March 30, 2023 | Samuel Chapman
In case you’ve been preoccupied with another totally transformative technological breakthrough and haven’t had the time to notice, generative AI is blasting its way into the annals of tech history. In what seems like a matter of short months, generative AI tools have already started to significantly disrupt the way we ideate, create, and manipulate digital content. By 2025, Gartner anticipates that 30% of outbound messages from large organizations will be synthetically generated, up from less than 2% in 2022. And by 2025, generative AI will account for 10% of all data produced, up from less than 1% in 2021. While the promise of generative AI is exciting, the challenge of managing the coming avalanche of content should be top of mind for everyone.
Digital asset management (DAM) is already an absolute essential to organizations that deal with the daily challenges of content chaos in a digital world – from trying to manage multiple repositories, content calendars, and content waste to running content operations at enterprise scale. DAM will become increasingly essential in the boom of generative AI for several reasons. In this post, we’ll explore how DAM helps organizations manage their content scarcity problems of today while preparing them for the content abundance challenges that will undoubtedly stem from this generative AI content bloom. We’ll also consider how a generative AI integration for DAM can deliver solutions for content discovery, prompt management, and intelligent creative workflows to help marketing teams achieve omnichannel personalization at scale.
Generative AI is a type of machine learning that uses neural networks to create new content. Unlike other types of machine learning that are trained on existing data, generative AI, while still trained – uses a generative model to create new data from scratch based on a large variety of data, including Common Crawl, webtexts, books, and Wikipedia. Additionally, generative AI can be trained to produce better output.
Generative AI has many applications, including the creation of text, images, and videos. In the film industry, generative AI is used to create special effects and virtual backgrounds, making it easier and cheaper to produce films. And marketing teams across industries are starting to use generative AI to create emails, social copy, collateral, and a variety of other digital content.
Organizations are under enormous pressure to create and deliver on-demand, personalized omnichannel content across customer segments and verticals. Despite constrained resources and this demand for higher volume and faster delivery, teams have achieved new economies of scale, often without making changes to people, process, or technology. They’ve grown accustomed to the “we’ve always done it this way” thinking which prevents them from being strategic. So, they produce a ton of content, but 70% of the time, it’s not the right content and goes to waste. In this sense, the challenge for most organizations today is content scarcity. New tools like ChatGPT, Dall-E, and Runway will rapidly move organizations from content scarcity to a world of content abundance.
The pace at which all this is happening is not without risk. It presents a slippery slope with potentially dire situations for organizations, which already face challenges in how their human-generated content is organized, stored, accessed, updated, and delivered. That’s where DAM comes in. DAM software is a foundational technology that is core to mature content operations. It acts as a system of record to help teams manage their rapidly multiplying digital assets. Streamlining the processes of asset creation, storage, and retrieval saves time and money while providing deeper insights into their content usage and performance.
There are many benefits of digital asset management that can address today’s content challenges:
Governance: Provides a single source of truth across global locations, internal businesses, and platforms. Capabilities like metadata, tagging, uploading, permissions for different user groups, and managing digital rights, facilitate consistent and efficient management.
Content Discovery: Allows users to quickly and easily search, find, access, collaborate, share, and deliver assets.
Improved Collaboration: Enables more effective collaboration by sharing assets and working on them together, leading to better communication and more efficient workflows.
Increased Efficiency: Streamline and automate workflows, reducing the time it takes to create, review, and approve creative assets, which increases productivity and cost savings.
Enhanced Security: Ensure that assets are stored securely and can only be accessed by authorized users, which helps to protect sensitive information and intellectual property.
Better Asset Quality: Ensure that assets are consistent and of high quality, leading to better customer experiences and improved brand reputation.
Improved Decision-Making: Get data analytics and insights that can be used to make better decisions about asset creation, distribution, and usage.
Composable Content Ecosystem: Encourages integration with a broad partner ecosystem to build composable content stacks and increase the value of your existing martech investments.
Combining generative AI with the right digital asset management platform has a lot of potential.
Auditing, Traceability, and Transparency: A DAM that can audit how content was created and which AI tools were used to do so. It can then track that content, its usage, and performance to understand which tools are producing the most effective content and identify opportunities where generative AI can be used to improve other aspects of the content lifecycle.
Prompting: Getting good generative output requires good prompts. When you hand a task to your copywriter, all sorts of campaign planning information can help feed a good prompt. For example, who is it targeting? What’s the aesthetic? What is the objective of the campaign? Asking a generative AI, “Can you write me a blog covering how great this toothpaste is?” can be greatly impacted with prompt modifiers, such as “Can you write me a blog covering how great this toothpaste is, focusing on the fact that 9/10 dentists recommend it, it leaves you with a minty fresh feeling, and the % chance it has to reduce cavities?” A DAM that also holds your campaign planning data can streamline a creative’s ability to produce generative AI output that’s consistently on brand and on message across your channels.
Brand Safety: DAM can help manage the varied and complex metadata for each AI-generated asset, which improves consistency and accuracy, and helps organizations adhere to responsible use policies that ensure brand safety.
Advanced Search: Traditional search methods, such as keyword searches, may not be effective in finding the desired assets in a generative AI context. Advanced search capabilities, such as computer vision for image recognition and computer hearing for auto-subtitling and content analysis, are needed to search and filter assets effectively, as well as using generative AI to automatically set metadata for better user experience on content uploads.
Responsible Use: Once generative AI is integrated into a solution like DAM, it’s essential that all instances of use must be flagged as such. The more content generated by AI, the higher your risk for situations like copyright infringement because it was trained on non-public domain content. Much like OpenAI’s code of conduct, organizations must implement meaningful human oversight, set limitations to reduce misuse beyond an intended purpose, and mitigate certain undesirable or scenario-specific behaviors.
Trained AI: The importance of training AIs to know what good content looks like will be critical. GPT today is trained on an internet dataset. It knows nothing about your tone, brand voice, or the values you want to speak to. You can influence GPT by prompting, but you can reduce the need for specific and complex prompting by training GPT to already be aware of those values or how to talk to a certain audience. And how do you do this? By training it on content usage and performance data models in the DAM.
Democratizing Creative Skills: Tasks previously reserved for creative roles, like removing objects from images, can be democratized so that a channel marketer can adjust content on the fly in new ways. While Aprimo today offers the ability to transform content on-demand (i.e., watermarking, grayscale, height/width adjustments, blurs, etc.), with generative-AI powered transformations, we could offer many more on-demand adjustments to the channel marketer without them having to tap a creative.
Organizations that invest in DAM will be better equipped to manage the complexity and volume of assets produced by generative AI, thereby increasing productivity and cost savings and delivering better customer experiences. Aprimo’s product vision is underpinned by its already mature AI, excelling in AI metadata extraction, AI starter packs, and the ability to train AI to recognize business-specific content. This allows our customers to put guardrails in place today, for the coming flood of AI-generated content. We must put the right processes in place for an ethical approach to AI that prioritizes quality, efficacy, and brand safety while also addressing the challenges of leveraging generative AI to meet omnichannel demand so businesses may maximize the potential gains while minimizing material risk.
At Aprimo, we take a practical and principled approach to AI. We believe AI should be seamlessly embedded into the user journey, affordable with a low barrier for entry and adoption, contextual, proprietary, and responsible. Amid all the hype and the rush towards the brave new world of generative AI, organizations need to look for a DAM that understands the nuances of content operations and generative AI together, both in today’s existing content stacks and those of tomorrow.
Sign up to receive our latest content on best practices, trends, tips, and more to elevate your content operations.