Overview

The Workforce AI Security Overview is the central interface for monitoring generative AI activity in your organization. It serves as an executive summary for administrators, providing a high-level view of generative AI usage and security posture across the organization. Displaying critical metrics related to AI application usage, it shows you:

  • Risk indicators for AI interactions

  • User activity and adoption trends

  • Policy enforcement status

  • Summary of detected violations

Top Applications

The Top Applications section highlights the total AI traffic and the four AI applications (chats) with the highest user traffic in your organization. This section provides a quick, actionable view of where users are most active and which applications may pose security concerns.

Key Metrics and Indicators

Each application is displayed with key usage and risk information to help administrators assess organizational exposure and adoption trends.

Risk-Level Color Indicators

Each tile is outlined with a color representing the application’s risk level:

  • Red — Critical risk

  • Yellow — High risk

  • Orange — Medium risk

This visual coding gives administrators an immediate sense of the organization’s security posture.

Application Details

For each top application, the section displays:

  • Management status: Whether the application is managed (covered by an organizational license) or unmanaged.

  • User count: The number of users who accessed the application.

  • Session count: The total number of sessions initiated

  • Number of prompts: Total prompts submitted by users.

  • File uploads: Number of files uploaded to the application.

Usage Trend Indicator

A trend marker shows how the application’s usage has changed over the selected timeframe. This helps administrators:

  • Monitor spikes or drops in adoption.

  • Evaluate user behavior.

  • Understand whether recently implemented security policies are having an impact - especially for higher-risk applications.

Operational Insights

The Top Applications section provides a clear and actionable snapshot of:

  • The most frequently used AI applications.

  • Their associated risk levels.

  • How usage patterns evolve over time.

This helps organizations quickly identify areas that may require policy enforcement, user education, or additional security controls.

Sensitive Data & Use Cases

The Sensitive Data & Use Cases section on the Overview screen provides a consolidated view of the types of sensitive information and use cases associated with AI application activity across the organization. This high-level summary helps administrators understand how AI tools are being used and what kinds of data they process, supporting informed risk assessments and policy decisions.

Key Metrics and Indicators

This section aggregates sensitive data trends across all monitored applications, offering an organization-wide perspective rather than application-specific details.

Sensitive data

The visualization highlights the five most common sensitive data types sent to AI applications. These categories reflect the data most frequently processed and help administrators identify potential risk hotspots, such as:

  • Corporate confidential information

  • Personal data

  • Financial or regulatory-sensitive content

Space is reserved to display additional categories as needed.

Use cases

Administrators can also view the top five use cases driving AI adoption across the organization. This breakdown helps clarify why users interact with these applications—for example, content generation, coding assistance, data analysis, or communication tasks.

Security Insights

By summarizing the most common sensitive data types and use cases, this widget helps administrators:

  • Quickly identify the kinds of sensitive content most often sent to AI tools

  • Understand how users are leveraging AI across the organization

  • Evaluate risk exposure related to data movement and usage patterns

  • Inform policy decisions and guide safe AI adoption

This consolidated view supports strategic oversight while maintaining consistency with more detailed analytics elsewhere in the platform.

Top-Risk Users

This widget identifies the users who exhibit the highest levels of risky activity when interacting with generative AI applications or MCP Servers. This view surfaces a small set of users - typically the top five - who accumulated the greatest number of sessions involving high-risk applications or agents. By highlighting these individuals, the section helps administrators quickly pinpoint where risky behavior is concentrated and where additional oversight, training, or policy enforcement may be required.

For each user, the dashboard displays the total number of high-risk sessions, allowing administrators to understand not only who the most impacted users are but also the relative scale of their activity. This targeted insight supports faster investigation and more informed decision-making, especially in environments where large user populations make manual analysis impractical. The section aligns with the broader overview by providing organization-wide visibility into risk exposure patterns while maintaining a clear focus on users whose behavior may require immediate attention.

Policy Enforcement

The Policy Enforcement widget provides an at-a-glance breakdown of how organizational policies are applied across AI-related user sessions. It presents a distribution of policy actions - such as Allow, Ask, Prevent, and Block - showing the percentage of sessions that triggered each enforcement outcome. This visualization helps administrators understand how effectively existing policies shape user behavior and where further adjustments may be needed.

The widget represents how policies influence activity: for example, how many sessions were blocked due to high-risk behavior, how many were assisted with guidance, and how many proceeded normally. This breakdown helps administrators evaluate whether policies are achieving having the intended effect, such as reducing risky interactions or increasing adherence to organizational guidelines.

The Total AI Traffic graph shows the proportion of organizational traffic flowing through applications that are formally managed versus those that are not. It helps administrators gauge how much user activity occurs in controlled, monitored environments compared to potentially riskier, unmanaged ones. An upward shift in managed application usage - and a decline in unmanaged usage - indicates stronger policy effectiveness and improved overall security posture.

Application Risk by Usage

The Application risk by usage matrix provides a visual summary of how AI applications are used across the organization. It combines each application's inherent risk level (based solely on the application itself, not on how users interact with it) with its overall usage volume. The matrix is divided into four quadrants formed by two axes: the X-axis represents usage levels, and the Y-axis represents application risk. Low-risk applications with minimal usage appear in the lower-left quadrant, while high-risk applications with heavy usage appear in the upper-right quadrant - representing the combination most relevant for administrative review.

Each application is displayed as a bubble whose size is proportional to its overall usage, allowing administrators to quickly identify which applications drive the most activity. The border color of each bubble corresponds to the application’s risk level, using the same color legend applied throughout the dashboard.

This matrix helps administrators understand how application usage is distributed, identify clusters of risky or heavily used tools, and determine where policy enforcement or additional controls may be needed. By mapping applications according to actual use and assigned risk, the matrix supports quick prioritization and clearer visibility into the organization’s exposure landscape.

The Top Used Agents by Invocations widget highlights the most frequently used pairings of AI platforms and their corresponding MCP Servers (agents), providing deeper visibility into how users interact with generative AI tools across the environment. Each widget row represents a unique combination of platform (the underlying AI provider or interface) and MCP Server. This view allows administrators to identify which agent–platform pairs drive the highest activity and therefore represent the most influential operational patterns within the organization.

For each combination, the dashboard displays the total session count, reflecting overall usage volume. This helps administrators understand which platform–agent pairs are most widely adopted, regardless of whether the applications themselves appear in the top applications listing. By surfacing these combinations, the section supports more effective monitoring of agent behavior, reveals where organizational workflows are concentrated, and helps pinpoint areas where policy enforcement or additional controls may be required.

This aggregated view aligns with the broader overview structure while offering an additional layer of insight beyond application-level metrics. It enables administrators to evaluate not only which AI applications are in use, but also how they are accessed—through which agents, in which configurations, and at what scale—strengthening their ability to manage risk and optimize AI governance.