Meta’s generative AI creative tools are features in Ads Manager that create new versions of your ad from what you upload: new backgrounds behind a product photo, wider or taller versions of an image to fit different placements, and alternative wordings of your primary text and headlines. Some also turn still images into short animated clips.
How Meta’s generative AI tools work
The tools sit inside the ad creative step, mostly under the Advantage+ creative enhancements. Background generation takes a product image, separates the product and offers several new backdrops. Image expansion extends the edges of a picture so a square photo can fill a vertical Stories or Reels frame without awkward bars. Text variations read your copy and suggest rewrites that Meta may rotate in delivery.
Some enhancements are switched on by default, and Meta may apply them at delivery without you approving each version. You can usually preview the suggestions and turn individual enhancements off. At the time of writing (October 2026), the set of tools, their names and their default settings change frequently, and some ads made or edited with AI may carry an AI label, so check the options each time you build a campaign.
None of these tools knows anything about your business beyond the creative and text you give it. It cannot tell that your shop is in a Victorian terrace, that your product only comes in three colours, or that a phrase would breach your industry’s rules. That knowledge has to come from the person reviewing the output.
Why it matters for a UK business
For a small business without a designer, these tools make it easier to supply the variety of sizes and versions Meta’s system rewards. A bakery with one good square photo can get a vertical version for Reels and a couple of background options without a new shoot, which supports creative diversity at little cost.
The risk is accuracy. UK advertising rules apply to the finished ad however it was made, and the ASA has made clear that advertisers are responsible for AI-generated content. An AI background that makes a room look bigger, a text variation that turns “may help” into “will fix”, or an expanded image that adds features your product does not have can all make an ad misleading under the CAP Code. For beauty and cosmetics ads, imagery that exaggerates what a product does is a known problem area.
Common mistakes
- Leaving every enhancement switched on by default and never looking at what was produced.
- Letting text variations add claims, prices or promises you cannot back up.
- Using generated backgrounds that misrepresent the product’s size, colour or setting.
- Applying the tools to regulated sectors, such as finance or health, without a human review of every version.
- Assuming the AI understands your brand voice, then finding American spellings or a tone that is not yours.
How to act on it
Turn the enhancements on deliberately, one at a time, rather than accepting all the defaults. Preview every generated version before publishing and reject anything that changes what the product looks like or what the copy promises. Check the text variations for British spelling and for claims you could not prove if asked.
Once ads are running, use the ad preview and reporting to see which versions Meta is delivering, and switch off any enhancement that produces work you would not sign off yourself. Reviewing AI-generated creative is part of how I run Facebook and Instagram ads management, and it fits naturally with a flexible ad format set-up.