Test a macro applying text and field updates together

The checks that matter

The macro makes five changes at once, and the four invisible ones are the ones that affect the queue.

Asserting the before state on priority and type is what makes the change provable. Both start at values the macro is supposed to overwrite, so a macro that silently skips its field updates cannot pass by coincidence.

The flow as a test file

The reply text is asserted in full rather than by a fragment, because a macro that inserts a truncated or outdated template is a customer-facing error.

---
mode: testing
url: https://my-testing-repo-main.vercel.app/zendesk-clone-app/agent?reset=true
max_steps: 45
tags: [zendesk, support, productivity]
---

# Zendisk 30.3: Macro application

Catalog objective: apply a macro to a ticket.
Key assertion: the macro's fields and reply text are both applied.

## Open the ticket
Click "Open" on ticket 4412 and verify "Priority" reads "Urgent" and "Type" reads "Incident".

## Apply a macro
Select "Refund - approved" in "Apply a macro".

## Verify the reply text was inserted
Verify the "Public reply" box now contains "I've approved your refund and it will land back on your original payment method within 5 working days."

## Verify the macro's field updates were applied
Verify "Status" now reads "Pending", "Priority" reads "Normal", "Type" reads "Question" and the tags include "refund" and "approved".

Tags are asserted as well as fields, since they drive reporting and automation downstream and are the part of a macro most likely to be dropped.

Step by step coverage

A starting state, the macro, the text, and the fields.

Splitting the text and the fields into separate steps means a macro that does half its job fails on the specific half, which is exactly the information an agent lead needs.

Adding it to a suite

kane-cli testmd run tests/zendesk/macro-application_test.md

Add a macro applied to a ticket already in the target state, since a macro that errors rather than being idempotent interrupts an agent mid-queue.

Macros are edited by support leads rather than engineers. Deriving a suite from written rules covers keeping tests aligned, and running flows in CI covers catching an edit that breaks one.