Explore how artificial intelligence (AI) and machine learning (ML) work alongside humans to get the most value out of enterprise digital asset management.
When we talk to teams who want to optimize their digital asset management (DAM) protocol, we hear a common story:
Both the time wasted looking for assets and the time and materials invested in recreating them has a considerable impact on operating costs. We’ve heard estimates of as much as $30,000 to $50,000 for each asset that needs to be recreated!
Your assets don’t have to be so hard to find.
Artificial intelligence (AI) is useful in identifying image attributes. Applying AI using machine learning (ML) models — especially when those models are trained on your specific business data — adds real value to your operation. Together, AI and ML:
Let’s go into depth about the power of AI and ML, and how simplifying the correct, consistent tagging of assets reduces image hide-and-seek.
Nuxeo Platform’s DAM offering was named a Leader in the DAM space in Omdia’s 2023 report. Our solution achieves the maximum score for advanced capabilities and solution breadth, and Omdia places our market momentum as above average for the field.
Many C-level executives consider content creation and management a vital capability for supporting their go-to-market strategies. Yet time and again, we hear from organizations that are stuck in their efforts to wrangle modern digital asset journeys with their current, legacy technology.
While there is often a large investment in managing structured data — usually through some sort of content services platform or ECM — unstructured data regularly goes undervalued, making it difficult to get necessary funding to support.
As a result, content ends up spread across multiple locations, in siloed systems, on employees’ personal drives or as email attachments — making it almost impossible to locate. Even when tools such as DAM platforms and federated search engines are used, they still need the content to be correctly tagged with consistent metadata.
Applying the right metadata manually is a time-consuming and repetitive task. It also traps a lot of important knowledge with a small amount of people.
— Sir Arthur Charles Clarke’s Third Law (1973)
We all encounter AI in some form or another every day, whether it’s ChatGPT, voice assistants, mapping services, or receiving personalized marketing messages or customized digital experiences based on our online behavior. Almost every enterprise is now using some form of AI in their day-to-day business operations as it becomes increasingly embedded within the tools and technologies they have deployed.
A recent McKinsey report on the state of AI details many enterprise benefits post-AI adoption, including cost decreases and revenue increases. AI is here to stay.
When you consider what AI, and ML in particular, are fundamentally designed to do — automate and assist with repetitive tasks that need some degree of human intervention and specialist knowledge —the benefits of their applications become clear.
AI and automation are traditionally leveraged to increase:
Marketers are tasked with managing an explosive growth of content to satisfy an increasing number of marketing channels. Many of these channels require specific and specialized formats, languages and variants on a global scale. The result is the prioritization of digital asset management as a foundation for any marketing content ecosystem.
For many years the key promise of DAM was that it helped organizations store and manage assets in a central location where they can be found and retrieved easily. However, this promise lost efficacy as marketers activated content in increasingly diverse digital and physical channels, leading assets to be stored and managed in disparate systems. Often, large enterprises create multiple DAMs that don’t communicate with each other and have little or no common metadata.
Additionally, new content types such as complex video and 3D models are often unsupported by older DAM solutions.
Learn more: DAM best practices to optimize efficiency
Bridging different content repositories and silos can be achieved using platforms that support the concept of federated search. To support that indexing, correct, relevant and consistent metadata is essential.
This is where AI and ML adds real benefits in eliminating the frustrations and cost of looking for assets.
The benefit and value of AI is not just about knowing what questions to ask or what models to build. It comes from connecting AI with the right sources of data.
The value of ML comes from helping to build those data sources in the right way, at scale. ML can apply the right information to the right assets at speeds that are impossible for a group of human subject matter experts to do manually, and in a consistent way.
Consider a hotel chain using first-generation asset management tools. It needs to categorize thousands of photographs to ensure the right ones are displayed at the right place in online and print promotional materials:
Metadata woes: While there is a general acceptance of the importance of metadata, no one really wants to be responsible for applying it. Even with DAM and other content management systems that have workflows to make metadata entry a required step when uploading an asset, the results are often minimal and inconsistent.
Paradoxically, the more you know about an asset and what it contains the more useful it becomes. If you take the hotel chain’s situation and apply AI and ML, the images could be automatically (and correctly) tagged with all the proper metadata — hotel location, date of shoot, photographer, area of hotel, lighting, campaigns it should be used in, etc., — immediately upon entering the digital supply chain. No more 10 seconds per picture, no more begging someone, somewhere to attach metadata.
Consider an apparel manufacturer that depends on its talented people to make its processes work smoothly downstream:
That’s a lot of people just to tag one image (assuming they take time to do it), and it all depends on their locked-up knowledge. Imagine trying to do that for thousands, or tens of thousands, of assets. It just doesn’t scale.
As we discussed earlier, without the right metadata, it is almost impossible to find the right assets at the right time, leading to time lost in searching and in recreating them/
However, with trained AI and ML models in place, a modern DAM solution would:
The sort of problems outlined above, where we need to automate repetitive tasks that need some degree of human intervention and specialist knowledge, are perfectly suited to the application of AI and machine learning.
Generic AI services connect to a broad set of public services for common use cases (general classification, enrichment, OCR, speech-to-text, etc.) and use commodity models to provide those generic services. Example: It may tag a particular vehicle as being a red pickup truck, but that is as detailed as it will go.
Custom models deliver more meaningful outcomes for the business by training machine learning models on your content and data to get highly relevant insights and enrichments. This enables specific business use cases across defined domains. Example: A custom model will identify that red pickup truck as a 2023 Ford F-150 Lightning, and maybe even recognize which types of accessories have been installed.
The result is not about training the ML to perform a repetitive task, it’s about developing a system where the output can be trusted by humans to deliver what they want, at the time they need it. From a DAM perspective, this means being able to locate the right image at the right time for the right use case.
With assets that have rich metadata models applied in a consistent manner, the business can use that information to connect with other systems across the digital supply chain.
Correctly configured, an enterprise DAM can be more than just a tool used by the marketing department. It can unify processes from product ideation through manufacturing, e-commerce, distribution and support while also delivering a consistent customer experience.
With a DAM solution powered by AI and ML, marketing teams can:
Thinking of the DAM as the center of the digital supply chain makes it more than just a place to store and retrieve media assets. It makes it a key platform for enterprise data strategy.
Organizations that have a clear data strategy with solid content platforms and pipelines will be able to operate more efficiently. Companies best positioned to deliver business benefits with AI- and ML-enabled metadata are those where the results are integrated into business systems and processes.
In this era of content and data, AI can play a key role in delivering more value from an organization’s existing content and data libraries. Given that each organization’s content and data is unique, AI models trained on individual organizational data can predict, classify and enrich content with much greater accuracy while avoiding potential legal issues with public AI models. Going beyond utilizing common but generic AI services, most organizations will benefit from developing their own custom and business-specific ML models that will better serve their unique business needs and data requirements.
AI is not going to take away jobs. Instead, it will make many activities less mundane and more satisfying. The desire for authenticity will see the rise of the cobot: Humans working in partnership with AI to produce quality services and products. AI will streamline processes and help with automation, while humans use their emotional intelligence to create meaningful content and experiences.
Many businesses have only just begun to capitalize on what AI can do, but Hyland’s AI-powered Nuxeo Platform for DAM has been paving the way for years.
Nuxeo for DAM is AI-infused, open source, extremely scalable and cloud-native, all of which sets it apart from other vendors.
Connect with a Hyland representative to see how AI can impact your organization’s asset management efforts.
