Every conversation about AI in marketing design eventually lands in the same place: speed. How fast can you generate a banner? How many ad variations can you spin up in an afternoon? How quickly can you go from brief to deliverable?<\/p>\n\n\n\n
Speed is real. I’ve seen it. But watching teams chase it has convinced me that speed is the wrong frame entirely \u2014 and that chasing it is how you end up with a brand that looks like it was assembled by committee, in a hurry, from a template.<\/p>\n\n\n\n
I run UX for marketing at Cvent, a B2B events platform. We operate at the intersection of scale and brand specificity \u2014 large enough to need systems, specific enough that off-brand feels immediately wrong. That context shapes how I think about what AI is actually doing to marketing design, and what it’s still nowhere near capable of.<\/p>\n\n\n\n
Let’s be honest about what’s genuinely different. A few things have shifted in ways that are hard to overstate.<\/p>\n\n\n\n
The first draft problem is mostly solved.<\/strong> For years, the most time-consuming part of producing marketing design wasn’t the thinking \u2014 it was the execution of the first pass. The blank canvas. The initial layout. The first five headline options. AI has largely absorbed that cost. Tools like Adobe Firefly, Midjourney, and generative features inside Figma have made first-draft creation close to instant. That’s real leverage, and I’d be lying if I said it hasn’t changed how my team operates.<\/p>\n\n\n\n Variation at scale is now viable.<\/strong> One of marketing design’s oldest constraints was the cost of producing variants. A single campaign asset used to become five variants at best \u2014 different sizes, maybe a colour flip, perhaps two headline options. Now teams can test dozens of meaningful variations without proportional increases in production time. Digital advertisers are already seeing this in practice \u2014 testing ad variations in seconds to identify the creative most likely to perform. For B2B marketing, where the audience is narrow and the stakes per impression are high, that testing depth changes what’s possible.<\/p>\n\n\n\n Personalisation has teeth now.<\/strong> Dynamic content that adapts by segment, by stage in the funnel, by industry vertical \u2014 this was theoretically possible before. It was practically unaffordable for most teams. AI has closed that gap. Salesforce data shows 78% of marketers report that personalisation strongly impacts customer relationships \u2014 and AI is finally making that personalisation executable at a level beyond “Hi [First Name].”<\/p>\n\n\n\n Here’s where the speed conversation starts to break down.<\/p>\n\n\n\n The more output you generate, the more opportunities you create for brand drift. And AI, left without strong creative guardrails, drifts fast. It gravitates toward the average \u2014 the visual language that appeared most frequently in its training data, the headline structure that “works” in the aggregate, the layout that feels familiar because it’s been used ten million times.<\/p>\n\n\n\n AI struggles to maintain consistency over time, which can be a problem if you’re trying to stick to a particular brand voice. That’s an understatement for anyone managing a brand with genuine distinctiveness. The problem isn’t that AI produces bad work \u2014 it’s that it produces competent, generic work at scale. And competent-and-generic, multiplied across a hundred touchpoints, is corrosive to brand equity in ways that are slow to show up and expensive to fix.<\/p>\n\n\n\n Brand consistency across all channels increases revenue by 10\u201333%. The inverse is also true: inconsistency actively damages the equity you’ve built. When your campaign emails sound different from your landing pages, which sound different from your event materials, customers stop knowing what your brand stands for. They just know it feels a bit off.<\/p>\n\n\n\n Coca-Cola’s approach to this is instructive. Their AI campaigns gave them access to thousands of creative variations while maintaining brand consistency through AI guardrails \u2014 the platform only allowed generation using pre-approved brand elements, colours, and compositions. The framework stays fixed. The variations happen inside it. That’s the right model \u2014 but it requires significant upfront design thinking to establish the framework in the first place. AI doesn’t create that. Designers do.<\/p>\n\n\n\n There’s a second-order effect of AI-accelerated marketing design that doesn’t get enough attention: everyone has it.