The Insights dashboard in F5 Workforce AI Security shows how your organization uses generative AI (GenAI). It reports usage, risk, and adoption for public generative AI services and for Private Models.

In OBSERVE, go to Insights. You need admin access to the Admin UI to open the dashboard.

To filter by group, set up Directory Sync groups or create manual groups. Both group types work with the Group Name filter.

The Insights dashboard has six tabs. Each tab groups tiles and charts for one part of GenAI use.

The Public Services tab shows totals and weekly trends for public generative AI services. The totals cover transactions, intercepted requests, active users, and blocked or redirected requests. They also cover requests with sensitive data violations and personally identifiable information (PII) entity detections. Charts rank the top services by access, interception, and blocks. Diagrams map groups and users to services.

The MCP tab shows requests where Workforce AI Security detected a Model Context Protocol (MCP) tool. It also shows MCP tool violations and the number of tools discovered. Other charts show the most common tools and servers, and the tools in violation of policy. Further charts rank the top users and groups by tool execution.

The Intent tab shows what people use generative AI for. Intent is the classified purpose of a request, split into domain, action, and topic. The tab shows intent topics and domains by group and by service, and the popular intents. Diagrams map groups to domains, domains to actions, and actions to topics.

The Risk tab shows intercepted requests, sensitive data violations, requests with PII, and blocked or redirected requests. It also shows the overall risk level of requests over time, and the high-risk services with their users. Other charts show the groups that make high-risk requests, the top sensitive data entity types, and the users who send sensitive data.

The Adoption tab shows workforce adoption. Workforce adoption is the share of directory users who were active in the last 30 days. The tab also shows service usage, active users, services used, and services per active user. A heatmap shows users who use more than one service. Other charts break down use by user and by group, and show daily active users.

The Private Model tab shows how users use Private Models. Users reach Private Models through the Gateway and Model Router. The tab shows requests to Private Models, active users, total tokens, and the usage of public services compared with Private Models over time. Other charts show token use by model, user, and group, input and output tokens over time, and the top models and users.

All charts on the Intent tab count only the prompts that Workforce AI Security intercepted. So do most charts on the MCP and Risk tabs.

If a user belongs to more than one group, charts that split data by group count that user once for each group. As a result, the group totals might be higher than the organization total.

On every tab, you can filter by Date Range and Group Name. The default Date Range is the previous 30 days. Other filters depend on the tab: Auth Type, Service Type, Event Type, Risk Level, and Model Name.

These charts always show the last 30 days, regardless of the Date Range you select:

The Multi-Service Users Heatmap (30 Days) also ignores all other filters.

Select a data point on a chart to see the transactions behind it. You can then filter or group those transactions further.

You can download the dashboard. You can also send the dashboard to email recipients on a schedule. For more information, see Creating Insights subscriptions.

A security lead sees more sensitive data violations and needs to know where they come from. The Risk tab shows requests with sensitive data violations, the top sensitive data entity types, and the groups that make high-risk requests. The security lead selects a data point to see the transactions behind it. Then the security lead filters or groups those transactions to find the users and services involved.

The owner of a generative AI program must report adoption to leadership on a schedule. The Adoption tab shows the share of directory users who were active in the last 30 days, and the services those users use. The program owner filters the dashboard by group. Then the program owner sets up an email subscription, so that leaders receive the dashboard by email.

A platform owner needs to know which MCP tools people use. The MCP tab shows the tools discovered, the most common tools and servers, and the tools in violation of policy. It also shows the top users and groups by tool execution. The platform owner uses these charts to find the tools that break policy and the users who run them.

For more information, see: