With so many different types of Artificial Intelligence available - and the disruptive pace of technological advances in both AI and DAM - it can be hard to understand your options.
This article explains nine ways Artificial Intelligence improves DAM performance, and helps you answer the question ‘Does my DAM need AI?’
(This is the first part of our series "Artificial Intelligence in Digital Asset Management".)
The first - and most lauded - use of AI in Digital Asset Management is auto-tagging. This is when you upload digital assets to your DAM system and the AI recognizes the content, automatically applying meaningful metadata.
The benefit of AI-auto-tagging in DAM is that you can quickly apply metadata to millions of assets without human intervention. This represents a significant improvement to DAM system management as it eradicates one of the major bottlenecks in asset ingest - applying metadata. Removing the manual, human element makes your DAM processes infinitely more scalable, allowing you to process millions of assets automatically.
AI can instantly apply more metadata than a human could (by manually selecting keywords from a master list or keying them in). This additional metadata can make your digital assets even more discoverable, opening them up to higher reuse. People can find and use assets based on keywords you might not otherwise have added. This increases the ROI of your assets and reduces the cost of commissioning or purchasing new ones.
Part of auto-tagging is facial recognition. Just like it sounds, this is when your AI can recognize and tag known people who appear in your digital assets.
The AI may have already learned famous faces - and you can train it to spot people specific to your industry. For example, your CEO. This can be applied at image upload or retrospectively to your entire collection, tagging the newly known individual wherever they appear.
Facial recognition can be especially helpful for weeding out your DAM system when someone leaves or withdraws permission for their image to be used. As long as your AI recognizes them, it can find them.
Like facial recognition, object recognition is about your AI recognizing the content of your digital assets. The AI recognizes known subjects - let’s say a dog - in your digital asset and will tag it accordingly. But that’s not all.
The AI will also determine where the subject appears in an image. This means AI can intelligently crop images, retaining the focus of the image intact, whilst meeting your desired size and dimensions (see more below).
But subject recognition doesn’t just apply to the actual physical subject of the asset. It can also apply tags for non-visual elements - such as whether an asset is happy or sad - by recognizing smiles, frowns, etc.
With AI, your DAM can create automatic transcripts of uploaded videos. This is a benefit in its own right - for example, by transcribing a webinar automatically, so you can easily create written content based on it.
But it also helps with discoverability in the DAM. You will be able to enter keywords or quotes into your search and your DAM will look through transcripts to surface relevant videos. And not just the file - the exact snippets relevant to your search.
For organizations that produce a lot of lengthy video content - such as webinars - the ability to locate and jump directly to relevant content is invaluable. It allows for faster discovery and repurposing. For example, quickly and easily creating teasers for social media, to increase engagement with on-demand content.
You’re probably already familiar with reverse image searching in Google - finding other images that look similar to an existing photo. AI makes that possible in your DAM system too. So when you have a digital asset and you’d like to see more of the same, you can simply drag-and-drop it into your DAM visual search. Your DAM system will identify existing digital assets that are similar to the one you want to use.
Say, for example, you want a photograph of the city where your business is headquartered for your annual report. You may commission a photographer to go out and take photos for you. Or find suitable photos from a stock photo library.
Imagine, instead, that you use that stock photo thumbnail to visually search your DAM. You may discover the perfect photo already exists within your own assets. And you can use that instead of spending money on something new and unnecessary. Although this may seem like a minor efficiency, cumulatively across your entire enterprise, it can be significant time and cost-saving.
Another benefit of AI-powered search for visually similar assets is the ability to find and eradicate duplicates from your system - meaning you deliver cleaner search results and a better user experience.
Manually editing images can be a real time-sink, so AI-powered automations can really accelerate and scale your processes. Especially if you’re receiving and using images from a range of contributors or suppliers.
Imagine you’re an e-commerce business, selling products from a number of manufacturers. For a consistent brand look on your website, you present your products as cut-outs on a white background. Because AI can recognize different parts of an image - such as the foreground and the focus - it can automatically remove the background on all product images, at the click of a button.
Or it can intelligently apply a brand treatment to all images. In our article on the benefits of AI-powered image optimization, for example, we talk about optimizing food photos for a more mouthwatering recipe book…
You need to be on all the channels your customers use - website, app, social media - and they all need different image crops and resolutions. Most DAM systems support automatic file renditions and crops for different platforms. What AI adds is the ability to crop automatically and intelligently.
AI can recognize the location of the subject of individual images - for example, if the dog is in the center or to the left of the foreground. This means it can ensure the subject is in the center of the crop, rather than applying a blanket crop to all images - which can result in images with faces, pets, or places sliced in half! AI means you can trust your DAM to not only do a job but to do it well.
Another benefit of AI-powered image recognition is that it can automatically check that uploads to your DAM system are compliant. Not just that they meet the minimum technical specifications - for example - but that they don’t contain anything that breaches your rules. The most obvious example is the ability to filter out adult content before it hits your site.
AI only knows what it has been taught. So whilst your AI plug-in will be able to recognize the Eiffel Tower and Leaning Tower of Pisa, it won’t immediately recognize your London HQ or international office in Toronto. Nor will it automatically recognize all of your products in order to tag them. This is stuff it will need to learn.
Fortunately, you can use custom training/machine learning to teach your AI to recognize what matters most to your business - so it CAN automatically tag product photos with SKUs and prices, recognize your staff members in images and video, or geo-tag photos of your physical locations around the world.
Whether you need AI-assisted DAM depends on the unique use case of your business. We’re sorry we can give you a definitive answer. But here are some questions to ask yourself, which might help you decide whether your DAM needs AI - and what type.
Not all DAM use cases require the addition of AI. For smaller businesses with lower numbers of digital assets, AI may add an additional layer of cost and complexity that isn’t justified. However, for organizations that create, manage and use large quantities of digital assets, AI in DAM will undoubtedly deliver benefits.
AI-powered DAM:
You’ll discover a wealth of AI tools built into WoodWing’s two Digital Asset Management systems - WoodWing Assets and WoodWing Swivle. To talk about the benefits of AI-assisted DAM for your use case, book a call back now!