Editorial
Digital Asset Managers and administrators spend an enormous amount of time and energy on metadata application.
It is tedious, thankless work, but absolutely necessary to find assets in the sea of digital petabytes.
What if we could simplify this process?
Time to Get DAM in on the AI Action
In the last few years, we’ve seen a growing number of applications — Facebook, Ancestry.com, Netflix — make use of machine learning to help perform specific tasks. Isn’t it time your digital asset management (DAM) system did as well?
Imagine if DAM learned how to tag and classify digital files as time went on, or recognized your face and made sure that your name was attributed in the metadata for a digital photograph.
Now extrapolate the possibilities artificial intelligence (AI) could have for your organization’s DAM system, where end users all over the world could potentially ask your DAM for your company's latest approved logo in vector format, ensuring an unparalleled brand consistency in terms of file format, color profile and currency.
The truth is every mechanical task on which DAM users spend time on can be improved through AI and machine learning. Let’s look at a few existing AI applications and speculate how the same solutions could apply to DAM.
Facebook’s Facial Recognition
Facebook, which open-sourced its image recognition software in 2016, uses an algorithm that assigns a number to your face based on unique data like the distance between your facial features. This allows Facebook to suggest you either tag yourself or your friend in photographs with confidence.
How Can We Apply This to DAM?
In the DAM scenario, think of a company that wants to identify and tag its CEO in every photograph. This single AI feature alone could potentially save days, if not weeks of manually applying metadata to files.
Skeptics of intelligent tagging and accuracy by machines cite that the “most sophistication auto-tagging falls into the ‘guessing’ class.”
Yet IBM’s Watson beats humans at Jeopardy and is also better at diagnosing cancer than human doctors by about 40 percent. While guesses may be involved, the machine’s level of confidence is extremely high. Facebook’s facial recognition has an over 99 percent accuracy rating for frontal shots.
Personal Assistants
You have an example of AI, which utilizes Natural Language Processing (NLP), right in your pocket: your smartphone's personal assistant. Popular personal assistants include Siri, Cortana and Google Now. NLP systems “translate ordinary human instructions into a language that computers can understand and execute.”
How Do You Apply This To DAM?
Imagine giving a presentation to your executive board and instead of showing a boring report, having up to date analytics available in real-time, activated by voice search.
“DAM System, how many people have logged in so far today?” or “DAM System, which assets are people downloading the most this week?”
Learning Opportunities
Webinar
Apr
18
Redefining Social CX: Drive Growth and Foster Loyalty at Every Touchpoint
Transform social interactions into a competitive edge. Learn how!
Register
Webinar
Apr
23
AI: The Essential Vitamin for a Healthy CX
Join us as we unravel the mysteries of AI adoption and discover the key to unlocking a thriving CX ecosystem.
Register
Webinar
Apr
23
Navigating Market Dynamics: Leveraging Reputation to Propel Business Growth
Join us to uncover the secrets of mastering reputation management in today's competitive environment.
Register
Webinar
Apr
24
The State of Outsourcing: At the Crossroads of AI, Quality, and Cost Savings
Learn how companies can maximize the results from their outsourcing investments.
Register
The Pensters Plagiarism Checker
The checker allows you to plug in any text and verify whether or not it came from a different source or if it is original content. While teachers use it to ascertain students are preparing original writing for academic papers, reimagined, it could be part of an approach to identify relationships between documents.
How Do You Apply This To DAM?
An example would be a company overview statement. The text from that document may end up getting reused in a variety of presentations and white papers used within the company. In this case, the related documents are not “plagiarized” or duplicates, but rather collateral produced using the original company overview statement as source material.
Finding similar documents can be done with really accurate algorithms: basically you analyze documents as points in a space (vectors) and find the distance between them. You can also use this type of artificial intelligence under a variety of algorithms to find categories, and automatically tag documents.
Ancestry.com
Ancestry.com entices metadata application by end users with specific knowledge about a specific domain: Family Heritage. Users can upload assets, but the real value lies in the metadata record itself.
Ancestry.com employs algorithms which scan for percentage matches in metadata and asks the end user to verify “adding” it to their tree (subsequently helping to teach Ancestry.com what data is likely to fit into their tree from its already large and ever-growing database of four petabytes of ancestry related metadata records and historical assets).
Ancestry.com scans this four petabyte information mine of public newspapers, public records and other Ancestry.com member trees to make connections between documents, user trees and metadata records.
How Do You Apply This To DAM?
Incentives go far when trying to convince end users to fill out data or upload new assets. Ancestry does this by offering new tips in the form of little leaves which logged in users see in their notifications. The user can follow these tips to continue to build their family tree. In a sense, it’s a treasure hunt. The end user is given more information in exchange for providing information.
A digital asset management system that allows users to build relationships between their assets and their work and the work of their colleagues (and thus opening up new information to them as they provide information) could incentivize metadata application in the same way that Ancestry.com does. Ancestry.com also constantly analyzes user behavior to make its algorithms even more intelligent.
Netflix
What could Netflix possibly teach digital asset management about AI? Its recommendation engine saves them on average $1 billion annually.
The recommendation engine brings titles to users who would not otherwise find those titles and also helps the company curate its catalog.
When produced and used correctly, recommendations lead to meaningful increases in overall engagement with the product (e.g., streaming hours) and lower subscription cancellations rates. That translates into higher lifetime customer values and lower revenue volatility.
How Do You Apply This To DAM?
Applied to DAM, the above quote looks more like: The digital asset recommendation engine increases DAM system engagement and user adoption, and reuse of niche content, targeted at users who show interest in similar assets. This effectively increases return on investment (ROI) in ways that traditional DAM systems without AI would not be able to do.
Going Mainstream
I’ve never particularly enjoyed explaining digital asset management over and over again to people.
Whether you are excited about the changes that will come to DAM through AI, or hesitant that they won’t live up to these projections, remember: the technology already exists. And every time you pull up a title on Netflix or let Facebook suggest tags for you when you upload a photograph it proves that it works.
Injecting AI into DAM could very well be the thing that makes DAM go mainstream and forever end the question, “What is DAM?”
Learn how you can join our contributor community.
