It wasn’t too long ago that artificial intelligence (AI) technology was merely a concept from science fiction movies. But today, AI is changing the status quo of entire industries, including marketing technology (martech).
In the marketing sphere, we are increasingly seeing AI capabilities leveraged in martech tools: the use of automation capabilities used to support chatbots, personalize content, and even manage social media platforms.
As consumer expectations shift, and agile, disruptive retail players enter the ecosystem, marketers need to ensure that they are still able to compete for consumers’ attention across an increasing number of channels. For these marketing teams, this means ensuring any visual assets used are both accurate and compelling enough in a competitive landscape to convert consumers into paying customers.
One way in which businesses have been streamlining their content strategies to allow for AI is through digital asset management (DAM), built to support marketing teams to organize, find, distribute, and analyze digital content. With AI capabilities on the rise, we ask how AI can be used to streamline DAM workflows?
Image recognition functionality
In order to be effective, DAM software relies on metadata, or descriptive information about each piece of content, to provide structure and information to make digital assets findable. Metadata is crucial to a system’s success but can often be time-consuming to add and is prone to error without a clear process.
When working with metadata, fields are used to answer different questions about each asset including identifiers like the file name, photography type, description, and usage rights. A system of metadata fields is called a metadata schema. Creating a schema with fields relevant to a specific company improves the search experience and helps ensure that metadata is added accurately. And with the help of image recognition software, AI has a valuable and growing role in system management through its ability to automatically tag assets with relevant metadata during the upload process.
The benefits of using image recognition software
Employing AI to automate part of the image-tagging process can improve categorization, power the accuracy of related assets, and offer advanced searching options for users. Not to mention, saving hours of manual tagging efforts.
The ability to automatically categorize and tag images extends to a range of scenarios. It can recognize faces or demographics — including the age, ethnicity, or gender of persons in the image. It can also apply keywords for specific industries, such as travel, food, and apparel.
Adding the power of image recognition to a DAM system makes metadata creation simpler, faster, and most importantly, better. Image recognition software can reduce human errors and inconsistencies, and avoid assets being uploaded onto the DAM system without any metadata.
Furthermore, AI can make searching more effective as thorough metadata allows search tools to return accurate results, quickly.
As well as creating accurate content, AI enables marketers to ultimately save time on the more remedial and security-based tasks, spotting anomalies and flagging potential errors. Image recognition software includes features that allow DAM administrators to:
- Moderate assets and automatically delete or hide uploads with inappropriate content.
- Enable auto-tagging functionality for specific upload profiles or metadata fields.
- Influence how the controls for strength, relevance, and variety are configured.
Automating a manual process reduces the time needed to tag assets, meaning workflows can be better streamlined. All of these efficiencies translate to business cost savings. After all, less time spent tagging and searching means more time for creative and strategic work.
Support brand agility:
Perhaps unsurprisingly, following the effects of Covid-19, it is estimated that marketing teams will lose around US$222 billion from budgets, and around 30 percent of their staff, by the end of 2021. This is according to a 2020 Forrester report, which studied the effects the pandemic is predicted to have on US marketers. In contrast, it is estimated that by 2023, marketing automation budgets will hit an all-time high of US$25 billion, up from the US$11.4 billion in 2017. As with any uncertainty, businesses will need to remain flexible and agile in order to not only stay afloat, but to keep up with consumer demands. This means that the automation of an organization’s workflow processes will soon be a necessity, not an option.
Bias in AI
In recent months, we’ve seen several examples where AI has been labeled as ‘biased’. Most recently, Twitter was accused of having biased artificial intelligence image-cropping algorithms. To stay away from such negative headlines, the tech giant announced that it would simply do away with the image-cropping software in favor of its users now cropping their own images as they see fit.
With such negative headlines regularly making press, there’s no doubt that we run the risk of marketers being too fearful to truly amerce their workflow process in AI in fear that they too will become subject to ‘biased AI’, resulting in a PR crisis.
And yet, there is a way that businesses can ensure they do not fall victim to such demises. Marketers and developers need to not only have a thorough understanding of the data being used, but also the patterns that the data generates. In doing so, AI tools can successfully be harnessed to automate a marketing team’s system management both effectively and efficiently.
Ultimately, it is not the artificial intelligence technology that is biased but the algorithms built by human teams that are its downfall.
What does the future hold?
Despite some people’s hesitations, there is no denying we are in fact moving into a digitally transformed, online world. But although technology is helping a lot of companies work with greater efficiency and speed, it is also clear that this cannot be achieved without some element of human intervention. Technology alone is not able to fix an organization’s workflow processes. A strategy must be put in place to align the workforce, processes, and technology — such as AI — in order to achieve the desired marketing outcomes.
Tools like image recognition require a balance between automation and human touch. While it can free up time and resources for other value-driven projects, the workforce is still needed to ensure accuracies across all data inputs, as well as informing and guiding both the technology and its users.
As we continue to see developments in this space, we will see the partnership between AI and the human workforce grow more important to marketing teams. Not only to streamline and scale DAM workflows but to successfully create cost savings, too.
Jake Athey, VP of Marketing and Customer Experience, Widen