# What Data And Access You Need Before QA Setup

## Definition

This checklist identifies the access, Zendesk data, service categories, and quality decisions needed for a first ZenOps QA workflow.

## Context

Use it before building a scorecard or enabling automated evaluation. You can prepare the standard and review process with a small pilot category before expanding QA coverage.

## Attributes

- **QA product and permissions:** Confirm the workspace has QA access. Identify a person with **Quality: Manage** for scorecards and evaluation setup, an administrator who can manage service categories and user access, and the reviewers or coaches who will use the results. Open [QA Settings](https://wfm.zenops.cx/settings/service/qa-analysis/general) and see [Roles And Access In ZenOps QA](/help/articles/qa-roles-and-access-in-zenops-qa/) for permission details.
- **Zendesk connection and conversations:** Confirm the intended support account is connected in [Zendesk integration settings](https://wfm.zenops.cx/settings/integrations/zendesk) and that representative tickets are available. For an interaction to be graded, it needs a public customer message, a public staff response, and an attributable Zendesk staff identity.
- **Service category:** Choose an enabled category that describes the pilot work. Review the taxonomy in [Categories](https://wfm.zenops.cx/service/categories). A category needs a published, assigned scorecard before automated QA can run. Decide how the tickets will receive that category and check a representative example.
- **Quality standard:** Agree on the behaviors to assess, the evidence reviewers expect, any critical checks, and who approves changes to the scorecard. Build the standard in [QA Settings · Scorecards](https://wfm.zenops.cx/settings/service/qa-analysis/scorecards) and publish the first scorecard version before enabling the category's QA policy.
- **Evaluation scope:** Choose eligible channels and any narrower filters, how much qualifying work to select, and when a selected interaction should be evaluated. Start with a scope small enough for reviewers to inspect individual results.
- **Agent identity and feedback:** If the pilot includes agent results, confirm Zendesk staff can be linked to the right workforce people. Decide who may publish results, coach agents, and resolve appeals. Set expectations for when agents will receive feedback and how they can respond.
- **Pilot outcome:** Pick a measurable first result, such as reviewing a set of graded conversations, publishing feedback to a small agent group, or completing one coaching cycle.

An enabled policy does not retroactively grade older conversations. Use a ticket test or a bounded batch when you need to check existing tickets during setup. The [QA workflow overview](/help/articles/qa-how-scorecards-evaluations-reviews-and-coaching-fit-together/) explains how the prepared inputs are used.

> Canonical page: https://zenops.cx/help/articles/qa-what-data-and-access-you-need-before-qa-setup/
> Audience: For evaluators and administrators
> Summary: Prepare Zendesk conversations, service categories, scorecards, access, and a first QA pilot.