Support Ticket Triage
Ticket updated in help desk with: category, priority, assignee, SLA timer started, internal classification note, customer acknowledgment sent. Slack notification for P1/P2. Weekly triage report: volume by category, priority distribution, SLA compliance rate.
Before you start
- Ticket categories and priority definitions documented
- Routing rules: which team/person handles which category
- SLA targets defined per priority level
- Incoming support ticket (subject, body, customer email, attachments)
- Customer account data (plan tier, lifetime value, open tickets count)
- Current team availability and queue depths
The steps
- Parse ticket content and extract intent — Read subject and body. Identify primary intent: bug report, feature request, billing question, how-to question, account issue, or complaint. If multiple intents detected, identify the primary (most urgent) and note secondary. Extract any error codes, URLs, or screenshots mentioned.
- Look up customer context — Pull customer record by email: plan tier, monthly spend, account age, number of open tickets, last interaction date, assigned CSM. Flag if customer is enterprise tier, has churned before, or has 3+ open tickets already. This context affects priority scoring.
- Assign priority and category — Priority matrix: P1 (service down, revenue impact, enterprise customer blocked) — respond in 1h. P2 (feature broken, workaround exists) — respond in 4h. P3 (how-to, minor bug, feature request) — respond in 24h. P4 (feedback, enhancement idea) — respond in 48h. Apply customer tier multiplier: enterprise bumps priority by 1 level.
- Route to correct team or agent — Apply routing rules: billing issues to billing team, technical bugs to engineering support, account issues to CSM, feature requests to product inbox. Within team, assign to agent with lowest current queue unless ticket requires specific expertise. For P1: assign AND notify team lead in Slack immediately.
- Send acknowledgment and set expectations — Send auto-reply confirming receipt. Include: ticket number, priority level, expected response time based on SLA, and self-service resources if applicable (link to relevant docs or status page for outages). For P1: include direct escalation path if not resolved within SLA.
- Update ticket metadata and log — Set ticket fields: category, priority, assigned agent, SLA deadline, customer tier, intent tags. Add internal note with classification reasoning. If this is a duplicate of a known issue, link to existing ticket and notify customer of existing workaround.
If it goes wrong
Misclassified priority (P3 was actually P1)
Add keyword triggers for critical terms (down, broken, can't access, blocking). Require human review for tickets from enterprise accounts until accuracy > 95%.
Wrong team assignment causing ticket ping-pong
Review routing rules monthly. Add 'uncertain' category that routes to a triage human instead of guessing. Track reassignment rate as a quality metric.
Duplicate detection triggers false positives
Require both subject similarity AND same customer account for auto-merge. Otherwise, flag as 'possible duplicate' for human confirmation.
All OpenLabor playbooks