Create custom attributes using Analyzer in Sign

Last updated on Aug 17, 2026

Create custom attributes to extract information that is specific to your documents, organization, or business processes.

A custom attribute defines the information Analyzer should identify in a document. After you create and validate an attribute, you can use its extracted values for review, filtering, comparison, and export.

Create a custom attribute

From Home, select Create attributes for analysis.

Analyzer Home page showing the Create attributes for analysis tile.
Starting from Home opens the attribute workspace with the creation panel ready for configuration.

Complete the attribute definition:

Attribute name
Enter a clear, descriptive name. The name identifies the attribute throughout Analyzer.

Attribute group
Select the group where the attribute belongs.

You can create a group by entering a new name in the field. You cannot rename the group after saving an attribute to it.

Attribute mode
Choose how Analyzer determines the extracted value:

  • Strict — Extracts values that closely match the detailed attribute description.
  • Interpretive — Uses related document content and context to infer the value.

Attribute type
Choose the type of value the attribute returns:

  • Text — Alphanumeric values, phrases, or longer text.
  • Number — Numeric values.
  • Date — Values in a recognizable date format.

Document type
Select the type of document you expect the attribute to analyze.

You can create a document type by entering a new name in the field.

Detailed attribute description
Describe precisely what Analyzer should extract using clear, natural language.

The description must contain at least 30 characters.

Create custom attribute panel showing the fields used to define an attribute.
The attribute definition determines both what Analyzer looks for and how extracted values are returned.

Refine the definition with examples

Expand Advanced options when the definition needs additional guidance.

You can provide:

  • Inclusion examples — Examples of values that should be extracted.
  • Exclusion examples — Examples of values that should not be extracted.

Start with a clear attribute description before adding examples. Use advanced options when testing identifies a specific extraction problem rather than adding complexity during initial creation.

Test the attribute

Select Test to preview extraction results before saving the attribute.

Review the results against representative documents and refine the definition if necessary. Creating an effective attribute may require several test-and-refine cycles.

Testing uses the same extraction behavior as other attribute processing.

Save the attribute

When the test results meet your expectations, select Save.

After saving, the attribute becomes available in Analyzer.

Apply the attribute to one or more collections

Select one or more collections where you want to apply this attribute.

Analyzer processes extraction in the background.

You can skip this processing at this point and apply the attribute at a later time if you are creating multiple attributes and can't wait for the processing to complete.

Extract for all confirmation dialog showing that the attribute will be extracted across multiple documents.
Applying an attribute broadly can process many documents, so validate the definition before starting extraction across the full document set.

What happens during extraction

While Analyzer processes the attribute:

  • You cannot edit the attribute.
  • You cannot use the attribute in filters or exports.

When processing finishes, the extracted values become available for use throughout Analyzer.

Validate custom attributes

You are responsible for creating, testing, and validating custom attributes for your documents and use cases.

Extraction quality depends on factors such as:

  • The clarity of the attribute definition.
  • The consistency and quality of the source documents.
  • Whether the selected mode and value type match the information being extracted.
  • How thoroughly you tested the attribute against representative documents.

Refining an attribute can cause previously extracted values to be recalculated.