As marketing agencies increasingly integrate generative AI into their workflows, a fundamental shift in pricing and operational models is underway. The core challenge stems from the token-based economics of large language models (LLMs), where every prompt and response incurs a computational cost. With campaigns like a recent Coca-Cola effort consuming millions of tokens, agencies face a critical decision: absorb these AI overheads or pass them directly to clients. Currently, there is no industry consensus, leading to a fragmented landscape of approaches. Some agencies, like Merge and Big Spaceship, treat AI compute as a metered, pass-through cost similar to traditional production expenses. Others, such as RPA and Anomaly, absorb the costs, arguing that the technology’s value is still too unproven to charge for, with concerns that early pass-through fees could be perceived as a „money grab.”
In response, new pricing frameworks are emerging to manage token economics at scale. Companies like Pencil (part of Brandtech Group) leverage bulk agreements with LLM providers to offer clients „generation credits,” with pricing tiers based on volume commitments. This model aligns incentives by charging for usage while ensuring larger clients don’t subsidize smaller ones. Media agencies are taking a different tack; Horizon Media, for instance, recovers costs for its AI platform Blu through a nominal fee focused on barrier-free client adoption rather than profit, while Kepler embeds AI costs into retainers, prioritizing delivered „impact” over token tracking. Full-service agency Lerma/ includes token costs as a transparent, non-marked-up line item in client estimates, viewing any surplus as an opportunity to deliver extra value.
Looking ahead, the industry is grappling with whether AI token management could evolve into a contentious practice akin to principal media buying, where agencies profit from arbitrage. However, most leaders argue this is a strategic misstep. The greater economic advantage lies in „labor compression”—using AI to streamline operations—which far outweighs marginal gains from token markups. The focus, therefore, is shifting from cost-tracking to value creation. As Ebiquity’s CEO notes, while token auditing may become standard, the true metric for agencies will be demonstrating how AI drives incremental business returns for clients, moving beyond a narrow focus on input costs to compete on transformative merit and outcomes.
Ez a cikk a Neural News AI (V1) verziójával készült.