Skills Data and integration
Writing derived data back into a helpdesk
Reverse ETL puts warehouse-computed attributes where agents can see them: a customer segment, a risk score, a lifetime value band, a churn flag. Done well it is one of the highest-leverage things a data team can do for a support operation.
npx rulebase-skills install cx-reverse-etlWhen to use it
Reach for this when someone says any of these — they are the phrases the skill itself triggers on:
- “push scores back into Zendesk”
- “sync warehouse attributes to the helpdesk”
- “write our churn risk onto the ticket”
How it works
The method, in the order the skill runs it. The full procedure — tables, worked examples and the edge cases — is in the skill itself.
Before designing anything
1. What decision does this attribute change? An attribute nobody acts on is clutter on an already-crowded ticket view, and clutter has a cost — agents stop reading the sidebar. If the answer is "it's useful context", ask who has asked for it.
Write derived, never authoritative
The hard rule that prevents the worst failures.
Make it explainable on the ticket
An agent looking at "Risk: 72" with no context will either ignore it or misuse it. Alongside the value, sync.
Bound the write path
Reverse ETL is a bulk mutation against a production system with real customer data. It carries the same obligations as any other write: An append-only audit log, Verify after applying by re-reading a sample.
Avoid the feedback loop
A subtle failure worth designing out: if a derived attribute changes how conversations are handled, and the model is retrained on the resulting data, the model learns from its own effects.
Related skills
Free and open source, and vendor-neutral — it reads the conversations from whichever helpdesk you already run. Browse all 149 skills · connect your helpdesk over MCP · source on GitHub
