AI assistance.
Clear boundaries.

On-demand ticket summaries, category suggestions, and public reply drafts. A company administrator must enable AI assistance first.

Nothing is sent to the requester automatically. Technicians review and edit drafts, and remain responsible for the final response and any applied category.

Generation context includes the ticket title, status, priority and version, up to 20 recent public messages with up to 2,000 characters each, and up to 200 active company categories.

Internal notes, attachments, People records, and asset records are excluded from this feature’s generation context. Public messages can still contain personal or confidential information.

Context is processed through Vercel AI Gateway and the selected model provider. Provider routing, retention, and training terms must be confirmed for your deployment; we do not make a blanket zero-retention or no-training promise here.

These are operational safeguards, not a monthly purchased allowance. Failed provider attempts count toward these short-term limits. When a limit is reached, further generations are blocked until earlier attempts leave the rolling window.

Provider availability and account funding can also affect whether a generation succeeds.

Commercial AI allowances are being validated before publication and enforcement. Confirm your pilot scope with our team; do not assume unlimited AI usage.

An administrator Usage page is planned for AI consumption, storage, and plan limits. It is not available yet. This page does not authorize or introduce automatic overage charges.

The AI feature stores usage metadata rather than prompt and generated-output text. A draft sent by a technician becomes a normal ticket reply; an applied category follows the existing ticket update workflow.

If ticket context changes during generation, the result is rejected so the technician can review the updated context. Generated suggestions can still be incorrect or incomplete.