What are Smart Tags?
Smart Tags let you automatically tag and categorize patent documents using natural-language descriptions. Instead of manually labeling training data, you simply describe what a category means, and the AI assigns tags to matching documents across the collection or search case you select.
Smart Tags work alongside the existing Technology Tags system in IPRally. Once enabled, any tag with a text description becomes a Smart Tag that can be applied automatically.
How to access Smart Tags
To use Smart Tags, your organization admin must
ensure that your organization has the Generative AI features (see Enabling AI Assistant Features) enabled, and that you are on Unlimited plan, and
ensure that the Smart Tags and Tagging permissions has been granted to your user account.
After the permission has been granted to your account, you need to refresh the page or log out and log back in to access the feature.
Setting up Smart Tags
1. Open the Technology Tags view
In Settings, go to the Technology Tags page under the Classification section. Here you can see all existing tag groups and tags in your organization.
The Technology Tags view shows all tag groups, their tags, descriptions, and visibility.
2. Create a tag group
Click Add and select Add new tag group to organize your tags. Enter a name for your new group and click Add. Tag groups help you keep related categories together.
The Manage Tag Groups panel lets you create, rename, and delete tag groups.
3. Add a tag with a description
Click Add and select Add new tag, and then fill in the details:
Tag Name — a short label for the category (e.g., "Office chairs").
Description — a natural-language description of what belongs to this category. This is what makes it a Smart Tag.
Tag Group — select the group this tag belongs to.
Visibility — toggle "Make public" off if the tag should only be visible to you.
Adding a tag with a description enables Smart Tags for that category. The description is used to perform the automatic categorization.
The tag description is key. Write a clear, specific description of what documents should receive this tag. The more precise your description, the more accurate the predictions. You can always edit the description later to refine results.
4. Importing tags from Excel
If you already have a set of tags that you'd like to import from an Excel sheet, you can do that by clicking Add and selecting Import tags. Drag and drop the file in the window, and then select which columns to import the tag name, description and tag group from. A preview screen will show up showing which tags will be imported, and you can review them before continuing.
Using Smart Tags
1. Select documents
Open a search case or collection, select the documents you want to tag, then go to Intelligence and choose Smart Tags.
Select documents and choose "Apply Smart Tags..." from the Selected menu.
Document limit. You can currently run Smart Tags on up to 2,000 documents at a time.
2. Select tags to apply
A dialog lists all available Smart Tags grouped by tag group. Check the tags you want the AI to evaluate against the selected documents, then click Start.
Select which tags the AI should predict. You can select individual tags or entire groups.
Tip: You can select all tags from a tag group in a single run by clicking the checkbox on the title row of that group. Each tag in the group is then evaluated against every selected document.
Reviewing results
View predictions
When the Smart Tags predictions finish, a summary popup shows how many tags were predicted. Predicted tags appear on each document with a ✓ and X icon for you to accept or reject.
The completion popup summarizes predicted tags. Predicted tags appear inline on each document row.
Read AI analysis
Hover over a predicted tag to see the AI's reasoning. The tooltip shows:
AI Analysis — an explanation of why this document matches the tag.
Tag Description — the description you provided when creating the tag.
Hovering over a predicted tag reveals the AI's analysis alongside the tag description.
Transparency by design. Every prediction includes an explanation, so you can verify the AI's reasoning and refine tag descriptions if needed.
Accept or reject individual tags
For each predicted tag on a document, click the ✓ to accept (confirm the tag) or the X to reject it.
Click the ✓ icon to accept a predicted tag, or the X to reject it.
Confirm or reject tags in bulk
To confirm or reject predictions across multiple documents at once, select the documents, open the Classify menu, and use Confirm/Reject predicted tags for selected. You can confirm/reject all tags at once or specific tags individually.
Use the Classify menu to confirm or reject predicted tags in bulk across selected documents.
Best practices
Writing effective tag descriptions
Be specific. The accuracy of the predicted tags will likely improve when you describe the category in more detail.
Include boundary conditions. Specify what should not be included to reduce false positives.
Iterate. Review the AI analysis and adjust your descriptions to improve accuracy over time.
Organizing with tag groups
Group related tags together (e.g., a "Vehicles" group containing "Automobile", "Bicycle", "Electric Car"). This keeps the tag selection dialog organized and makes it easy to apply an entire category set at once.
Visibility settings
Use public tags for organization-wide categories that all team members should see and use. Use private tags for personal categorization needs, such as tags specific to your own monitoring workflow.