<\/p>\n\n\n\n If your team can generate fifty ad variants in an afternoon, so can every competitor in your category. If AI lowers the cost of personalised creative, it lowers it for everyone simultaneously. The baseline rises. Differentiation gets harder. The surge in AI-powered content creation has brought a new challenge \u2014 the proliferation of low-quality, mass-produced content, sometimes called “AI shovelware,” posing risks including brand dilution and audience fatigue.<\/p>\n\n\n\n This is the paradox at the heart of AI in marketing design: the tool that promises competitive advantage is available to all your competitors at the same price. What you do with it \u2014 the brief you give it, the taste you apply to its outputs, the creative framework you’ve built for it to work inside \u2014 that’s where the differentiation actually lives. And that part is still entirely human.<\/p>\n\n\n\n AI campaigns deliver 32% more conversions, but human-created content gets 5.44x more traffic. Read that twice. The two things are not in conflict \u2014 they point to different roles. AI excels at optimising and scaling within a defined creative direction. Humans are still far better at originating creative directions that actually capture attention.<\/p>\n\n\n\n I’ve been thinking a lot about where this leaves people who do what I do \u2014 designers embedded in marketing organisations, responsible for brand quality at scale.<\/p>\n\n\n\n The uncomfortable truth is that some of what marketing designers spent time on is genuinely automatable. Production tasks, resize variations, first-draft layouts, initial copy options \u2014 AI handles these competently. Teams that haven’t reckoned with this yet are going to face the reckoning eventually.<\/p>\n\n\n\n But the work that moves up to fill that space is more interesting, not less. It looks like this:<\/p>\n\n\n\n Creative direction becomes the core skill.<\/strong> If AI is executing, the value shifts to those who can direct it well. What brief do you give? What do you accept, reject, and push further? What constitutes “on-brand” versus “close enough”? These are judgment calls that require deep knowledge of the brand, the audience, and the competitive context. They can’t be prompted.<\/p>\n\n\n\n Systems thinking becomes non-negotiable.<\/strong> Building the guardrails \u2014 the approved elements, the tone guidelines, the explicit definitions of what on-brand means \u2014 this is design infrastructure work. It’s not glamorous, but it’s what separates teams that use AI well from teams that use it chaotically. 65% of marketing teams now have dedicated AI governance structures. The ones that don’t tend to be the ones with the brand consistency problems.<\/p>\n\n\n\n The brief is now a design artefact.<\/strong> How you prompt AI is a creative skill. A vague brief produces generic output. A precise, well-considered brief that encodes brand voice, audience context, and creative intent produces something you can actually use. Writing good briefs \u2014 for AI and for human collaborators \u2014 is underrated as a design capability and is becoming central to the job.<\/p>\n\n\n\n When I look at the marketing design work I’m most proud of \u2014 the campaigns that landed, the assets that felt right, the brand moments that actually moved people \u2014 none of it came from a tool. It came from a point of view.<\/p>\n\n\n\n A clear opinion about what this brand should feel like in this moment, for this audience, given everything else competing for their attention. AI can execute against a point of view. It cannot originate one.<\/p>\n\n\n\n Creative AI is maturing into a hybrid, modular model \u2014 human concepts, AI scaffolding, rapid iteration, and new workflows that unlock experimentation without linear timelines. That framing feels right to me. The most honest version of what AI does for marketing design is remove the friction between an idea and its execution. It’s a better pencil, not a replacement for knowing what you want to draw.<\/p>\n\n\n\n The designers who are going to thrive in this environment are the ones who have a strong enough point of view that AI becomes leverage rather than a crutch. The ones who can look at fifty AI-generated options and know immediately which direction is right \u2014 and why. The ones who treat taste as a skill worth developing, not something to outsource.<\/p>\n\n\n\n Speed was never the point. The point was always the idea.<\/p>\n\n\n\n
\n\n\n\nThe Brand Consistency Problem Nobody Talks About<\/h2>\n\n\n\n
\n\n\n\nThe Noise Problem<\/h2>\n\n\n\n
\n\n\n\nWhat This Means for Marketing Design Teams<\/h2>\n\n\n\n
\n\n\n\nThe Question I Keep Coming Back To<\/h2>\n\n\n\n
\n\n\n\n