What if the biggest problem with AI-generated creative work isn't the AI?
I've been looking at the state of AI-powered advertising lately, and there's a pattern that's hard to ignore. The outputs are technically competent — clean layouts, decent copy, passable imagery — and yet they're almost universally… forgettable. Bland. The kind of work that slides right past you in a feed without leaving a trace. And the instinct is to blame the tool. But I don't think the tool is the problem.
What happened
The conversation around AI-generated adverts has shifted. The early excitement about speed — "we made 50 ad variants in an afternoon!" — is giving way to a harder question: why do most of those variants feel so generic?
The emerging consensus is surprisingly simple. When you feed AI a vague brief, you get vague work. When you skip the strategic thinking — who's this for, where will they see it, what should they feel, what makes this brand different — the AI fills in the gaps with the most average, middle-of-the-road option it can find. Because that's what language and image models do. They converge on the mean.
The fix isn't better AI. It's better inputs. Detailed prompts that specify audience, platform context, tone of voice, campaign objectives, and brand constraints. In other words: doing the design thinking before you hit generate.
The bigger picture
This isn't just an advertising problem. It's a design problem — and it's one that touches every discipline where AI is entering the workflow.
We're in a moment where execution is getting cheaper by the day. AI can generate layouts, write copy, produce illustrations, even rough out motion graphics. The mechanical cost of making things is plummeting. And that means the value is shifting. Hard.
It's shifting upstream. To strategy. To the brief. To the ability to articulate — with precision — what a piece of design needs to do, not just what it needs to look like.
This is uncomfortable for some designers, because in design we've long tied our identity to craft. To the making. But here's the thing: craft isn't going away. It's just changing shape. The craft now is in the clarity of your thinking. Can you define the problem tightly enough that an AI tool — or a junior designer, or a collaborator in another timezone — can produce something that actually works?
I'd argue that writing a brilliant creative brief has always been undervalued in our industry. It's about to become the most important skill in the room.
What this means in practice
Think about your own workflow for a moment. When you use an AI tool — whether it's for generating visuals, exploring copy, or prototyping layouts — how much time do you spend on the input versus evaluating the output?
Most of us spend almost all our time on the output side. Generating, reviewing, regenerating, tweaking. But the leverage is on the input side. A prompt that specifies "a confident, slightly irreverent headline for a fintech app targeting freelancers in their 30s, for an Instagram Story, emphasising security without sounding corporate" will outperform "write me a catchy headline" every single time.
This applies beyond prompts, too. If you're building AI into a design system or a production pipeline, the quality of your constraints — your brand guidelines, your component libraries, your content strategy docs — becomes the ceiling for what AI can produce. Garbage in, gorgeous garbage out.
Tool spotlight
This is less about a single tool and more about a practice. But if you're looking for somewhere to start, I'd recommend building what I call a prompt brief template for your team.
It's dead simple. Before anyone touches an AI tool, they fill in a short doc: audience, platform, objective, tone, constraints, and — crucially — what success looks like. Not what the output should look like. What it should achieve.
You can build this in Notion, a shared Google Doc, even a Figma frame. The format doesn't matter. What matters is that it forces the strategic thinking to happen before generation begins. I've seen teams cut their AI iteration cycles in half just by being more intentional about this step.
The takeaway
AI is an amplifier. Feed it sharp thinking and you'll get sharp work. Feed it mush and you'll get polished mush.
So the next time you're disappointed by what an AI tool gives you, don't reach for a different model or a cleverer prompt hack. Reach for a clearer brief. Ask yourself: did I actually do the design thinking here, or did I outsource it to a machine that has no idea what my audience cares about?
The designers who'll thrive in this era aren't the ones who are best at prompting. They're the ones who are best at thinking — and then translating that thinking into constraints an AI can work with. That's always been the job, really. AI just made it obvious.