Assigning a precise dollar value to the challenges faced by AI-assisted creative teams is difficult, but each represents a potential obstacle threatening their success. A Gartner survey from October of 400 global marketers found that 58% use generative AI for content production. Many advertisers aim to develop semi-automated systems like Unilever’s, although building such an AI-driven production line can take over a year. Craig Elimeliah, chief creative officer at Code & Theory, likens AI production to “building your own house instead of renting someone else’s. ” This “house-building” involves legal consulting, selecting appropriate large language models (LLMs) for a brand, and painstakingly compiling brand guidelines and past content into briefs that generative AI can understand. It also requires extensive trial and error to ensure the system reliably handles brand-sensitive material. This setup demands significant time investment, ironically at a moment when 81% of marketers measure success by the time saved through AI, according to the Gartner study. Dave Rolfe, global head of production at WPP’s Hogarth, adds that the most costly element is adapting to this new process. Additionally, recruiting skilled personnel to design, implement, and operate such AI systems is challenging amid competition for AI talent from tech giants and advertisers. James Thoams, global CTO at Dentsu Creative, emphasizes that “AI talent is hard to come by. ” Generative AI is often accessed via subscriptions, but some companies, including OpenAI, cap usage and sell credits in a pay-as-you-go model, which can make intensive testing with premium models expensive. Ómar Thor Ómarsson, CEO of Optise—a B2B AI tool provider improving organic search performance—warns that companies generating large volumes of content on demand face escalating costs. Though individual prompts cost only fractions of a cent, campaigns involving tens of thousands—such as Coca Cola’s Christmas ad which used 70, 000—cause expenses to quickly add up. Ómarsson notes that “undisciplined testing with big prompts and premium models can quietly add up. ” Legal concerns also persist amid ongoing copyright disputes between AI firms and authors, causing some clients to hesitate over which LLM to deploy. Larger agencies address this by offering indemnification. For instance, WPP integrated compliance checks into its WPP Open platform from early last year. Brands working in-house typically lack such protection, making compliance a significant consideration.

Rolfe stresses the need for contained systems tied to compliance standards. Beyond technology, traditional human workflows pose delays. Approval procedures within agencies and client organizations often consume more time than the AI-driven content creation itself. Elimeliah explains that the “real cost isn’t generating assets, it’s generating your assets, ” as someone must review, judge, and refine the multiple options a single AI prompt generates, turning previously invisible decision work into explicit tasks. This leads to “client approvals” becoming bottlenecks, as AI produces content in minutes but clients take weeks or months to approve. Elimeliah highlights this time gap as “the most expensive part of the pipeline. ” In response, some organizations have revamped their briefing processes and use generative AI with specialized templates to elevate initial creative briefs. Gartner analyst Nicole Greene observes that many clients now deploy generative AI early to develop higher-quality strategic creative inputs for their agencies. Rolfe at Hogarth describes a shift in production philosophy: using AI enabled the company to cut production time for a telco client’s promotional campaign materials from seven weeks to two. However, gaining such efficiencies requires adopting a “component mindset, ” prioritizing post-production workflows over traditional rapid bulk content capture. Other firms are employing automated quality control systems that apply pre-flight checklists assessing factors such as aspect ratio, lighting consistency, logo timing, and performance benchmarks. Despite such automation, marketers remain cautious about fully hands-off AI processes. 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. ” In sum, while generative AI presents great opportunities for content creation efficiency and quality improvement, its effective implementation demands significant time, investment, talent, legal safeguards, and adjustments to human workflows—each representing potential risks that creative teams must navigate carefully.