (Audit search over the unified audit log, activity explorer, the DLP Alerts dashboard and Purview alerts in Microsoft Defender XDR) and content (Microsoft Defender for Cloud Apps file policy matches, and eDiscovery searches that find in-place content such as email, documents, and instant messaging conversations). To protect data used by AI, Protect data used by AI services starts from the rule that all Microsoft 365 Copilot prompts run in the security context of the user who initiates the prompt, then stacks access controls (who can open an item), processing controls (whether an AI app may use what the user can open; examples are the Microsoft 365 Copilot and Copilot Chat location in DLP, called the Copilot DLP location below, and Endpoint DLP on onboarded devices) and compliance controls (what record of the interaction remains). Microsoft Purview Data Security Posture Management (DSPM) is the front door there: it monitors risks associated with AI apps and agents and creates one-click policies (preconfigured policies made from its setup tasks), and the earlier version, DSPM for AI, is now named DSPM for AI (classic) in the portal. The decision tree below routes a requirement through these steps and the detect-versus-block split, asking about AI first to route, not as the reading order: content reaching Copilot or another AI app goes to the controls for AI services and DSPM; stopping an action only for risky users is Adaptive Protection with DLP or Conditional Access; stopping it for all users in scope is an ordinary DLP policy, built in Create and configure data loss prevention policies; scoring one user's risk over time is an Insider Risk Management policy; and anything else is investigation with Audit, activity explorer, alerts and eDiscovery.
When evidence is missing, check setup and scope before concluding nothing happened
An empty score, search or dashboard in this domain is a setup or scope question before it is an answer. A user included in a policy but without a triggering event isn't evaluated as a potential risk; for core services such as Exchange and SharePoint, audit record availability is typically 60 to 90 minutes after an event occurs; a restricted administrator (one assigned administrative units) can only search and export user-generated audit logs within the scope of their assigned administrative units; and DSPM first shows an initial setup task for Auditing and analytics that is required to see meaningful data. Licenses follow the same instinct, because they attach to the people whose activity is recorded rather than to the investigator: retaining an audit log for longer than 180 days requires that the user who generates the audit log (by performing an audited activity) holds an Office 365 E5, Microsoft 365 E5, Microsoft Purview Suite or E5 eDiscovery and Audit add-on license, and DSPM insights for Microsoft 365 Copilot need the monitored users to be assigned a license for Microsoft 365 Copilot.
The three steps: what each reads, what it produces, and where enforcement comes from
| Step | Evidence it reads | What it produces | Where enforcement comes from | Drill into |
|---|---|---|---|---|
| Score risky users | Activity that global indicators collect, scored by policy indicators once a triggering event makes the user active | Risk scores, alerts, one-user cases, and insider risk levels | Adaptive Protection: DLP and Conditional Access policies that read the insider risk level | Implement and manage Microsoft Purview Insider Risk Management |
| Investigate activity and content | Activity records (unified audit log, activity explorer, DLP and Defender XDR alerts) and content (Cloud Apps file policy matches, eDiscovery searches) | Findings, with each outcome recorded in its own workflow, such as a DLP alert's Status | Response actions you choose on a finding, such as Remove public sharing on a Cloud Apps file policy match | Manage information security alerts and activities |
| Protect data used by AI | Copilot and other AI app interactions, recorded by Audit and collection policies | DSPM insights, one-click policies, and stored prompts and responses | Access controls and processing controls such as the Copilot DLP location; compliance controls only record | Protect data used by AI services |