Introducing AI Assistant for Enterprise Administrators

  • Updated
Follow

This AI-powered reporting and insights assistant helps Enterprise administrators turn learner data into action.

Using natural-language questions, administrators can quickly explore learner progress, engagement, and training data without having to manually work through reports. The AI Assistant brings the relevant information together and transforms it into clear, actionable summaries to help administrators identify who is engaged, who may need support, where outreach may be most valuable, and what trends deserve attention.

These insights can also help administrators take the next step, from prioritizing learner outreach to designing custom learning paths that better align training with individual learner goals.

AI Assistant Demo Video

Who can use the AI Assistant?

It is designed for Enterprise administrators. 

It is not a learner-facing feature. Learners cannot view the classifications, recommendations, or outreach priorities generated for administrators.

Availability may depend on your organization’s account settings and the current product rollout.

Where can I find it?

It is accessed through the Admin Console. Selecting Assistant opens a panel on the right side of the page. When accessing the AI Assistant for the first time, you will be prompted to provide consent for the AI Assistant to access your data.

This allows you to interact with the AI Assistant while reviewing the underlying Dashboard or Report.

What can the AI Assistant help me do?

The AI Assistant can help administrators turn training data into useful insights and next steps. You can use it to:

  • Understand learner engagement and progress and quickly identify meaningful patterns in training activity.
  • Identify learners who may be stalled or at risk and understand where additional support may be needed.
  • Compare activity over time to see how learner engagement and progress are changing month over month.
  • Analyze hands-on lab participation and identify learners who may be engaging with content but not applying their skills in labs.
  • Review assessment and exam history to better understand learner outcomes and continued progress.
  • Prioritize learner outreach by identifying who may benefit from attention and why.
  • Understand skills and content engagement to see where learners are focusing their training.
  • Spot team-level training trends and turn reporting data into clear, actionable summaries.
  • Design custom learning paths based on a learner’s progress, activity, and goals.

Instead of manually working through raw reporting data, administrators can ask questions, explore insights, and use the results to determine what action may be most valuable.

Learner status classifications

The AI Assistant may classify learners as:

  • Active: The learner has demonstrated recent hands-on engagement.
  • Non-active: The learner shows extended inactivity, limited meaningful progress, or repeated disengagement.

Hands-on activity is the primary indicator of active engagement. Reading course material or watching videos may provide useful context, but those activities alone do not count as hands-on engagement.

The same learner-status logic is used across summaries and outreach recommendations so classifications remain consistent.

What hands-on lab activity is included?

It considers hands-on activity across supported OffSec lab content, including:

  • Course and module labs
  • Challenge labs
  • Proving Grounds Play
  • Proving Grounds Practice
  • OCR
  • DCR
  • Shiny Labs
  • Grimoires

Where a lab includes a flag, proof, exercise, exam, or assessment submission, that activity may count as hands-on engagement.

Where a lab does not include a submission, a meaningful lab session may count. Simply opening and immediately closing a lab may not qualify as engagement.

It only considers lab products available to your organization. If your account does not include access to a particular product, it will not be included in your results.

Suggested questions

When starting a new conversation, AI Assistant may display four suggested questions based on the page or report you are viewing.

You can also enter your own question.

Examples include:

  • Which learners need attention?
  • Who has been inactive or has made limited progress?
  • Which learners are engaging with course content but not completing hands-on labs?
  • How has learner activity changed compared with the previous month?
  • Which labs appear to be causing difficulty?
  • Which learners should I contact first?
  • Can you summarize this report for my monthly review?

NOTE: The more specific and information-rich the question, the more relevant and accurate the response will be.

Ranked outreach recommendations

The AI Assistant can generate a prioritized learner-outreach list.

The list may consider:

  • Length of inactivity
  • Lack of progress
  • Current learner status
  • Proximity to a training milestone
  • Engagement with hands-on content

Recommendations may include a brief reason and a suggested message that administrators can adapt for email or direct outreach.

It does not send messages or notifications to learners. Any outreach must be reviewed and sent by the administrator.

Assessment and exam history

Where data is available, the AI Assistant can help summarize a learner’s assessment and exam history, including:

  • Passed attempts
  • Failed attempts
  • Repeated attempts
  • Time between attempts
  • Activity after passing an assessment or exam

This can help administrators understand whether a learner continued progressing, moved to another area, or disengaged after an attempt.

