Blaze Fish editorial team · Report a correction. Estimates and reported outcomes are not guarantees.
Auto-close thank-you replies and prioritize support tickets with ChatGPT
Allen Lai's customer experience team at Otter.ai used Zapier and ChatGPT to spot reopened Zendesk tickets that only said thanks and close them, and to tag new tickets by urgency and category. Over 1,000 tickets were auto-solved in three months and more than 10,000 were enriched and routed.
- Difficulty
- Beginner
- Estimated setup time
- 45 min
- Indicative monthly cost
- $20/mo
- Reported or estimated benefit
- several hours per agent per week
How to do it
- Create a Zap triggered when a help-desk ticket is reopened or gets a new customer comment.
- Send the latest comment to ChatGPT and ask whether it needs any reply.
- Add a filter: if the answer is no, mark the ticket solved and leave an internal note.
- Build a second Zap for new tickets that asks ChatGPT for urgency, sentiment and category and writes them as tags.
- Spot-check auto-closed tickets for the first two weeks before trusting it fully.
The prompt
Read this customer message from a support ticket. Answer with JSON only: {"needs_reply": true/false, "sentiment": "positive|neutral|negative", "urgency": "low|medium|high", "category": one of [billing, bug, how-to, account, other]}. A message that only says thanks or confirms the issue is fixed does not need a reply. Message: [comment]
Where this comes from
One person reported it. Promising, not yet confirmed. Original report: Zapier blog (vendor case study).
Tried this? Tell us how it went and help move its proof level up.