Docs
AI & BYOK

AI Feature Toggles and Spend Caps

Enable or disable individual AI features and set budget limits.

Each AI feature in Ticket0 can be turned on or off independently. Go to AI Agent → Features to manage them.

Available toggles

FeatureDefaultDescription
Auto-triageOnClassifies each inbound ticket (topic, urgency, sentiment, escalation flag, churn risk). The signals drive inbox filtering, routing rules, and the data shown on each ticket.
Draft repliesOnGenerates a suggested reply for each ticket during the processing window and pre-fills it in the composer when the operator opens the ticket.
Auto-replyOffWhen enabled, the AI sends its own replies without operator review — but only when the confidence score meets the threshold below. Below threshold, the draft is still produced and waits for the operator. Starts off by design: admins opt in once they've reviewed enough drafts to trust the quality.
Auto-reply threshold85The confidence score (0–100) the draft must hit for auto-reply to fire. Only applies when Auto-reply is on. Tune per-category too — see Confidence thresholds.
Similar ticketsOnSearches past resolved tickets by semantic similarity and shows matches in the right sidebar of each ticket. Used for pattern matching and as a signal into draft quality.
TranslationOffDetects the customer's language and translates inbound messages + outbound drafts so operators can work in their own language. Off by default — turn on if you handle multi-language support.
Knowledge retrievalOnGrounds AI drafts in your knowledge-base articles. The retrieved articles are passed to the draft prompt as context.

Spend caps

Set a workspace-wide monthly cap to prevent runaway AI costs:

  1. Go to AI Agent → Spend
  2. Enter a USD amount
  3. Choose an alert threshold (default: 80%)

When the cap is reached, AI features pause for the remainder of the billing month. Tickets continue to arrive and agents can reply manually.

Spend caps apply to BYOK usage tracked by Ticket0. If you set limits directly in OpenRouter or your provider's console, those limits apply independently.

On this page