# Data Suitability Recipe Permissions

Source: https://userguide.qualytics.io/recipes/data-suitability/deep-dive/permissions/

An AI Use Case has no teams of its own. Every permission derives from the datastores it binds, and the requirement must hold on **all** of them: a use case is only as accessible as its weakest binding. On top of that, the platform role and, for Members, ownership of the record decide who can change it and who can sign it off. For the canonical matrices, see [User Roles](https://userguide.qualytics.io/settings/access/users/roles/overview/) and [Team Permissions](https://userguide.qualytics.io/settings/access/teams/team-permissions/overview/).

> **Note: Two Permission Systems**
>
> **User Roles** (Member, Manager, Admin) control what a user can do across the platform. **Team Permissions** (Reporter, Viewer, Drafter, Author, Editor) control what a user can do on the specific datastores they reach through team membership. They are independent systems with different scopes. For AI Use Cases, the team permission is checked on every bound datastore at once.

## User Roles (Platform-Level)

| Action | Member | Manager | Admin |
| --- | --- | --- | --- |
| Open the recipe and view the records you can access | Allowed | Allowed | Allowed |
| Create a use case | Allowed | Allowed | Allowed |
| Edit, delete (through the API), baseline, generate evidence, submit for review, or save recipe progress on a use case you created | Allowed | Allowed | Allowed |
| Do the same on a use case someone else created | Not allowed | Allowed | Allowed |
| Certify or revoke a use case | Not allowed | Allowed | Allowed |

Each row still requires the matching team permission on every bound datastore, listed below. Ownership is set when the record is created and never changes.

## Team Permissions (Datastore-Level)

The team permission must hold on **every** datastore the use case binds. Creating a record that binds no datastore yet, the usual shape of a fresh draft, still needs the **Author** team permission on at least one datastore somewhere.

| Action | Reporter | Viewer | Drafter | Author | Editor |
| --- | --- | --- | --- | --- | --- |
| View a use case, its baseline, its recipe run, and its attestations | Not allowed | Allowed | Allowed | Allowed | Allowed |
| Compare and export attestations | Not allowed | Allowed | Allowed | Allowed | Allowed |
| Create a use case and bind assets (Model, Inputs, Standard) | Not allowed | Not allowed | Not allowed | Allowed | Allowed |
| Edit or delete a use case; save recipe progress | Not allowed | Not allowed | Not allowed | Allowed | Allowed |
| Capture a baseline (Baseline) | Not allowed | Not allowed | Not allowed | Allowed | Allowed |
| Submit for review, or send back to Draft (Evidence) | Not allowed | Not allowed | Not allowed | Allowed | Allowed |
| Generate an attestation (Evidence) | Not allowed | Not allowed | Not allowed | Allowed | Allowed |
| Certify or revoke (Evidence), with the Manager role | Not allowed | Not allowed | Not allowed | Not allowed | Allowed |

> **Note: Two people, by design**
>
> Preparing evidence and signing it off are different jobs. An **Author** on every bound datastore builds the record and submits it for review. Certifying it takes the **Manager** role together with **Editor** on every bound datastore, or the **Admin** role. Being the author of a record never grants the ability to certify it, and the Manager role alone does not bypass the team permission.

## How Both Layers Work Together

To build a use case and generate its evidence, a user must satisfy all of these:

1. **User Role**: at least **Member**, and for a Member, ownership of the record
2. **Team Permission**: **Author** on every bound datastore

To certify or revoke, a user needs the **Manager** role plus **Editor** on every bound datastore, or the **Admin** role. Ownership plays no part in certification: writing a record does not grant the right to approve it, and an author who also meets these requirements is not blocked from approving their own record.

> **Info: Admin Bypass**
>
> Users with the **Admin** role bypass team permission checks and can reach, edit, and certify every AI Use Case.

## UI Behavior Without Permission

| Scenario | What the User Sees |
| --- | --- |
| User lacks Viewer on one of the bound datastores | The record does not appear in the record list and its page is not found. A partial view of a model's governed inputs is never shown. |
| User can view but not edit the record | **Edit in Recipe**, **Continue Recipe**, and **Adjust Model or Inputs** are not shown. On the Evidence step, no transition buttons appear and the step explains that authoring needs Author on every bound datastore and certifying needs a Manager with Editor, or an Admin. |
| User can edit but not certify | **Submit for Review** and **Send back to Draft** are offered; **Certify** and **Revoke** are not. On a record In Review, the step asks you to share the page with an approver. |
| Member working on someone else's record | The record is visible but treated as read-only, as above, even with Author on every bound datastore. |
| Creating a record without Author on any datastore | Creation is refused with a permission error, even when no datastore is bound yet. |

## Important Notes

- **Losing access to a bound datastore hides the record.** Losing Viewer on any one bound datastore hides the whole use case until access is restored.
- **Attestations follow the record.** Reading, comparing, and exporting them needs the same Viewer access as the record; there is no separate permission for evidence.
- **The recipe's autosave needs edit rights.** A user who can view but not edit a record can browse it in the recipe, but progress is not saved and the side panel footer reports the failed save.

## See Also

- **How It Works**

  The two phases, the Use Case picker, how progress is saved on the AI Use Case, and what the recipe writes.

  [How It Works](https://userguide.qualytics.io/recipes/data-suitability/deep-dive/how-it-works/)
- **Define the Use Case**

  Model, Inputs, and Standard: naming the model, binding its assets, and pinning the check set and expiry schedule.

  [Define the Use Case](https://userguide.qualytics.io/recipes/data-suitability/deep-dive/define-the-use-case/)
- **Capture the Evidence**

  Baseline, Evidence, and Complete: fixing the drift referent, certifying, generating the attestation, and exporting it.

  [Capture the Evidence](https://userguide.qualytics.io/recipes/data-suitability/deep-dive/capture-the-evidence/)
- **AI Use Cases**

  The governance record the recipe writes: certification lifecycle, who can do what, attestations, and audit obligations.

  [AI Use Cases](https://userguide.qualytics.io/recipes/data-suitability/deep-dive/ai-use-cases/)
- **Entry Points and Resuming**

  Get Started, Open in Recipe, Continue Recipe, Edit in Recipe, and how a run resumes from the AI Use Case.

  [Entry Points and Resuming](https://userguide.qualytics.io/recipes/data-suitability/deep-dive/entry-points/)
- **Examples**

  Real scenarios showing a model's inputs turned into a certified evidence record.

  [Examples](https://userguide.qualytics.io/recipes/data-suitability/deep-dive/examples/)
- **Best Practices**

  Guidelines for naming, binding scope, baselines, certification, and keeping evidence trustworthy.

  [Best Practices](https://userguide.qualytics.io/recipes/data-suitability/deep-dive/best-practices/)

*Last updated: October 7, 2026.*
