5 Ways to Make Your Content Management AI Work Better and Faster

Artificial intelligence can be a virtual assistant to users in the content management and digital asset management space. It can speed up processes and make it easier to focus on the work that really matters. 

Even if you already use digital asset management AI to manage content, there are ways to tweak your settings so that it does even more for your users.

Breaking down Digital Asset Management AI Terms

Machine Learning: Artificial intelligence programmed to learn “on its own” as it gets new information. (Ex: Translation)

Computer Vision: Artificial intelligence that tries to replicate human vision so it can recognize objects and people. (Ex: Auto-tagging; facial recognition)

Get better auto-tagging results

Picture an organization with more than 10 million assets to tag. If they had a person tag all those assets, even with batch editing capabilities, it would take about a hundred years and several million dollars to catalog them all. 

When it comes to content management, artificial intelligence can simplify that process. A computer vision system can look at all those assets in a fraction of the time (months instead of a century) and turn in a bill in the thousands rather than millions.

The trouble is machine learning can’t beat the human eye, and so sometimes digital asset management AI will add tags an admin might not approve of.

Auto-tagging does work, though — at least if you do it the right way. There are a few simple fixes that ensure auto-tagging returns reliable results.

• Use multiple AI providers: Some digital asset management systems can look at the combined results of multiple computer vision systems. While you might doubt the results from just one provider, if you average the results from several, your odds of accurate results go up significantly.
• Set an accuracy threshold: You want your digital asset management AI to tag the best possible results. Your DAM should be able to let you know how confident the AI is of a given term, and let you auto-tag assets beyond a certain percentage.
  This works even better with multiple computer vision services. For example, if Google and Microsoft Azure both are more than 90% certain an image has a house in it, then you can feel confident that when considered together they’re probably right.
• Review new terms: Digital asset management AI services may suggest terms that are synonyms for terms you already have in your system. To prevent your thesaurus from getting bloated with multiple terms that mean essentially the same thing, check out any new terms your AI suggests.
  If you use OrangeDAM, you can route new terms to be vetted by administrators; then admins can approve those terms or set them as synonyms to existing terms. If you use another system, it may have a similar set of automations you can put in place.

Make videos more searchable

Videos with captions have more value than those without. Beyond providing accessibility to those who need it, using speech-to-text to create captions also makes the videos themselves easier to find and navigate. Certain content management artificial intelligence features make that easier:

• Indexed captions: Make sure your DAM can index captions to increase the amount of information available to your users. Indexing makes the results of speech-to-text searchable. That means any line of speech in a video can be searched, rather than users trying to find information based on just an asset’s title and description.
• Edit speech-to-text: Auto-captioning gets you a lot further a lot faster than manually captioning videos. But if a part of the transcript is a little off base, you want a content management AI that lets you easily make and save edits.
• Transcript navigation: A good digital asset management AI will connect the captions to time-stamped parts of the video. So if you click on text, it’ll jump to that section of the video, letting you find the section of the video you need faster.
• Autotagging & video: Some DAMs (like OrangeDAM) not only let you search for video assets, but let you jump to a specific spot in a video where a tag is mentioned. For instance, if you're searching for a segment of a video talking about pizza, you can select that tag and go straight to any part of the video discussing pizza.

Connect people to assets with facial recognition

Facial recognition works by mapping the features of a person’s face. Research has shown that the AI for facial recognition is up to 99.97% accurate. That accuracy makes facial recognition useful for identifying public figures, individuals within your organization, donors for nonprofits, etc. Adding that to your content management artificial intelligence stack can help you speed up your time to market.

For instance, say a public figure is in the news. You can quickly search for assets featuring that person and share them out to be a part of the conversation. 

A few other ways to leverage facial recognition:

• Batch edit the metadata or visibility for any assets featuring a certain person. For example, say you end a contract with a certain model. You can search for their name and then remove their images from public view.
• Search for a specific model or actor featured in a photoshoot with several people. If using OrangeDAM, you can then submit the assets that include them to their agent for approval. In other systems, you can upload the assets to a talent approval or photo-culling software for review.
• If your digital asset management system can run facial recognition on a video, you can use that feature to jump to sections with a specific person.

Translate text with artificial intelligence

A global enterprise needs to have assets that are accessible to offices around the world. Being able to quickly translate content without a human translator handy is vitally important for keeping sites accessible to multi-language audiences. Luckily, the quality of A.I. translation has only been improving, with services like Google Translate improving constantly as it analyzes 150 billion words a day.

A digital asset management system can integrate with machine learning systems like Google Translate to give you:

• Auto-translates captions: Take the captions we talked about in the speech-to-text section above and run them through a translation tool for captions in multiple languages.
• Translated metadata: If your users speak more than one language, translate your metadata so that it is searchable in all the languages your users need.
• Translated interface: Offer your interface in multiple languages so that you can reach a broader audience.

Protect your brand with computer vision

An asset has a Coca-Cola logo on a bottle in the background of a video — and you’re not Coca-Cola. Computer vision can let you spot the logo and make sure you don’t get into legal trouble. And a DAM system or content management system can integrate with those AI tools, keeping your brand safe. A few features to add to look for:

• Identify logos: Make sure you can find your logo — and anyone else’s — in your assets.
• Explicit content: Artificial intelligence can identify explicit or violent content that might be inappropriate.
• External usage: See how your assets are being used outside of the DAM.

Change your AI outlook

The frustrations you feel when using the artificial intelligence in your DAM is often just how it’s set up. Change how you’re using the AI resources you have, and you can accomplish so much more.

Orange Logic has more than 20 years of experience crafting digital asset management systems that go above and beyond. Find out how we can guide you to an A.I. strategy that makes sense for your assets and your business.