Does the AI Assistant follow report filters?

When supported, it uses the context of the page you are viewing, including active filters.

This may include:

  • Reporting period
  • Team
  • Learner classification
  • Job title
  • Content type
  • License type
  • Selected learner or report tab

The AI Assistant panel displays the scope used for the response so you can understand which information was included.

When you navigate to another page, the AI Assistant panel will close. Your conversation is automatically preserved and can be accessed again when you reopen the AI Assistant. Your chat history is available in the upper-left section of the panel.

Does the AI calculate its own numbers?

No. The AI does not rely on the language model itself to perform mathematical calculations.

Numbers, totals, percentages, comparisons, and classifications are generated using platform tools, data services, and data already available in existing reports. The AI then uses those results to explain the information in natural language.

Responses that include numbers, calculations, estimates, classifications, or projections may also include a reminder that AI-generated information can contain errors and should be reviewed and verified.

What happens when there is not enough information?

The AI Assistant is designed to clearly state when the available data is not sufficient to answer a question.

It should not:

  • Invent missing learner activity
  • Create unsupported calculations
  • Present an estimate as a confirmed fact
  • Generate a learner trajectory that the available data cannot support

When possible, the response will indicate what information is missing.

Is the AI Assistant read-only?

Yes. It is read-only.

It can summarize and interpret data and may assist with recommendations, such as helping you design a custom learning path based on available information. However, it cannot directly apply or save changes in the platform.

It cannot:

  • Assign or remove licenses
  • Edit learner information
  • Change account settings
  • Modify reports
  • Send emails, messages, or notifications
  • Make changes to training records

What information can Admins access?

The AI follows existing Enterprise account-access rules.

Administrators can access information available within their organization, including aggregate learner progress, outcomes, and supported learner-level training information.

It does not provide administrators with the private conversation or session history of a learner’s interactions with Kai or another AI agent.

Information is restricted to the administrator’s own organization. It does not provide cross-customer comparisons or shared data from other organizations.

AI access and organization settings

NOTE: If AI features are disabled at the organization level, users within that account will not be able to access the applicable AI experiences.

Where required, users may also be prompted to provide consent before an AI feature can access and use the data available to them.

Can I export a response?

Yes, for supported response types.

The AI can generate PDF files that can be downloaded for sharing or reference. CSV generation is also planned for supported data and list-based responses.

Export availability may depend on the type of information returned, so not every response will necessarily support the same download format.

What should I do if an answer appears incorrect?

Compare AI Assistant’s response with the relevant Admin Dashboard or Report.

If the answer appears incorrect or misleading, contact OffSec Support and include:

  • The question entered
  • AI Assistant’s complete response
  • The Admin Console page used
  • Any active filters
  • The learner or report scope
  • A screenshot showing both the response and the underlying page

This information will help our team investigate the issue.

AI Assistant Usage Limits

To help manage AI Assistant usage, each Enterprise account has a daily usage limit.

What is the daily limit?

By default, an account can use up to $10 of AI Assistant model usage per day.

This limit is:

  • Shared across the entire account, rather than assigned to individual administrators
  • Reset each day at midnight UTC
  • Applied only to usage of the AI Assistant in Admin Console

Usage from other OffSec AI experiences, including KAI, OSCAR, AI Path Creation, and Upskilling Preferences, does not count toward this limit.

Is the limit per administrator?

No. The daily limit applies to the account as a whole.

All administrators using AI Assistant within the same account contribute to the same daily allowance. This means higher usage from one administrator can affect how much AI Assistant usage remains available to other administrators in that account.

What happens when the limit is reached?

Once the account reaches its daily AI Assistant usage limit, the Assistant will temporarily stop generating new responses for that account.

Access automatically becomes available again when the daily usage period resets at midnight UTC.

Where can administrators see usage?

In the left-side navigation, go to:

Admin Console → Team Management → Assistant Usage

The usage page shows:

  • Today's total AI Assistant usage for the account compared with its daily limit
  • A breakdown of usage by administrator for the previous 30 days

Because the limit is shared at the account level, individual administrator usage may appear significantly lower than the overall account total.

Powered by Zendesk