# Data annotation

URL: https://aidataforsale.com/glossary/data-annotation
Publisher: AI Data For Sale (Marketing Signals Ltd)
Updated: 2026-11-10

Adding labels, tags or notes to raw data, such as categories, boxes around objects or transcripts, so a model can learn from it.

## Annotation and labeling

The terms are used interchangeably. Annotation covers any added information; labeling usually means assigning a category or value. Both turn raw data into training examples.

## Common types

| Data | Annotation |
| --- | --- |
| Text | Categories, sentiment, entities, extracted fields |
| Images | Bounding boxes, segmentation, captions |
| Audio | Transcripts, speaker turns, sound events |
| Video | Actions, object tracks, step labels |
| Model outputs | Ratings and rankings ([preference data](https://aidataforsale.com/glossary/preference-data)) |

## Who does it

Annotation vendors with large workforces, specialist expert networks, in-house teams, and increasingly models with people checking samples.

## Labels in licensed data

Data that already carries labels from real work is often more valuable than data labeled after the fact. Resolution codes on tickets, outcomes on sales calls and the fields entered from invoices are all labels a buyer can train on. Keep them when you prepare data for licensing.

## Go deeper

- [Expert-annotated data](https://aidataforsale.com/data/expert-annotated-data)