Fashion brands can use Generative AI to produce more adaptable e-commerce content, but human creative judgment is still needed to protect image quality and brand identity, Pixel Moda CEO Fabio Loparco said in a conversation with Fairchild Media Group CEO Amanda Smith presented by Pixel Moda.
Pixel Moda works with more than 900 fashion and luxury brands to create content for their channels. Loparco said brands often turn to generative AI to cut costs, but he sees its greater potential in helping them create more content and generate sales.
He emphasized the need for careful quality checks. AI-generated images can contain errors, such as a model with extra fingers or garments with incorrect details, and brands need to ensure that their identity is not lost. Loparco also warned that brands risk producing similar content if they rely on the same tools, prompts and models. Tailoring styling to different markets, including New York, Dubai and Tokyo, can help make content more relevant to local audiences.
Generative AI can also support a move from static e-commerce imagery toward content adapted for different countries and consumer niches. Loparco said brands can test the effect on sales and track results for specific stock keeping units.
As an example, Pixel Moda produced business-to-business photography for a client, using more editorial photos and videos to enrich product detail pages. The company said the work helped increase the brand’s revenue by more than 15 percent, cut overall costs by 49 percent and reduce product returns by 12 percent.
Loparco advised brands to begin testing rather than spend weeks or months developing an AI strategy without practical trials. He suggested starting with one category of 20 to 30 products, reviewing the results, then adjusting or expanding the approach. He described testing as the most important phase.
He said he views the technology as improving, estimating its accuracy at about 90 percent now compared with 60 to 70 percent last year. His preferred approach, he said, is to combine human expertise with AI rather than use AI in place of people.
Sources: WWD.
