How to Read the Unsanctioned AI (Shadow AI) Apps Section

Use this topic when you want to understand what the Unsanctioned AI Apps section of the AI Governance Dashboard shows, and how to use it to find the unapproved AI tools — the Shadow AI — already in use across your organisation.

Before you begin

Your organisation's web traffic must flow through the Forcepoint Secure Web Gateway for unsanctioned AI activity to be detected. No app connector or additional configuration is required on the AI Security side.

To understand where this data comes from and why an app may or may not appear, see Understand How Shadow AI Detection Works.

The Unsanctioned AI Apps section lists the AI tools your employees are using that you have not approved and connected. This is also called Shadow AI.

An app appears here precisely because no connector has been configured for it. Everything that is not in the Sanctioned AI Apps section shows up here instead, which makes this section your discovery list — the starting point for deciding what to approve and what to restrict.

Steps

  1. Sign in to the Forcepoint Data Security Cloud Platform.
  2. In the left navigation, under APPLICATIONS, click AI Security, then click Dashboard.
  3. Click the AI Tool Landscape tab.
  4. Set the time range using the filter at the top of the page.
    Note: Every figure in this section reflects the selected time range. Widen the range if you are looking for a trend rather than current activity.
  5. Scroll to the Unsanctioned AI Apps section.

    Each detected AI tool appears as a row in the table.

  6. Read the columns in the table.
    Column What it tells you
    AI Application The name of the detected tool, with the domain your users reached it on beneath it. The domain is useful when two tools have similar names.
    Risk The risk level assigned to the application — High, Medium, or Low. Use this to decide which rows to investigate first.
    Category The kind of AI tool it is, such as Conversational AI Assistant or AI Meeting & Transcription. This is the cloud app category used to identify the traffic as AI in the first place.
    Active Users The number of distinct people who used the app during the selected period.
    Activity Mix How the activity splits across Visits, Uploads, and Downloads, shown as a bar with percentages. A high upload share means data is leaving your organisation, not just being browsed.
    Total The total activity count for the app across all three activity types.
    Note: Sorting is not available in this Early Access (EA) release. Although arrow indicators appear on the column headers, clicking them does not reorder the table. Column sorting will be enabled in a future release.
  7. Check the colour key above the table to interpret the Activity Mix bars.

    The key identifies which colour represents Visits, Uploads, and Downloads.

    Note: Uploads are the most important figure for data protection. An app with a low risk rating but a high upload percentage still represents data leaving your organisation through a channel you do not monitor.
  8. Use the paging controls at the bottom of the table to see the remaining apps.

    The control on the left shows how many apps are being displayed out of the total. Use Rows per page to show more apps at once.

  9. Click a row to open the full detail view for that app.

    The detail view opens with the app name, its risk badge, and a breadcrumb trail showing AI Governance > Shadow AI > the app name.

  10. Review the Users Accessing table.

    This is where you find out exactly who has been using the app.

    Column What it tells you
    User The email address of each person who accessed the app during the selected period.
    Activity Mix How that individual's activity splits across visits, uploads, and downloads. This is the same colour key used on the main table.
    Total That user's total activity count for this app.
    Note: The figure in the top right of the panel shows the total number of users accessing the app — for example, All 20. This matches the Active Users figure for the app on the main table.
  11. Identify the individuals with the highest upload percentages.

    A user whose activity mix is weighted towards uploads is sending data out to the app rather than simply browsing it. These are the people to talk to first.

  12. Use the paging controls at the bottom of the panel to see the remaining users.

    Adjust Rows per page to show more users at once.

  13. Click Shadow AI in the breadcrumb trail to return to the dashboard.
  14. Decide what action to take for each app that concerns you.
    • To approve the app and gain full visibility, connect it as a sanctioned app, if a connector is available for it. Its row is then removed from this table and it appears as a card in the Sanctioned AI Apps section. See Connect AI Apps via Easy Setup.
    • To restrict or block access to the app, use your web security policies.
    • To leave the app available but keep watching it, note its Active Users and upload share and review the figures again after widening the time range.
    Note: Connecting an app is what gives you prompt-level and file-level visibility. This section tells you which app is being used, by whom, and how heavily — but not what was typed or which files were uploaded.
  15. (Optional) To inspect the raw web records behind these figures, go to Insights > Transaction Viewer and select the web collection.

    Scroll right to the Cloud App Category column to see how each individual request was classified.

Result

You can see which unapproved AI tools are in use across your organisation, how risky each one is, how many people use it, and how much data is being uploaded to it — so you can decide which tools to sanction, which to restrict, and which to keep watching.

Next steps

For an organisation-wide view of shadow AI usage rather than a per-app view, check the Shadow AI Users figure in the Identity Landscape section. See How to Read the Identity Landscape Section.