Use attributes to extract consistent, comparable information from documents so you can review, filter, compare, and export the results.
An attribute defines a value you want Analyzer to identify across your documents. Instead of opening each file and locating the same information manually, define the information once and let Analyzer extract it from the files you process.
Each extracted value includes attribution to the source content so you can review where the result came from and validate it against the document.
Understand attribute types
Analyzer provides two kinds of attributes:
Predefined attributes
Ready-to-use attributes designed to extract commonly needed information. These are provided by the system and cannot be deleted.
Custom attributes
Attributes you define for information specific to your organization, documents, or workflows.
Use predefined attributes when they match the information you need. Create custom attributes when your documents require terminology, logic, or values specific to your use case.
How attributes work
For each attribute, Analyzer uses its definition to extract a value from a document.
Depending on the attribute, the extracted value can be text, a number, or a date.
After extraction, you can use attribute values to:
- Review information across documents.
- Filter documents.
- Compare values between documents.
- Export extracted data.
- Identify values that need further validation or correction.
Verify extracted values
Extracted values include attribution to the source document.
Use the attribution to review the supporting content before relying on a result for important decisions.
Extraction quality depends on the attribute definition and the documents being analyzed. Custom attributes may require testing and refinement before you apply them broadly.
Build effective custom attributes
Start with a clear definition of the information you want to extract.
When creating a custom attribute:
- Use a specific, descriptive name.
- Clearly describe the value Analyzer should identify.
- Test the attribute against representative documents.
- Review both correct and incorrect results.
- Refine the definition when results are inconsistent.
- Validate the attribute before applying it broadly.
For more complex requirements, examples can help clarify the results you expect.
Maintain extracted data
Attribute definitions and business requirements can change over time.
You can:
- Refresh attributes to apply current definitions to documents.
- Override an extracted value when a specific document needs a correction.
- Re-extract a value to remove an override and return to the extracted result.
- Export extracted values for use outside Analyzer.
- Download and upload attribute definitions to reuse them.
- Disable attributes you no longer need.
Use attributes responsibly
Validate extracted information before using it for decisions that require a high degree of accuracy.
Your organization is responsible for determining whether Analyzer is appropriate for the documents and use cases you process, including content subject to legal, regulatory, privacy, or internal compliance requirements.