Disclaimer: This is an independent creative test made with AI, for demonstration only. It isn't an ad for any brand shown or named, and isn't affiliated with, sponsored or endorsed by them. All trademarks belong to their owners
Nano Banana Pro - why your AI product shots look fake
The same product shot twice on one model, a lazy prompt against one that names the material
Tested by Nouamane Benbrahim
Ever wonder why your AI product shots look fake even when the model is good?
I ran a test. I took five products and generated each one twice on the same model, with the same settings. The first time I typed a lazy one-liner. The second time I described what the thing was actually made of. That was the only change. None of these are real products, this was just a test.
Here's my exact process.
1. Name the material, not the object
My first perfume prompt was a lazy one-liner, and I got a generic bottle back. Fine, but flat. So the second time I described what the glass actually does: the thickness, the meniscus where the liquid meets the surface, the way it bends the light behind it, the crisp highlight down the edge. Same model, and it came out way closer.
Exact prompt
a glass perfume bottle with amber liquid, on a table, product photo
Exact prompt
a crystal-clear glass perfume bottle, refraction bending the light behind it, fine surface detail along the glass, visible glass thickness and a real meniscus where the amber liquid meets the surface, sharp edge definition with crisp highlight lines down the sides, a polished glossy finish catching a soft studio key light, deep near-black background
2. Add the small real-world flaws
A real product photo is never perfectly clean. For the soda can I asked for cold condensation beads, a few droplets sliding down, and the embossed logo catching a little shadow. The lazy can looked like a video game prop. The second one looked like it just came out of a fridge.
Exact prompt
a can of soda, product shot, high quality
Exact prompt
a chilled aluminum soda can, brushed metal with fine micro-texture and specular highlights running down its side, a soft key light from the left catching the edge, a rim light separating it from the dark backdrop, cold condensation beads catching the light, the embossed logo catching a subtle shadow, a few droplets sliding down, every element feels tactile and present, hyperreal, highly detailed
3. Match the flaws to the material
Different materials wear in different ways. On the whiskey I added a faint condensation haze near the base, a few dust particles floating in the light, and a real meniscus on the amber.
Exact prompt
a whiskey bottle, product photography, studio, high quality
Exact prompt
a whiskey bottle, a faint condensation haze near the base with a few droplets, the embossed lettering on the glass catching shadow depth, a few dust particles floating in the backlight, sharp edge definition on the glass with crisp highlight lines, visible glass thickness and a real meniscus where the amber liquid sits, moody studio backlight
Why it works
The model draws what you tell it to draw, and a vague one-liner tells it almost nothing. Every lazy prompt came back generic the same way: a clean object with no history. Once I described what it was made of and how it wears, the model had something real to render. The words that did the work were all material words: what it's made of, how thick the glass is, where it beads, where it collects dust.
Your turn: pick one product and write down what it's made of and three things that would be true of it in real life. Prompt those.
And if you'd rather skip the prompting, Emero writes the prompts for you and remembers your brand so you're not repeating yourself every chat. Everything's at cost.
Get one of these every day
I test new workflows daily and send them with the exact prompts, the models and the costs
5,979 generations tested so far
More workflows
See all
GPT Image 2.5 Flare - fix one thing in an image without losing the rest
Change one detail in an AI image, like size, mood or pose, while the rest of the picture stays put
GPT Image 2.5 Flare · 2 min render
GPT Image 2.5 Flare - Gen Z slang taken literally
Turn slang like spill the tea, glazing and catfished into photos that show the words literally
GPT Image 2.5 Flare · 2 min render
GPT Image 2.5 Flare - old sayings turned into deadpan photos
Turn everyday sayings into realistic photos where one impossible thing happens and nobody reacts
GPT Image 2.5 Flare · 1 min render