The Metadata Enhancement tool provides GAC partners with direct control to run and tune existing Optical Character Recognition (OCR) features. This allows to extract text from images and documents within the collections, further enriching the metadata (Transcripts field) and enhancing discoverability and usability of the assets on GAC platforms.
To start, you will need to provide a list of maximum 500 assets IDs using one of the following buttons.
The tool will run a preliminary analysis of the assets and confirm their validity. After that you can press ok and the preview will be generated.
As for all the Metadata Enhancements functionalities, there are 2 options for the handling the data output, "Apply" and "Run and manually approve":
- Apply: This option automatically accepts and saves the AI-generated content into the missing metadata fields on the CMS. It's suitable when you are confident in the AI's accuracy and want to quickly enrich your collection's metadata.
- Run and manually approve: This option generates the AI-suggested content for missing fields but requires your manual review and approval before the changes are saved. This gives you direct control to fact-check and verify the accuracy of the generated content, ensuring it aligns with your standards before publishing. This is particularly useful for critical fields or when a higher degree of human oversight is desired. You can then copy and paste the output into separate documents for sharing and editing, before eventually reporting the text into the “Transcript” field of the CMS.
If you then export your data to a CSV file from the CMS, these suggestions are included. If you edit this CSV file and re-import it, all "requires review" suggestions from the CSV will be automatically approved.
They will keep their AI-generated status and association with the enhancement job. You will also see an extra line in the confirmation dialog after the re-import.
Please be mindful of this if you use a CSV export/import workflow.