With over 100M active users in just four months after its launch, ChatGPT claimed nearly half of all unique visitors to Microsoft’s Bing search engine within its first month.
Moreover, Snowflake recently announced a number of capabilities to bring generative AI and large language models (LLMs) directly to customers’ proprietary data through a single, secure platform. The allure of this generative AI technology is its speedy, seemingly effortless results that allow for creative assets to be created at scale.
So, what’s the catch?
One challenge is the bottleneck being created for the human-powered and analyzed data sets that accompany content and assets. More specifically, marketing analytics, data operations and ad ops teams are overwhelmed with an increased need to categorize and standardize associated data, which is compounded by the accelerated pace of generative AI-produced content.
Here are some of the biggest implications of generative AI’s looming content overload.
Inconsistencies and silos within the enterprise
Improved compliance and consistency are critical as AI delivers larger amounts of creative assets. In fact, the American Marketing Association cites that 90% of marketing materials are never even put to use because they’re irrelevant, out of date or inaccessible.
Successful AI tools need quality data to operate effectively. Without the correct metadata or naming conventions as inputs, AI doesn’t have the proper resources to provide relevant results. Thus, irrelevant, incomplete or inconsistent assets become plentiful, and data gaps develop from content creation all the way to measurement.
While it is very easy to assume that a simple folder system for categorization will solve content retrieval issues, it’s also very easy for this type of organizational structure to become just as muddled and inconsistent. In fact, a 2020 Gartner study reported that poor data quality costs organizations an average of $12.9M in lost revenue annually.
And there is more to consider. AI-generated content causes an overload for the analytics and operations employees who take on the burden of sifting through large amounts of data to find a specific asset. Forrester found that teams can spend 2.4 hours per day searching for the correct data, amounting to 30% of a person’s week. The projected growth of AI will only increase this overload.
The allure of more content is a desirable outcome on the surface. However, large brands run the risk of sharing more content without any control, which can present problems for businesses. When exploring the content supply chain for any team or brand, it can become more challenging to govern the people, teams and various technologies at play to deliver the right customer experience.
Measurement and attribution challenges
AI-fueled content overload can also have trickle-down effects on measurement and attribution, especially when it comes to metadata and taxonomy inconsistencies. When marketers are unsure how well various assets in a campaign perform because of the lack of data organization and tracking, it is challenging to measure campaign results.
Since measurement and analytics teams are at the receiving end of the content supply chain, they are often left to interpret what other teams or systems have provided. The goal is to feed better inputs into AI models. Otherwise, attribution and personalization can be daunting with more accelerating amounts of content. Even more so, as more content is created with the help of AI, being able to measure the performance of a human-created asset against that of a computer-generated asset could surface very valuable insights for optimization decisions.
With some recent announcements from Snowflake and other platforms, LLMs are becoming more accessible and commoditized. This means differentiation is going to be squarely on data access and the ability to consistently and accurately feed data to finally turn on the promise of AI for commercial use.
Yet data inconsistencies, employee burnout, measurement and attribution challenges are important considerations. The growth of AI-powered technology can act as an incredible resource for marketing teams, but it will become even more important to stay agile and be ready to update current practices as the AI revolution continues.
“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Follow Claravine and AdExchanger on LinkedIn.
For more articles featuring Chris Comstock, click here.
