AI product imagery or a studio shoot — which should you use?
Both, for different jobs. The useful question is not which is better, but which failures you can live with on a given shot.
We use AI for parts of the imagery and video we make, and we photograph things when photography is the answer. That makes this an awkward page to write honestly, so here is the disclosure first: we are not neutral, we sell both, and the reason we do both is that neither one covers the whole job.
Where generated imagery genuinely wins
Volume across variants. A catalogue with forty colourways is a scheduling problem before it is a creative one, and the ninetieth frame is where a studio day quietly stops being affordable. Generated imagery does not get tired at frame ninety, and holding one grade across a whole range is easier when the range was made in one pass.
It also wins on worlds. Putting a bottle in a setting that does not exist, or a location you would need three days and a permit to reach, becomes a decision rather than a budget line.
Where it still fails
- Anything a customer will hold up against the real product. Stitching, weave, closures, hardware, the specific way a fabric falls. This is where generated frames go subtly wrong, and subtly wrong is worse than obviously wrong.
- Text on the product. Lettering on packaging, a logo on a shoe, a label. The answer is to composite the real artwork, never to generate it.
- Anything sold on provenance. If the story is that a person made it by hand over fourteen weeks, a generated photograph of that person's hands is not a shortcut, it is a lie about the only thing you were selling.
- Faces you intend to present as customers. Not a craft limit, an honesty one.
The line we do not cross
We never produce synthetic testimonials or fake customer reviews, and creator-style content is always presented as brand content. This is not squeamishness. Fabricated endorsements are treated as fake reviews by regulators, and the exposure sits with the brand that ran them, not the studio that made them.
The perception problem nobody mentions
There is a measurable trust penalty on advertising that reads as AI-made, and it is worst on considered purchases — the couture, fragrance and audio end of the market, which is most of what we work on. That is a reason to be careful about where you use it, not a reason to hide that you did. A brand caught being coy about it pays twice.
Our position is boring on purpose: a person directs every frame, and if a shot has to be photographed we say so and photograph it. What you are buying is the direction and the result.
How to decide, shot by shot
Ask what the frame has to survive. If it will be compared against the product in someone's hands, photograph it. If it carries the brand's central claim, photograph it. If it is the fortieth colourway of a shoe on a plain ground, generating it is not a compromise — it is the correct use of everyone's money.
If this is the problem you have, send the brief. First concepts back in 48 hours, watermarked so you can judge the work before anyone talks about money.