# Classification (/docs/classification)

![Transcend Classification](/images/product/classification.svg)

## How Classification works [#how-classification-works]

Transcend Classification discovers schemas and samples content in connected data systems, then recommends data categories and processing purposes for what it finds. You configure and run this work as **scans** on each integration’s **Scan Configuration** tab.

### Scan types [#scan-types]

- **Discovery Scan** — Scans schemas (tables, collections, files) without classifying contents. Useful when you need schema coverage for DSRs or Preference Management.
- **Discovery + Classification Scan** — Discovers schemas and classifies sensitive contents. Recommended for systems that may hold personal data.
- **Classification Scan** — Re-samples and reclassifies without rediscovering schema. Available only after at least one Discovery or Discovery + Classification scan has completed.

### Ad-hoc vs recurring scans [#ad-hoc-vs-recurring-scans]

- **Run Ad-Hoc Scan** — Start a one-off scan from the Most Recent Scan widget to test configuration.
- **Recurring Scans** — Enable a schedule so Transcend keeps discovering and classifying as your systems change.

### Where to configure [#where-to-configure]

1. Go to **Infrastructure → Integrations** and open a connected data system.
2. Open the **Scan Configuration** tab.
3. Choose a scan type, set optional scope/sampling options, then run an ad-hoc scan or enable recurring scans.

For step-by-step setup, see [Classification Quickstart](./quickstart.mdx) and [Configure a Classification Scan](./scans/configure-classification-scans.mdx).

![Scan Configuration overview with Discovery + Classification selected](/images/sanity/2026-04-21-configure-scans-scan-type-options.png)

<FolderIndex />