# Governance and Lifecycle

Athenic supports a governed workflow so teams can move fast with AI-generated definitions while still keeping a clean, consistent metric and dimension layer.

<div data-with-frame="true"><figure><img src="https://2503620627-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FVqTQOKxSAPs7Xr7cb4Ms%2Fuploads%2FH8L5mD5zVgZ22XzVwZNK%2FGoverned%20Metrics.png?alt=media&#x26;token=8050852b-e717-4dca-bf03-d381b4307001" alt=""><figcaption><p>The Governed Metrics page for team admins can be found in the AI Analyst setup </p></figcaption></figure></div>

### Admin Approval Page

Admins manage advanced metrics and advanced dimensions from **Governed Metrics**:

* **Advanced Metrics** tab: create, review, edit, and manage advanced metrics
* **Advanced Dimensions** tab: create, review, edit, and manage advanced dimensions

Items that are pending review will appear **in orange** and are labeled **Quarantined**.

### Quarantined vs. Governed

**Quarantined**

* Created by users via **Generate with AI**
* Highlighted **in orange** to indicate they are not yet fully governed
* Intended for review to ensure correct definitions, naming, and consistency

**Governed (Approved)**

* Standard metrics/dimensions used across Athenic
* Serve as the canonical definitions for reporting and decision-making

#### Lifecycle (end-to-end)

1. A user requests either a new metric or advanced dimension while editing an AI proposal.
2. Athenic **generates a draft** and, if confirmed by the user, marks it as quarantined for your team admin to review.
3. The team admin reviews it in the **Governed Metrics** page (Advanced Metrics / Advanced Dimensions) of the AI Analyst setup.
4. After admin approval, the item becomes a **normal governed** metric or dimension and will appear alongside the rest.

#### When to create new metrics or advanced dimensions

Create a **new metric** when:

* The calculation should be reused often (e.g., “Gross Margin”, “Conversion Rate”)
* It needs a shared, consistent definition (filters, inclusions/exclusions)
* You want it selectable by name in AI and dashboards without rewriting logic

Create an **advanced dimension** when:

* Users repeatedly need a **derived grouping** that isn’t stored as a column
  * CASE WHEN revenue < 100 THEN '0-100' WHEN revenue < 500 THEN '100-500' WHEN revenue < 1000 THEN '500-1000' ELSE '1000+' END
* As an admin, you can always reject the advanced dimension and add your own version by editing the dataset instead.

Avoid creating new governed items for:

* One-off explorations that won’t be reused
* Highly specific, temporary filters better expressed directly in a question

<br>