We agree with Theresa Regli's "Auto-tagging technology can it replace humans" and others that machine tagging is not a panacea for all tedious metadata tagging.  A person needs to check to see if the tags are correct or useful.  While auto-tagging cannot replace the librarian or DAM Manager who must add the metadata tags, it can serve as a time-saving mechanism for staff to review tags and to make corrections. The key is the quality of the automated tags and the metadata/taxonomy structure. If the auto-tags are not useful, then the auto-tags can either duplicate or create more work. If the auto-tags are useful then there is a potential for some labor reductions, however, this largely rests upon the groups' metadata and taxonomy structure requirements. The metadata and taxonomy structures are key to answering the next question: which is better for automating metadata tagging- Automated Tagging via Image Recognition or Integrating the DAM and PIM? The business requirements for metadata and taxonomy structure determine the answer.  Since metadata and taxonomy structure vary between businesses, there is not a single answer, however, there are key elements that will help guide you.

The Case For Auto-Tagging With AI Image Services

If the tags generated from an  AI Image Recognition Service - such as Google, Amazon, or Azure - is accurate and applicable to a company's metadata requirements, then the Image Recognition Service can serve as a time-saving tool.  Auto-tagging is not a complete solution, but if it can save staff time by reducing the number of fields to enter or correct, then automation is worth the investment. There are several AI Image Recognition tools on the market with their own advantages for tagging certain types of assets including facial recognition, speech, and audio transcription, and text reading. We recommend researching the tools to see what is the best fit.

The Case For Auto-Tagging Through DAM-PIM Integration

Product Information Management (PIM) systems serve as a repository of information related to assets with lots of useful metadata for specific business cases.  Syncing the information between the DAM and the PIM saves time but automating and improves metadata accuracy, however, there are some restrictions:

• Not every business needs a PIM or has an equivalent.  PIMs are usually found in retail, consumer product goods (CPG), pharmaceuticals, construction, health sciences, and manufacturing.
• The quality of the data of the PIM determines the success of the metadata enrichment.  Before considering a PIM-DAM integration, a business must examine the quality of the metadata within the PIM.  Sometimes, the metadata in the PIM will need to be cleaned before integration.  Good quality data will make automation easier.  In addition, the frequency in which the metadata is updated affects the quality.  Metadata stored and gathered by PIM must be updated regularly.
• The quality DAM-PIM Integration affects the accuracy and quality of the metadata.  To accomplish this goal requires extensive business analysis, a health metadata and taxonomy structure, quality APIs, and the ability for data to follow back and forth between the two systems.
• A PIM's data will not cover 100% of the DAM's metadata.  Usually, marketing and creative departments will have specific data such as marketing campaign information that will not be in the PIM.  The related metadata for related metadata will need to be entered.

The Answer Depends Upon Your Business Needs

In sum, your business needs will determine which tool - auto-image recognition, DAM-PIM integration, and/or both - is the best solution as a time-saving measure but not a complete replacement for the human element. From CyanGate's experience, the DAM-PIM integration has the greatest potential as a time-saving measure by improving the accuracy of the information in several industries for our customers.  If you want to access what method is better for your business needs, contact us. Over the years, CyanGate helped several customers improve their metadata quality through both DAM-PIM integrations and Image Recognition solutions to improve our customers' metadata across different industries. CyanGate specializes in OpenText Media Management (OTMM), SAP DAM, 4ALLPORTAL, and DAM solutions.

DAM and PIM In One Solution

If you are looking for a DAM and PIM in a single solution, please check out CyanGate's partner, 4ALLPORTAL. 4ALLPORTAL provides both DAM and PIM in a single solution. By having the DAM and PIM in a single product, time and cost-saving when upgrading and maintaining the system. Usually, the API calls connecting a DAM and PIM must be updated each time either the PIM or DAM is upgraded. Do you want to learn more about 4ALLPORTAL? Contact us here.