This FAQ answers common questions about iManage AI Enrichment, including how it classifies and enriches documents, which file types and sizes are supported, how you can tell when a document has been enriched, how enriched metadata behaves as documents change, and how enriched properties can be searched and filtered. Use it to quickly find answers to specific questions, or refer to the topics in the guide for full details on any of these areas.

In this section:

Overview and purpose

This section covers what iManage AI Enrichment does, the benefits it offers, and how the underlying AI models are trained.

What does iManage AI Enrichment do?

iManage AI Enrichment runs a categorization service that adds one of three appropriate categories to supported document types: Agreement, Court Document, and Undetermined. For documents categorized as an Agreement, iManage AI Enrichment next identifies the agreement’s specific sub-type, then analyzes the document content and, where applicable, populates the following fields:

• Agreement Date

• Parties

• Heading

• Jurisdiction

• Governing Law

For more details, refer to What does iManage AI Enrichment do?

What is agreement sub-type identification?

For documents classified as Agreement, iManage AI Enrichment identifies the agreement’s specific sub-type (66 supported sub-classes) before extracting metadata. This step runs automatically and doesn’t produce a separate visible property.

What benefit does iManage AI Enrichment offer users?

iManage AI Enrichment adds valuable metadata to new content as it is generated in real time, so that users can more easily search for and filter to find content.

Is customer data used to train the AI models?

No customer data is used to train iManage AI Enrichment models.

Eligibility and identification

This section explains how to identify which metadata properties have been populated by iManage AI Enrichment.

How do I identify AI-enriched properties?

The AI-enriched properties are automatically populated by the system and are marked with an AI icon. This visual indicator remains until a user manually edits any of the enriched metadata properties; this causes the AI icon to be removed from all the fields. This ensures that the metadata is recognized as enriched and allows the system to continue updating it as needed.

For more information, refer to How to know when iManage AI Enrichment is active.

Supported formats

This section lists the file types iManage AI Enrichment supports and the criteria a document must meet to qualify for enrichment.

What types of documents does AI Enrichment support?

To be enriched, documents must meet the following criteria:

• Supported file types:

  • .doc, .docm, .docx

  • .dot, .dotm, .dotx

  • .odt, .ott

  • .pdf

  • .rtf

• File size limit: There's no file size limit for documents.

• Minimum document size: There's no minimum document size, but documents with fewer than 100 characters won't be enriched.

• Maximum document size: 1 million characters. Content beyond this limit is skipped and excluded from metadata enrichment.

• Document language: Delivers accurate results for documents with at least 80% English-language content. Metadata values, such as Agreement Date or Heading, aren't enriched if they're in other languages.

Enrichment behavior

This section answers questions about when and how the enrichment service runs, including what triggers it and how to control it manually.

Does AI enrichment run against all versions of supported document types?

Yes. After AI enrichment is enabled, it runs against all new versions created. For more information, refer to What actions trigger the AI enrichment service?

For versions created before AI enrichment was enabled, refer to How can historical documents be enriched?

What actions trigger the AI enrichment service?

The following actions cause a document version to be classified and enriched:

• A document has its content modified and is then saved back to iManage Work. After it is checked back in, enrichment will run against the modified document.

• A new version is created (including by using Promote previous edit to latest version).

• A document is copied. None of the enriched metadata properties are copied to the new document. Enrichment runs against the copy as a new document once it’s saved.

• A new document is created.

• A Restore from Trash, for a single version or the entire document.

These actions also ensure that Insight+ receives updated metadata. If enrichment is disabled or skipped, Insight+ may not reflect the latest document changes.

What actions don’t trigger the AI enrichment service to run?

The following actions won’t trigger the enrichment service to re-run:

• Changes to the document Security.

• Changes to the document Profile.

Why are some emails classified while others aren't?

Native emails (such as .msg or .eml) aren't classified. However, when an email is saved as a PDF and added to iManage Work, it is treated as a document. In this case, the document is displayed under Email, reflecting how the PDF is stored and displayed.

Can I manually select a document version and ask that it be enriched?

Yes. You can use the Enrichment Disabled field in the Properties panel to control whether and when enrichment runs for a specific document version.

To allow enrichment to run again, update the Enrichment Disabled field for that version (for example, clear the field or set it to No, depending on when you want enrichment to run).

For full details, refer to Manually editing an enriched property.

Can I exclude specific documents from being enriched?

Enrichment is designed to be a background process. However, manually setting the Enrichment Disabled property to YES in the Properties panel of a document prevents enrichment from running against this document version.

How long does it take to enrich a document version?

This depends on the size of the document and the complexity of its content, though enrichment of newly uploaded or edited documents typically completes near instantaneously. (Retroactively enriching an entire existing document corpus is a separate process. For more information, refer to How can historical documents be enriched?)

AI enrichment fields in iManage Work Web

This section covers how enriched metadata properties appear, behave, and can be used within iManage Work Web.

Visibility and auditability

These questions cover how to confirm whether and how a document has been enriched.

How can you tell if a document has been enriched?

Documents and their versions that have been enriched display values in the enriched properties of the document Properties panel, and in Grid or List view (if enabled by your administrator).

AI-enriched properties are automatically populated by the system and are marked with an AI icon.

