When AI Replaced the Photoshoot

We needed recipe images that hit an exact nutritional ratio.

The project was a nutrition-focused guide built around the plate method — a framework where half the plate is vegetables and fruit, a quarter is grains or healthy carbs, and a quarter is protein. Every recipe had specific ingredients, specific proportions, a specific look. The imagery needed to show all of it accurately, or the guide wouldn't work.

Stock photography doesn't work that way. You search, you scroll, you compromise. The protein is wrong. The portion sizes are off. The greens are there but the grain is missing. After enough searches you start negotiating with yourself — this is close enough — because the alternative is an expensive food photography shoot and a weeks-long production timeline.

I stopped searching and started generating.

For each recipe I input the dish name, the specific ingredients, the plating style, and the plate method ratios directly into Adobe Firefly. What came back was a dressed, styled plate image built around that exact meal — the right protein, the right grain, the right proportion of vegetables. Every time. In minutes.

The image above is one of them. Greek yogurt chicken with saffron and garlic, quinoa, and a mixed green salad. Overhead shot, white background, plate method ratios accurate to the guide. It looks like it came from a food photographer's studio. It came from a prompt.

No shoot. No location. No licensing fees. No back-and-forth over whether the broccoli should be roasted or raw. The image matched the content because I built it directly from the content.

What changed for me wasn't just the time saved — it was the creative control. I wasn't browsing a library hoping something close enough existed. I was describing exactly what I needed and getting exactly that. The tool became part of the design process rather than a workaround for it.

Close enough is not a forced choice anymore. That shift is bigger than it sounds.

---

Tools used: Adobe Firefly