Skills CX operations
Behavioural customer effort
CES asks customers how hard something felt. The answers arrive late, from a biased sample, and conflate product pain with support pain. Behavioural effort uses what they already did: repeats, switches, transfers, reopens, and time-to-resolution.
npx rulebase-skills install cx-effort-scoreWhen to use it
Reach for this when someone says any of these — they are the phrases the skill itself triggers on:
- “customer effort score”
- “behavioural CES”
- “our CES doesn't match operational data”
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.
Signals that indicate effort
Each signal should be observable in ticket/contact data with stable definitions.
CES survey traps (when you still have survey CES)
If behavioural effort rises while CES falls, trust the behaviour until survey solicitation, timing, and response bias are audited.
Constructing a composite (without fake precision)
Step 1 — Binary flags per episode. For each resolved issue episode, mark whether each signal exceeded a threshold (e.g. ≥2 transfers, reopen within 7 days, ≥2 channels). Thresholds should be set from distribution elbows on your own data, not industry benchmarks.
Data requirements
Without episode linking, every ticket looks like low effort.
Traps
Counting transfers as skill failure. Some drivers require specialist routing; measure avoidable transfers (return to previous queue, wrong team) via taxonomy.
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
