The Hidden Costs of AI in Marketing

Marketers have increasingly turned to artificial intelligence (AI) technology in hopes of streamlining operations and reducing costs. By automating the creation of marketing assets, they envision cutting weeks of work down to mere hours. However, while the allure of increased efficiency is strong, many are finding that hidden costs linger beneath the surface.

The Scarcity of Top Talent

One of the most significant challenges marketers face is the scarcity of qualified AI talent. The competition isn’t limited to the advertising world; tech giants are also on the hunt for skilled professionals who can navigate the complexities of AI systems. "AI talent is hard to come by,” states James Thoams, global chief technology officer at Dentsu Creative. Thus, while attempting to implement AI-driven solutions, many marketing teams find themselves in a battle for the human resources necessary to make these systems effective.

Building a Creative Assembly Line

In the quest for efficiency, marketers aspire to establish semi-automated production systems. Unilever is a prime example of a company that has initiated such an assembly line, but setting it up took over a year. Craig Elimeliah, chief creative officer at Code & Theory, draws an apt analogy: “AI production is the equivalent of building your own house instead of renting someone else’s.” This “house-building” involves a multitude of tasks, from consulting legal experts to sifting through brand guidelines to make them digestible for AI tools.

The time commitment required to develop these systems can quickly eat away at the very efficiencies they aim to create. In fact, 81% of marketers in a Gartner survey cite time saved as their principal measure of success for AI investments, a somewhat ironic twist given the extended development periods common in these projects.

Unforeseen Costs: Usage and Compliance

The costs skyrocket when considering usage fees for generative AI tools, which often operate on a subscription basis or a pay-as-you-go model. OpenAI, for instance, offers credits for business use, and these expenses may rapidly accumulate with high-volume usage. Ómar Thor Ómarsson, CEO of Optise, points out the potential pitfalls: “If they’re using it to create content on the fly, a lot, that’s going to cost the company money each time.”

A single prompt may only cost a few cents, but large-scale campaigns can demand tens of thousands of individual prompts. For example, Coca-Cola’s Christmas ad utilized a staggering 70,000 prompts, illustrating how quickly those small costs can add up.

Additionally, legal concerns around copyright and compliance are constantly at play. As the battle for AI ownership and rights rages on, brands need to be cautious. Larger agencies, like WPP, have begun embedding compliance measures into their platforms, but in-house teams lack that level of protection.

The Time Sink of Human Oversight

Despite the speed that AI can provide, the human elements of approval and oversight often act as bottlenecks. Elimeliah notes, “The real cost isn’t generating assets, it’s generating your assets.” While generative AI can churn out hundreds of options in minutes, the subsequent decision-making processes can prove cumbersome. Approval pathways and review meetings designed for traditional creative workflows are now outdated. The disconnect between the rapid production of content and the glacial pace of approvals can result in substantial delays.

As Elimeliah illustrates, “AI collapses the time it takes to make content but does nothing to collapse the time it takes to approve it.” This gap becomes one of the most expensive aspects of the creative pipeline, challenging marketers to rethink their approaches to both production and approval.

Reevaluating Briefing Processes

In response to these challenges, some brands are innovating within their briefing processes. By employing generative AI to refine briefs early on, they can better prepare teams for efficient creative production. “Most of my clients are now using generative AI specifically to develop higher quality, strategic creative groups to pass to their agencies,” explains Gartner analyst Nicole Greene.

At Hogarth, Rolfe describes a transformative shift in production philosophy, allowing his team to reduce the time taken to produce marketing materials significantly. Instead of focusing on capturing as much content as possible in a short time, they are adopting a "component mindset."

Quality Control at Scale

While many marketers express excitement about the possibilities of AI, concerns around quality control persist. Automating the creative review process could lead to problems if not handled meticulously. A typical approach might involve a series of pre-flight checklists to ensure assets meet standards; however, many marketers remain hesitant to automate these reviews entirely. Ómarsson notes, “AI content can be awesome and it can be awful. We’re all learning to trust what’s coming out of the LLMs.”

As the landscape of AI in marketing continues to evolve, the dialogue around efficiency must be accompanied by a deeper understanding of the intricate complexities and hidden costs that come into play. The hope for streamlined processes and reduced expenses is valid, but marketers must navigate the nuanced realities of implementing AI technologies effectively.