AI Ticket Assistant - FAQ’s

A guided chat experience that helps staff submit IT tickets with clear, actionable, and IT-friendly data. It improves ticket quality, reduces back-and-forth, and supports faster resolution without replacing human agents.
​

No. AI assists with ticket intake and guided troubleshooting; humans retain full control over review, resolution, and decision-making.
​

Any user who submits IT tickets can use the AI Ticket Assistant, including teachers, staff members, agents, and administrators, provided they have the correct permissions.

Yes. AI adheres to the same permission rules as the current ticket wizard and will not surface users outside a user's permitted scope.
​

Yes. Access can be filtered by 'Location' and 'Role' using the Permission policy engine. Multiple locations use an OR relationship, while Location is an AND relationship with other active filters (e.g., Role AND Users).
​

Yes. AI access is managed via the 'Enable AI Access' permission within specific permission policies. This allows for controlled rollout, such as enabling it for Agents/Admins only before broader district-wide deployment.
​

If a user lacks permission for a ticket type (e.g., Facilities), they are redirected to the legacy submission workflow and shown an error message: “You do not have permission to submit a 'Facilities' ticket. Please check with an administrator to request elevated permissions.”
​

The AI will ask for clarification or present multiple assets for selection to prevent incorrect device association.
​

The AI Ticket Assistant cannot directly deactivate, lock, or deprovision a device. It can, however, help submit a ticket that initiates an existing device lock/disable workflow.
​

Yes. The AI Ticket Assistant attempts to match issues using the district's existing issue categories and issue types.
​

The AI uses semantic search for selection, so it can account for renamed issue categories when Custom Names are used.
​

Yes. Users can review and modify all ticket details before submission, including Issue Description, Requestor, Location, Asset/Model, Issue Type, and Has Sensitive Information.
​

Yes. All custom fields, including required and issue-specific fields from the standard ticket workflow, will appear in the AI flow.
​

Yes. AI-created tickets behave like manually created tickets; rules and automations run as usual.
​

Yes. There are two Source filters, “AI Chat” and “AI Email,” which allow filtering for AI-created tickets.
​

Not by default. They can be identified using the "Source" filters (“AI Chat,” “AI Email”). You can optionally configure Rules to update a ticket status or color for visual differentiation.
​

Yes. Historical ticket imports contribute to the AI context, as their descriptions and comments form the majority of the data the LLM uses for analysis.
​

The AI Ticket Assistant does not change visibility enforcement. The user’s existing permissions govern access, and AI will not bypass these restrictions for issue categories or types.
​

Yes. This is supported by configuring Required Custom Fields for the relevant issue type. The AI Ticket Assistant will enforce these fields during ticket creation (e.g., asking for justification for a URL unblock request) to capture essential details upfront.
​

Yes. Users can upload one or more images (e.g., a cracked screen) using the attachment icon. AI will attempt to describe the issue based on the image(s). Pasted images are not supported. Images are attached to the Ticket Activity/Timeline section of the ticket.
​

Yes, by referencing the other user's name, email, or identifier, provided the submitting user has the required permissions (List Users and Submit Tickets), and the target user is within their allowed location scope.
​

Only for themselves, unless explicitly granted “student-led help desk” permissions.
​

No. Quick Tickets follow the legacy ticket flow and do not involve the AI Assistant.

To prevent users from submitting multiple open tickets for the same issue, it automatically detects close matches to existing open tickets submitted by the same requester.
​

After the user confirms the ticket preview, but before a new ticket is created.
​

The system evaluates only open tickets from the same requester, using issue matching combined with contextual attributes (description, issue type, issue category, asset, model, and requester).

Ticket creation is paused, and the user is shown details of the existing ticket, including the ticket number, summary, linked ticket count, current status, and resolution summary (if available).
​

The user can choose:

No, the system does not automatically merge, close, or reassign duplicate tickets.

It provides guided troubleshooting steps for users to resolve simple issues themselves before the ticket reaches an agent.
​

The ticket is forwarded to an agent, who can view the attempted steps in the ticket's Timeline history.
​

Yes. Diagnostic steps and self-service actions are added as comments to the ticket Timeline so agents can see what was attempted.
​

Generally, no. AI uses the email subject, body, and other context to select the correct issue and associate the user/asset, reducing the need for extensive keyword-routing rules.
​

Yes. Rules always override AI. AI sets the initial ticket details, and then rules fire and can overwrite fields.
​

Yes. If the AI Ticket Assistant is used, assets referenced in the email will be automatically added to the ticket.
​

No. AI-created tickets do not replace or override existing rules. AI populates initial fields, but configured Email-to-Ticket rules always run afterward and override AI-selected values if applicable.
​

No. Existing rules continue to work without reconfiguration because AI processing is layered into the existing workflow.
​

Yes. AI can associate assets based on the sender’s email address and assets assigned to that user, even without asset tags or serial numbers in the email.
​

It creates a single ticket and associates only one asset (which may be selected at random if reliability is uncertain). It does not split the request. Users should submit separate emails/tickets for each asset for accurate tracking.
​

No. AI processing only applies to the initial inbound email. Replies are added as comments, but the AI will not reprocess the content to update ticket metadata fields (such as Location or Issue Type). For example, if a user replies with “The printer is in room 101,” the comment will appear on the ticket, but the Location/Room field will not automatically update.
​

Yes. It supports tickets from guest or unrecognized email addresses, consistent with existing Email-to-Ticket behavior.

Yes. The ticket will be generated "on behalf of" the other user, using the context provided in the email body.

CC’d users are added as Followers, but not as the ticket's Requestor.
​

No. BCC does not affect ticket creation or follower behavior.

Yes. AI evaluates the full ticket content for sensitive information (e.g., student IDs, personal data) and may flag the ticket. Existing site options, rules, and custom field configurations for sensitivity are still respected.
​

Currently, no. AI flags content for sensitivity but does not prompt users for confirmation.
​

The system includes moderation safeguards. Inappropriate or harmful prompts are flagged, and the user is given options to resubmit appropriately or submit a ticket manually.
​

Yes. Moderation blocks unsafe or disallowed content in prompts/images, such as weapons, violence, and explicit content.