The following three fields (if exposed by your administrator) can also tell you if a document version has been enriched, even if no metadata was added to the AI enrichment fields:

• Enrichment Date

• Enrichment Status Code

• Enrichment Version

For more information, refer to How to know when iManage AI Enrichment is active.

Is there a way to tell which version of iManage AI Enrichment has been run against a given document?

Yes. Administrators can include the Enrichment Version field in the document Properties panel in iManage Work Web. When enabled, it shows which version of the AI enrichment app processed the document. iManage Work increases the version number when significant updates are deployed (for example, a new model or retraining to improve accuracy).

Are there Timeline entries for the AI enrichment service?

There are no iManage Work Timeline entries for the enrichment service itself. If a user modifies an enriched field, this will be captured in the timeline using the standard “changed profile of version X” text.

Field availability and editing

These questions cover how enriched properties are displayed and edited in iManage Work Web.

Are the enriched properties available in iManage Work Web?

Yes. If configured by an administrator in iManage Control Center, the enriched properties can be displayed on the Properties panel for documents in iManage Work Web.

Can users edit the enriched property values in iManage Work Web?

Yes. The enriched properties Category, Agreement Date, Parties, Governing Law, Heading, and Jurisdiction are editable by any user with read/write access to a document.

What happens if I remove an enrichment value?

If you remove an enrichment value, the Enrichment Disabled property is automatically set to Yes for that document version. This prevents further automatic enrichment for that version and, by default, subsequent versions. When enrichment is disabled, the AI icon is not displayed in the Properties panel, indicating that the properties are no longer being AI-managed.

To allow enrichment again, a user must manually change Enrichment Disabled back to Not set (cleared) or No. For details, refer to Manually editing an enriched property.

Behavioral questions

These questions address common questions about changes to enriched property values.

Why did the enrichment properties on my document change?

The iManage AI Enrichment service automatically updates metadata properties when a document is first saved to iManage Work, modified and then closed, a new version is uploaded, or when an administrator triggers a reprocessing event. These updates ensure the metadata reflects the most current document content and classification standards. You can see when enrichment last occurred by checking the Enrichment Date property in the document Properties panel.

Why are some enriched properties empty?

Enriched properties are only populated if the AI model can confidently extract the information.

Searching and filtering

These questions cover using enriched properties in search and filtering.

Can the enriched properties be used to search for content in iManage Work Web?

Yes, if configured by an administrator in iManage Control Center, the following properties are available to search in Advanced Search in the iManage Work Web:

• Category

• Agreement Date

• Parties

• Governing Law

• Heading

• Jurisdiction

• Enrichment Date

• Enrichment Status Code

• Enrichment Version

For more information, refer to Searching and filtering with enriched metadata.

Can the enriched properties be used to filter content in iManage Work Web?

If configured by an administrator in iManage Control Center, the following properties can be applied as filters in iManage Work Web:

• Category

• Agreement Date

• Governing Law

• Jurisdiction

For more information, refer to Searching and filtering with enriched metadata.

Special scenarios

This section addresses less common scenarios, including historical documents, co-authoring, and how specific properties behave in search and filtering.

How can historical documents be enriched?

A request can be raised to iManage by an administrator or your IT team, and a job will be scheduled and communicated accordingly. Enrichment of historical documents is included with your iManage AI Enrichment subscription.

How does enrichment work when using next-generation co-authoring, or exclusive checkout with Save & Sync in iManage Work?

Any interim changes to a document won't cause enrichment to run. Only after the document has been saved back to iManage Work and is showing as checked in will the enrichment service run against this document version to look for new metadata.

Why does iManage AI Enrichment extract detailed metadata only from agreements?

To date, we’ve focused on enriching agreements because they are widely used across our customer base.

Why is the Jurisdiction property hierarchical, but the Governing Law property is flat?

Jurisdiction uses a hierarchical taxonomy because there is a clear parent-child relationship between legal authorities. For example, cases heard in the courts of England and Wales may ultimately be appealed to the Supreme Court of the United Kingdom. Likewise, the courts of Scotland and Northern Ireland sit alongside England and Wales as siblings within the broader UK judicial structure. This layered authority model supports hierarchical reasoning, faceted search, and inheritance logic (for example, “all UK appellate courts”).

By contrast, Governing Law is not hierarchical. It represents the body of substantive law chosen to govern a legal relationship and can be selected independently of geography or court system. For example, “the law of England” functions as a complete legal system in its own right and is not meaningfully nested under a broader category such as “UK law.”

The region is included in the name of the governing law dictionary entry (for example, “United Kingdom” in “United Kingdom | England & Wales”) to improve findability through search and browsing, and to clearly distinguish similarly named jurisdictions such as “Georgia” the country and “Georgia” the American state.

When I search or filter by a value, are documents tagged with child values also included?

It depends on the type of metadata property:

• Hierarchical taxonomy properties (for example, Jurisdiction): Selecting a parent value returns documents tagged with that value and any of its child values.

• Non-hierarchical classification or lookup properties (for example, Category and Governing Law): Selecting a value returns only documents assigned to that exact value. These properties don't support child values.

• Text, date, and Boolean properties (for example, Parties, Heading, Agreement Date, and Enrichment Disabled): These properties don't use hierarchies. Search results are based solely on direct value matching or filtering criteria.