Skip to main content
AI and Automatic Asset Processing

This article explains several features that, through asset processing, can improve your portal's metadata and search capabilities.

Updated over 2 months ago

This article looks at several features you can manually activate for your libraries. For some of your libraries you may want to take advantage of these capabilities. We outline several use cases below, as well as provide a step-by-step guide on how to activate them.

Asset content search (OCR search)

The asset content search feature analyzes documents (pdf, docx, and pptx) in your projects and libraries for their text content. We make use of AWS Textract and Comprehend for the recognition and content analysis.

Main use cases

Once the text has been recognized in the files, it is stored and used to enhance the search results. All three main searches - workspace search, library page search and guideline search - can access the text content.

Being able to search through the text content of your assets can have multiple benefits.

  • In the workspace it can help find the documents that need updating or editing faster. There is also less manual metadata entry required on that asset for users to be able to find it.

  • In library pages, your teams can quickly find the right contract or documentation they need for a customer.

  • Finally, in the guideline search, users can quickly search across libraries to find the right document faster.

Activating the asset content search

Each project or library has a toggle in its main settings. The default is set to "off". Upon activating the feature by toggling "on", you will immediately see the progress of the asset processing.

Once the processing is complete, you will be able to use the workspace-, library page-, and guideline searches to find documents based on their text content.

Deactivating the feature is as simple as toggling it "off" in the same settings.

Auto-tagging (AI- or suggested tags)

Auto-tagging is a way to generate suggested tags and categorize your assets in an intelligent manner. This feature is available in our Media-, Logo-, Icon-, and Document library types. AI-based Auto-tagging reduces the amount of work needed to properly tag assets in the platform. The Auto-tagging service is based on Amazon (AWS) Rekognition.

(We have a separate help article describing Metadata and (manual) Tags.)

When you set up the feature, it will thoroughly analyze each asset. Suggested metadata tags are then displayed based on the results.

Please note that the file formats supported by the auto-tagging feature are: .png, .jpg, .tif, .tiff, .jpeg, .psd, .psb, .svg, .webp, .heif, .heic, .mov, and .mp4.

By following the steps described below, you’ll be able to use the AI auto-tagging feature to enhance the assets in your libraries.

Setting up Auto-tagging when Creating a New Media Library

When creating a new Media Library under ‘New’ in the top-right corner of your Libraries section, you’ll be instructed to name and choose a Workflow template, or link it to an available Guidelines. You’ll have the option to activate the auto-tagging toggle in the same modal.

Auto-tagging in Existing Libraries

To start using auto-tagging in your Library, you would need to activate it and also generate it for the existing assets. This can be done by clicking on your Library Settings on the Power Bar and then toggle to activate the Automated Tagging field under the Suggested Tags section.

The auto-tagging feature should begin working after a few minutes. Generated tags will appear under the ‘Metadata’ section in ‘Suggested tags’ in the lower information panel within the asset viewer.

When clicking on the ‘Suggested Tags’ you will discover newly generated tags. The recommended tags are easy to accept or reject. If you find there are some that don’t actually fit with the image or video, you’re able to easily eliminate them from the list by clicking to deselect them.

If you deselect a tag by mistake, you can recover it by clicking on the tag once again. However, if you refresh, all the unchecked tags will be completely erased.

Whenever the feature of auto-tagging is switched off, the suggested tags will not appear on the right-side panel.

Understanding Auto-tagging Notifications

There are three possible status' for the auto-tagging processing.

  1. "There are no tags assigned yet"

    This status will appear when the feature is enabled, but tags were not yet generated for the asset.

  2. "In progress"

    When the feature is enabled and tags are being generated, then the ‘Suggested Tags’ section will have a label saying that the tags are ‘in progress' and therefore being generated.

  3. ‘Suggested tag generation for this asset has failed..’

    When the feature is enabled and the tag generation was not successful for some reason, then an error message saying "Suggested Tag generation for this asset has failed” will appear. In this case, don’t hesitate to contact us: support@frontify.com  

If the file format is not supported, the ‘Suggested Tags’ section will not be displayed in the metadata sidebar.

In case you have any further questions, please don’t hesitate to contact our support team. 

Did this answer your question?