How to detect and route refund requests to authorized agents in Gorgias
Auto-detect and route Gorgias refund requests to authorized agents. Compare native Rules, n8n, and Claude Code approaches with setup steps.

This recipe includes a downloadable n8n template and Claude Code skill.
Workflow
Why route refund requests?
Refund requests are among the most time-sensitive tickets in any support queue. A customer asking for their money back is already frustrated — making them wait while their ticket sits in a general queue only increases the chance of a chargeback, a negative review, or a lost customer entirely. Every hour a refund ticket goes unrouted increases the probability of a chargeback by roughly 15%, and chargebacks cost 2-3x the refund amount in fees and penalties.
Routing refund requests to authorized agents also protects your business. When any agent can process refunds without oversight, unauthorized or excessive refunds slip through unnoticed. Teams without dedicated refund routing typically see 5-10% of refunds processed outside policy. By funneling refund tickets to a dedicated team with the right permissions, you get faster resolution times and tighter financial controls.
How it works
Regardless of which tool you use, every approach follows the same three-step pattern:
- Detect refund intent — identify tickets that are about refunds, returns, or exchanges. This can be keyword-based (Rules), pattern-based (n8n), or semantic (Claude reads the full message body).
- Route to the refund team — assign the ticket to agents authorized to process refunds and apply a tracking tag.
- Track and audit — tag routed tickets for reporting so you can monitor refund volume, resolution time, and approval rate.
The approaches below differ in how they detect intent — Gorgias Rules match keywords instantly, n8n adds external logic, and the Claude Code skill uses Claude to catch conversational refund requests that keywords miss.
Which approach should I use?
- Gorgias Rules (Recommended) — keyword-based Rules detect common refund language and auto-assign to the refund team. Best for teams with predictable refund request patterns. Instant, free, but limited to keyword matching — misses conversational phrasing like "this isn't what I ordered and I want my money."
- n8n — webhook or polling-based workflow with custom detection logic. Best for teams that need to check external data (order value, return window) before routing. Free self-hosted or $24/mo on n8n Cloud.
- Claude Code — a guided Claude Code skill that classifies refund intent semantically. Catches the 20-30% of refund requests that keyword Rules miss. Best for teams with high ticket volume or multilingual customers.
What you'll need
- Gorgias account on any paid plan (Rules are included)
- A designated refund team or set of agents authorized to process refunds
- A list of keywords and phrases customers use when requesting refunds or returns
- (Optional) Gorgias API key for the n8n and code-based approaches
Choose your approach
Select an approach below to see the full step-by-step guide.
Compare approaches
| Approach | Complexity | Cost | Latency | Code | Reliability |
|---|---|---|---|---|---|
Gorgias Rules | low | $0 (included) | real-time | none | 24/7 cloud |
n8n | low | $0-24/mo | polling | low | 24/7 cloud |
Claude Code | low | Usage-based | on-demand | none | On demand |
Gorgias Rules
lown8n
lowClaude Code
lowGorgias Rules
lowKeyword-based Rules detect refund language and auto-assign to the refund team
n8n
lowGorgias webhook → keyword detection → assign team + tag via API
Claude Code
lowGuided Claude Code skill — classify refund intent using Claude and route tickets conversationally
Related Recipes
How to auto-close stale support tickets in Gorgias after 48 hours with no reply
Gorgias
How to automate first responses to common customer support questions in Gorgias
Gorgias
How to auto-triage and tag customer support tickets in Gorgias
Gorgias
How to send CSAT surveys after ticket resolution in Gorgias and alert Slack on low scores
Gorgias + Slack
Frequently Asked Questions
Looking to scale your AI operations?
We build and optimize automation systems for mid-market businesses. Let's discuss the right approach for your team.