Everyone in AI advertising is talking about prompts. Which model, which keywords, which settings, which seed. And it makes sense — for a while, knowing the tools was the edge. But that edge is disappearing fast. When anyone can generate a beautiful image in thirty seconds, a beautiful image stops being valuable. What becomes valuable is the thing the tool can’t give you: knowing what the image should mean.
That’s where my work actually starts — not with a prompt, but with a story that’s older than any brand.
The prompt is not the craft
A prompt is a set of instructions. It’s how you talk to a machine. But advertising isn’t about talking to a machine — it’s about talking to a human being who has half a second to decide whether they care. That decision isn’t made on resolution or lighting. It’s made on recognition: I know this feeling. I know what this is about. This is for me.
My background isn’t technical. It began in Persian literature, linguistics and anthropology — years of studying how humans build meaning, how symbols carry weight, and why the same image can feel sacred in one culture and empty in another. Long before I generated my first frame, I was trained to read stories. That turns out to be the part of this job the AI can’t do for you.
What mythology teaches an advertiser
Myths are the oldest advertising we have. For thousands of years they did exactly what a great campaign does: they took an abstract idea — temptation, abundance, protection, rebirth — and gave it a face, a scene, a symbol you couldn’t forget. That’s not decoration. That’s engineering meaning into a form the human brain is built to hold onto.
When I built a brand story around the Garden of Eden — Adam, Eve, the fruit, the moment of desire — I wasn’t reaching for something clever. I was reaching for something everyone already carries. A myth like that comes pre-installed in the audience’s mind. You don’t have to explain temptation; you just have to trigger it. The AI can render the apple beautifully. Only a human decides that the apple is the right symbol in the first place.
This is the difference between generating an image and directing a campaign. One fills a frame. The other loads it with meaning the viewer completes on their own.
Semiotics: why the same visual travels — or fails
Semiotics is just the study of how signs make meaning — colors, gestures, objects, compositions. It sounds academic until you’re launching the same product in five countries and watching one visual win in London and die in Riyadh. A color that reads as premium in the Gulf can read as generic in Europe. A gesture that feels warm in one market feels wrong in another. None of that is visible in a prompt. It’s only visible if you know how to read the signs.
This is exactly where most AI creative work quietly breaks. A freelancer delivers a technically flawless image with no idea whether its symbols mean the right thing to the specific people who’ll see it. It looks finished. It just doesn’t land. And a campaign that doesn’t land is expensive no matter how cheap it was to produce.
Why this matters more as AI gets better, not less
It’s tempting to assume that as the models improve, the human role shrinks. The opposite is happening. As production gets effortless and everyone has access to the same tools, the output starts to look the same — a flood of polished, interchangeable, forgettable content. The scarcity moves. It moves from can you make it to do you know what’s worth making.
That’s a strategy problem and a culture problem, not a software problem. No amount of model progress tells you which myth your brand should own, which symbol will survive crossing a border, or which half-second feeling will make someone stop scrolling. That judgment is the work. The tools just execute it.
How I put it to work inside an FMCG company
I don’t do this in theory. Every day I work inside the creative team of one of the region’s largest confectionery groups, making AI-driven campaigns for products that sell on real shelves in real markets. That context forces the meaning to earn its place: it has to move product, respect the brand, and hold up under a deadline. Mythology and semiotics aren’t a flourish I add at the end — they’re the strategy I start with, before a single frame is generated.
So when a brand asks me what makes AI advertising work, my answer is almost never about the tool. It’s this: figure out the story first. Find the symbol that’s already living in your audience’s mind. Make sure it means the right thing to the specific people you’re selling to. Then — and only then — pick up the AI and produce it. The prompt is the last step, not the first.
Tools will keep changing. Meaning won’t. That’s the part worth getting right.
Working on an international campaign and want it to actually land across cultures? Tell me about your brand — I reply within 24 hours with ideas and a quote.