# Action Map — Rate Guard
> Kit 02 · Rate Guard · deliverable · designed and built by I Wayan Usada — instructional design portfolio · fictional composite

## The goal (proposed, not measured)

In launch month, each audience answers its own Rate Guard questions: tickets fall week on week, escalations arrive complete, and a question that appears twice gets an article.

## What people do instead

Escalate anything unfamiliar; answer from memory in a private reply; close anything that looks like "working as designed"; demo the feature with promises the product does not keep.

## The actions that close the gap

| # | Behaviour | What it looks like | Where the artifact trains it |
|---|---|---|---|
| 1 | **Search the KB before escalating** | a search with words from the ticket; the article linked in the reply | practice queue step 1 (KB point) |
| 2 | **Log a gap when nothing answers** | zero-result question logged, not improvised | white-label tickets t7 → t9; gap radar in the KB |
| 3 | **Route through triage** | the questions answered against the record, not the customer's framing | practice queue step 2 (route points); critical ticket t3 |
| 4 | **Escalate with everything attached** | ID, screenshot, one sentence of impact | practice queue step 3 (discipline point) |
| 5 | **Demo without over-promising** | arm before send; window, revision and toll answers right | demo rehearsal in the tour (/4) |

## What is not a training problem

- A revision screen that does not say "this ends protection".
- Partner questions in languages the KB does not have.
- A ticketing form that lets escalations go out without IDs.

## Revision note

The day-30 "≥70% self-service" goal was stated as a pre-committed target with no baseline; it is now a proposal. The earlier map pointed at a troubleshoot widget with no tickets; the queue now practises each action. Nothing has been measured.