Overview
The Insights page gives admins a shareable view of GenAI usage, risk, and adoption across the organization. It combines filtering with drill-through access to the underlying transactions. This lets admins move from trend to investigation in a single click. Dashboards export as PDFs, and email subscriptions keep stakeholders informed automatically.
Insights reports sensitive data exposure across GenAI usage and adoption rates and usage patterns by department and service. It also reports policy enforcement outcomes. These dashboards update in real time, replacing manual analysis.
Prerequisites
Administrative access to the F5 Workforce AI Security admin interface.
Directory Sync (SCIM) recommended for Group-based filtering.
Available Insights
The Insights page has four tabs. Each tab includes tiles, charts, and tables. These can be reconfigured and filtered.
Public Services: Usage, adoption, and risk for public GenAI services such as ChatGPT and Claude. Shows volumes, top services, intents, violations, and policy outcomes. This includes shadow AI usage across the organization.
Risk: Quantifies and monitors risk across both Public Services and private models. This includes sensitive data detections, violation rates, and blocked requests.
Adoption: Who is using GenAI and for what, across Public Services and private models. Breakdowns by group, user, intent, and service show growth. They also show use cases.
Private models: How users are using private models through the Gateway. It also covers the Model Router.
Working with charts and tables
Charts and tables are interactive and customizable.
Click a chart or table title to change the grouping or update summarizations. You can also switch visualization types.
Hover or right-click on chart segments to access context actions such as "See these transactions" or "Break out by…".
Use tool tips to see exact values. They also show date buckets.
Drill-through: See these transactions
From any data point, click it, then select See these transactions. This opens a tabular view of the underlying data used to render the visualization. From there, admins can:
Add additional filters.
Change groupings and summarizations.
Export rows, or pivot to investigate specific users, services, or violations.
Filtering the dashboard
Insights supports filtering with unlimited look-back.
Date Range: Choose from presets like Today, Previous 7 days, Previous 30 days, or Previous month. You can also select Specific dates and Relative dates. Unlimited look-back lets admins analyze any time frame across the tenant's history.
Groups: Filter by any Group synced via SCIM or created manually in the tenant. This gives focused views by department or region (for example, Engineering or HR).
Filters apply to the entire dashboard, including tiles, charts, and tables. PDF exports and email subscriptions respect these filters too.
Share and automate
Export as PDF
Download the current dashboard, with all active filters, as a PDF. Use it for distribution to leaders or for inclusion in status reports.
Steps
Apply the desired Date Range and Group filters.
In the top-right of the Insights page, select Export as PDF.
Save or share the generated file.
Email subscriptions
Subscriptions send the dashboard on a schedule to a specified address and other stakeholders. Each subscription preserves its own Date Range and Group filters. This lets different leaders receive the view they need.
For information on how to create subscriptions, see Insights Subscriptions.
Example use cases
Identify security risk trends: Use Risk insights to identify spikes in sensitive data detections across departments. Then drill through to the underlying transactions to investigate.
Report AI adoption to executives: Generate weekly PDF reports filtered by group. These reports show public service and private model adoption trends to stakeholders.
Review software costs: Analyze Public Services usage trends to see which GenAI services teams use most. These trends can inform license and pricing decisions.
Support compliance audits: Export filtered dashboards showing policy enforcement rates and data protection controls. Use them as evidence for auditors.
Terminology
Intent: The classified purpose of a request (for example, General, Marketing, Product management).
Policy Decision: The action taken by Workforce AI Security (for example, Intercept, Redact, Block).
Sensitive Data Detection: Identification of PII or other restricted content. This is based on configured policies.