Art Direction: Background Design for Campaign Assets
AI is still often generally referred to as "slop" and doesn't always find its place in creative work. It's viewed as cheap or a shortcut to save money and speed up the process. Through a great deal of work with Generative AI I can tell you this is not always the case. AI is still in need of educated prompts, knowledge and experience of the user, taste, point of view and even empathy- at least for now.
Under Armour's willingness to create with Gen AI has brought criticism, but like all new technologies there will be naysayers and detractors. So when they came to me to create backgrounds for their campaign I jumped at the chance. I'd worked in the past with their Creative Director and knew the level and talent she worked with and cultivated so I was on board right away. And this was the task AI image generation, in the here and now, was built for. Abstract environments and backgrounds that support the products and offer support to the energy, messaging, technical innovation, look and feel. The canvas was wide open and the imagination could run wild generating and cultivating the creative vibe of the shoot.
Brands are willing to use Gen AI now to extend campaigns and supplement work already being done. While this is an excellent use case and plays with assets already created via human hands, these sorts of creative digital paintings as backgrounds and environments is a true Gen AI sweet spot. There is no uncanny valley when no one expects photorealism. The margin for iteration is wide. Acceptance is easier to achieve. You can push color, texture, and mood aggressively without the image breaking. I built a system of prompts around the brand's visual language and the season's athletic wear offerings and ran variations until each background carried enough specificity to feel intentional but enough flexibility to work across formats and aspect ratios. We as a team were building the emotional container the shoot would be contained within.
What made this project valuable had less to do with the images themselves and more to do with what it proved. A major sportswear brand needed creative assets produced at speed, at scale, with a level of quality that could sit inside a campaign ecosystem alongside traditionally produced work. Utilizing the technology to deliver something new and novel that in no way felt off brand, cheap or distracting.
The engagement also clarified something about where AI-generated work fits inside a larger creative operation. Working alongside Under Armour's in-house team we were simultaneously producing assets using AI generation tools combined with traditional post-production: compositing, refining, integrating AI output into production-ready files. That hybrid workflow, generation plus traditional post-production, is where the industry finds its creative comfort zone. The brands that move fastest will be the ones that are flexible, pay attention to innovation around Gen AI while continuing to use Generative AI as a tool inside an existing creative pipeline, not a wholesale replacement for it.
For me, the project confirmed a positioning I had been building toward: the value is not in generating images. Anyone can generate images. The value is in knowing what to generate, understanding a brand's visual language and identity well enough to produce work that belongs inside it, and delivering with the reliability that makes a creative director trust you with their campaign. That reputation, built over years in the industry, is the thing no tool can replicate.
