Work with attributes in Adobe Acrobat Analyzer

Use attributes to turn large sets of files into clear, comparable data you can act on.

Attributes help you turn the details you care about across a set of files into values you can actually work with, scan, filter, and export. Instead of opening files one at a time to hunt for the same information, you define an attribute once and Adobe Acrobat Analyzer extracts that value across the collection. Each extracted value includes an attribution link (citation) so you can jump to the source in the file and spot-check what was found. As your needs evolve, you can refine the attribute definition and refresh the collection to apply the updated logic. And if a single file needs a correction, you can override the extracted value for that file, then re-extract later to remove the override and return to the extracted result.

There are two types of attributes:

• Standard attributes
  Predefined and available out of the box, these attributes help you get started quickly and provide examples of how Acrobat Analyzer extracts common contract terms.
• Custom attributes
  Defined by you to capture information unique to your organization, workflows, or file types.

Every extracted value includes an attribution link, allowing you to verify the source directly within the file.

Why attributes matter

• Move faster through large file sets by surfacing key values without reading every file end-to-end.
• Validate results quickly using attribution links that show where each value appears in the file.
• Work at scale by filtering, exporting, and comparing extracted values across collections.
• Capture organization-specific concepts by defining custom attributes that match your workflows and terminology.

Important guidance

• The standard attributes included with Acrobat Analyzer are designed to help you get started quickly and understand how attribute extraction works across general file types. These attributes have been developed and rigorously tested using a broad range of data sources, including Adobe’s own files and language, supported by ground truth from our product, legal, and expert teams. They have also been cross-validated against publicly available datasets and industry benchmarks such as CUAD. 
• Definitions for specific terms can vary by organization. For the highest accuracy, create custom attributes tailored to your files.
• You are responsible for creating, testing, and validating custom attributes tailored to your specific files and use cases. Acrobat Analyzer provides tools for evaluation and feedback; however, accuracy depends on your input and subsequent refinement.

How attributes appear in Acrobat Analyzer

When you open a file, attributes display in the Document Attributes panel. You can:

• Review extracted values.
• Jump to source passages using citation links.

Creating and improving custom attributes

You can create attributes by defining the name, type (text, number, date), and description. Start simple, test on a small set of files, and refine the definition as needed.

Helpful practices include:

• Review the tips for creating a successful attribute.
• Use the attributes walk-through to step through the process for building attributes.
• Create custom attributes with well-thought-out descriptions and values, as if explaining the concept to a new colleague.
• Evaluate your attribute to test and refine its accuracy.
• Review results regularly and adjust when values are incorrect.
• Validate the attribute to hone it before scaling up its use.
• Provide concise positive or negative examples if needed (3–5 examples recommended).
• Begin with a small sample of files before expanding to your whole dataset.

Managing and maintaining your attributes

Once you’ve created an attribute and you’re getting useful results, these tasks help you keep values reliable, correct edge cases, and move data where you need it.

• Refresh a collection’s attributes when you add new attributes or update an attribute definition, and you want the latest logic applied to a collection.
• Manually override an attribute’s value when a particular file is an outlier, and you need the corrected value to drive filtering, review, or comparison without changing the attribute definition.
• Re-extract an attribute’s value when you want to remove an override and return to the extracted result for that file.
• Export attribute values when you need to share results, archive a point-in-time snapshot, or use the data outside Acrobat Analyzer.
• Download attribute definitions when you want a backup of your definitions, or you’re standardizing a set of attributes across projects.
• Upload attribute definitions when you want to reuse a known-good set of definitions instead of recreating attributes one by one.
• Disable an attribute when an attribute is no longer relevant, or you want to stop using it while you rework the definition.