Write an effective Zuora AI prompt

Before you write the prompt, confirm the following:

You have access to the Zuora AI experience in your tenant.
You know the identifiers, accounting periods, or filters that define the scope of your request.

Zuora AI interprets the request as written. A prompt that names the outcome, the records, and the scope produces a response you can validate. A broad prompt produces a broad response.

Your existing permissions and tenant settings continue to apply.

State the outcome you want. Name the deliverable: an explanation, comparison, report, export, or recommended next step.

Identify the business area and records involved. Name the objects, such as an account, subscription, invoice, payment, revenue contract, quote, workflow, or report.

Add scope that narrows the request. Include a date range, accounting period, currency, status, identifier, or filter. For a comparison, name both sides of the comparison and the period or condition to compare.

Specify the output format. Request a table, CSV-ready output, executive summary, or step-by-step explanation.

Ask Zuora AI to state its assumptions. Request that Zuora AI identify missing information and list the checks to complete before you act.

Split a complex request into a sequence of prompts. Ask Zuora AI to investigate in the first prompt, validate the findings in the second, and prepare an export or proposed action in the third.

Review the response before you act. Validate the findings against your source records. Review any proposed change before you approve it.

Zuora AI returns a scoped response in the format you requested, together with its assumptions and the items that require validation.

Refine the prompt if the response is broader or narrower than you need. For context-aware prompt suggestions, use the suggestions available in the Zuora AI experience.