> ## Documentation Index
> Fetch the complete documentation index at: https://docs.abbyy.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Document skills in Advanced Designer

> Build Document skills in Advanced Designer for structured, semi-structured, and unstructured documents — with notes on variants and training data.

A **Document skill** extracts field values from one type of document. You can build Document skills in either ABBYY Vantage (cloud) or Advanced Designer (Windows desktop).

Use Advanced Designer when you need to combine multiple Vantage technologies, add NLP, or branch the processing flow on document type — see [Use cases](/vantage/documentation/advanced-designer/use-cases/use-cases) for typical scenarios. For background on document categories, see [Extract data from documents in Advanced Designer](/vantage/documentation/advanced-designer/use-cases/extracting-from-docs).

## Document type variants

Documents of the same type usually share the same fields, validation rules, and structure, but variants differ in small ways — for example, by the year a tax form was issued. One Document skill can be trained on multiple variants. The technology you choose depends on how many variants you need to handle:

| Variants                 | Best fit                                                                                                                                                                                                                     |
| :----------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Up to \~10 (fixed forms) | [Forms activity](/vantage/documentation/advanced-designer/activities/forms/forms-activity) — see [Process structured documents in Advanced Designer](/vantage/documentation/advanced-designer/use-cases/process-structured). |
| Most essential variants  | [Fast Learning](/vantage/documentation/advanced-designer/activities/fast-learning) and/or [Extraction Rules](/vantage/documentation/advanced-designer/activities/extraction-rules/extraction-rules) activities.              |
| Hundreds                 | [Online Learning](/vantage/documentation/runtime/online-learning/online-learning) in Vantage refines the skill from manual-review feedback.                                                                                  |
| Thousands                | [Deep Learning activity](/vantage/documentation/advanced-designer/activities/deep-learning-semi-structured) extracts with \~80–90% accuracy depending on document complexity.                                                |

<Note>
  If a fixed form has many more than \~10 variants, treat each as a separate document type.
</Note>

## Training and testing a Document skill

For best results, train and test the skill with three different document sets:

* **Training set** — used to train the skill.
* **Test set** — used to measure accuracy during development.
* **Blind set** — an additional test set the skill has never seen, used to evaluate true generalization.

<Note>
  Use different documents in each set. Reusing training documents in the test set inflates accuracy estimates.
</Note>

### Training set

Aim for a representative set with 2–3 documents per variant. If you can't cover every variant, the [Deep Learning activity](/vantage/documentation/advanced-designer/activities/deep-learning-semi-structured) generalizes from image patterns and surrounding labels, so it can process variants it wasn't explicitly trained on.

Recommended document counts depend on the activities you use:

| Activity                                                                                                                         | High-variability documents         | Low-variability documents                 |
| :------------------------------------------------------------------------------------------------------------------------------- | :--------------------------------- | :---------------------------------------- |
| [Deep Learning for semi-structured documents](/vantage/documentation/advanced-designer/activities/deep-learning-semi-structured) | At least 200–300 (2–3 per variant) | At least 10 (2–3 per variant)             |
| [Segmentation](/vantage/documentation/advanced-designer/activities/segmentation/segmentation-activity)                           | At least 100                       | At least 20                               |
| [Deep Learning for NLP](/vantage/documentation/advanced-designer/activities/deep-learning-nlp)                                   | At least 150 (2–3 per variant)     | Can start with 1; aim for 2–3 per variant |

<Note>
  Even if you can't hit the recommended counts, one document per variant is better than none.
</Note>

### Test set

Match the test-set distribution to your production document flow so the accuracy estimate is meaningful. For example, if invoices from one vendor make up 30% of production traffic, the test set should contain about 30% of that vendor's invoices. The simplest way to hit this ratio is to test against random samples of production documents.

### Blind set

Use documents the skill has never seen during training or testing. The blind-set results are your best estimate of real-world quality.

## Configuring a Document skill

After you create a Document skill on the start page, configure it in this order:

<Steps>
  <Step title="Skill settings">
    Click the settings button next to the skill name to view and adjust [skill settings](/vantage/documentation/advanced-designer/document-skills/skill-settings).
  </Step>

  <Step title="Upload documents">
    On the **Documents** tab, [upload the documents](/vantage/documentation/advanced-designer/document-skills/documents) the skill will work with.
  </Step>

  <Step title="Define fields">
    On the **Fields** tab, [create the fields](/vantage/documentation/advanced-designer/document-skills/fields) you want to extract and label their locations on sample documents.
  </Step>

  <Step title="Configure activities">
    On the **Activities** tab, [build the document processing flow](/vantage/documentation/advanced-designer/document-skills/activities).
  </Step>

  <Step title="Test the skill">
    On the **Results** tab, [test the skill](/vantage/documentation/advanced-designer/document-skills/results) on sample documents and review extraction quality.
  </Step>

  <Step title="Publish">
    On the **Publish** tab, [publish the skill](/vantage/documentation/advanced-designer/document-skills/publish) to make it available in the [Skill Catalog](/vantage/documentation/skill-catalog/skill-catalog) in ABBYY Vantage.
  </Step>
</Steps>

After publishing, your skill appears alongside built-in skills, read-only skills, and any [derived skills](/vantage/documentation/advanced-designer/document-skills/derived) in the Skill Catalog.

## Next steps

<CardGroup cols={2}>
  <Card title="Skill settings" icon="sliders" href="/vantage/documentation/advanced-designer/document-skills/skill-settings">
    Configure recognition, training, and processing options.
  </Card>

  <Card title="Activities" icon="diagram-project" href="/vantage/documentation/advanced-designer/document-skills/activities">
    Choose and combine activities for the processing flow.
  </Card>

  <Card title="Derived skills" icon="code-branch" href="/vantage/documentation/advanced-designer/document-skills/derived">
    Build a new skill on top of a built-in or read-only Vantage skill.
  </Card>

  <Card title="Use cases" icon="lightbulb" href="/vantage/documentation/advanced-designer/use-cases/use-cases">
    See worked scenarios for common document types.
  </Card>
</CardGroup>