After an internal preview in May
, Meta is finally introducing its first-ever set of generative AI features for advertisers. The new features available within Meta’s Ads Manager include Background Generation, Image Expansion, and Text Variations.
allows advertisers to create product image backgrounds to complement different campaigns and audience segments. While similar to Meta’s Backdrop, an image editing tool that can edit photo backgrounds based on text prompts, the new ad-focused feature will create “simple backgrounds with colors and patterns” that are “generated based on original product images.” This feature is available for Advantage+ catalog users.
Advertisers can repurpose images and videos to fit different aspect ratios, like Feed or Stories, with the Image Expansion tool. This feature is also available in Advantage+ within Meta’s Ads Manager.
Lastly, advertisers can generate up to six versions of their original ad copy with just a click of a button. Text Variations
, also available within Meta’s Ads Manager, allows advertisers to highlight keywords or phrases from ad copy and edit the raw AI results, or simply use the version they like the most.
Advertisers can run the same ad with different text variations to see what drives the best results. While they can’t track the results for each variation separately, this option gives advertisers more opportunities to connect with their audience, ultimately improving their overall ad performance.
Meta developed the new AI-powered ad features in collaboration with select advertisers. Feedback has been positive, with most of its early adopters saying that they “expect saving time,” while some estimate that they’ll save “five or more hours a week,” which translates into one month per year. Almost all advertisers believe they will achieve quicker ad development at scale and improve their campaign performance.
Meta said its first generative AI ad toolkit
is “just a start” and it plans to offer a broader range of AI-powered tools for advertisers to increase productivity and performance. The company recognized that customization is still an area that needs improvement. Namely, Meta is working on finding ways to enable brands and agencies to train its AI models to produce content unique to their voice and visual style.