Generative AI has done more to change the creative production side of advertising in the past two years than anything since desktop publishing in the 1990s. Concepts that used to require a photographer, a studio, and a post-production team can now be prototyped in an afternoon. Copy variants that used to take a copywriter a day can be generated in minutes. The economics of creative testing have fundamentally changed. But the marketing industry, prone to overcorrection, has moved from dismissing AI creative entirely to treating it as a complete replacement for human creative judgement — and that's where things are going wrong.
What's Working: Volume, Speed, Variation
AI-generated creative delivers clear value in three specific contexts. The first is creative volume for testing. Performance Max and Advantage+ campaigns reward advertisers who supply many creative combinations — headlines, images, video cuts — so the platform's AI can test and optimise. Generating 20 headline variants or adapting a base image across ten different backgrounds to test different visual approaches is now fast and cheap. The creative testing velocity this enables would have been cost-prohibitive a few years ago.
The second is background generation and product photography enhancement. Tools like Adobe Firefly and Amazon's AI background generator allow e-commerce brands to take a single product image and place it in multiple lifestyle environments — on a kitchen bench, in a living room, outdoors — without a photoshoot. For brands with broad catalogues, this is genuinely transformative. The output quality for simple product shots is now indistinguishable from photography in many cases. Our AI content studio uses exactly these tools to produce scaled creative assets for social and advertising campaigns.
The third is rapid iteration on copy. Using AI to generate multiple framings of the same benefit — three headlines emphasising price, three emphasising convenience, three emphasising social proof — and then testing them against each other is exactly the use case the technology was built for. The AI is a fast, tireless copywriting assistant for generating raw material. The human's job is to select and refine.
What's Not Working: Brand Coherence at Scale
The failure mode with AI creative is homogenisation. When everyone uses the same tools with similar prompts, the outputs start to look similar. The hyper-saturated lifestyle imagery, the particular rendering of people and environments that comes from the same foundational models, the cadence and structure of AI-generated copy — these are recognisable to any trained eye, and increasingly recognisable to consumers. The brands running fully AI-generated creative without human creative direction are publishing ads that look and feel like everyone else's ads.
Brand-specific creative — work that reflects a genuine visual language, a consistent tone of voice, a specific aesthetic point of view — still requires human creative judgement to define and maintain. AI can execute within a well-defined creative brief. It cannot generate the brief. The businesses winning with AI creative are using it as production infrastructure for a human-defined creative strategy, not as a replacement for having a creative strategy at all.
Meta Advantage+ Creative: What You're Actually Getting
Meta's Advantage+ Creative features — which automatically add enhancements to your uploaded ads including brightness adjustments, text overlays, background variation, and music — are on by default and require deliberate action to turn off. For many advertisers, the platform is automatically modifying their creative without them knowing. Some of these enhancements help. Others introduce inconsistency with brand guidelines. You need to check your active Advantage+ campaigns and understand which enhancements are enabled.
Meta's AI is also now generating entire ad variants from your existing creative library and testing them automatically in Advantage+ Shopping Campaigns. The performance data shows these AI-generated variants sometimes outperform human-designed ads on pure conversion metrics, particularly for direct-response e-commerce. But "converts well" and "builds brand effectively" are different objectives, and the AI is only optimising for the first one. Managing this tension requires the kind of strategic social media oversight that keeps brand consistency front of mind even as the platform takes more creative control.
Google Performance Max: Creative as Fuel
Performance Max is fully dependent on your creative assets. Google's AI assembles combinations of headlines, descriptions, images, and video and tests them across all Google inventory. The quality of the outputs is directly proportional to the quality and diversity of the inputs. If you provide two images and three headlines, Performance Max has limited material to work with. If you provide fifteen images across multiple contexts, five videos at different lengths, and ten headlines targeting different aspects of your value proposition, the campaign has genuine creative material to optimise.
The practical implication: treat your Performance Max asset group like a creative brief, not a form to fill in. Every input you provide is an instruction to Google's AI about how to represent your brand. Poor inputs produce poor outputs.
What's Coming: Video Generation
The next frontier is AI-generated video at a quality level suitable for advertising. Tools like Runway, Sora, and Google's Veo are producing video content that, for certain use cases — abstract visual content, product demonstration, background video — is approaching broadcast quality. The constraint in 2026 is still consistency of human representation (faces and body movements remain a tell) but for a significant proportion of ad creative that doesn't require realistic human footage, AI video generation is already viable. Expect this to change the economics of video ad production within 12 months.
If you need an AI creative strategy that delivers performance results without sacrificing brand coherence, our AI services team can build the framework and production workflow that makes it work. The tools are available — the strategic direction is what most businesses are missing.