Product imagery used to mean a studio, a photographer and a day per product. CGI changed the economics for some products and not for others, and generative tools have moved the line again. This guide gives you the decision, not the sales pitch.
01When CGI wins
- The product doesn't exist yet: launch imagery from CAD before manufacturing
- Many variants: every colour, finish and configuration from one model, perfectly consistent
- Impossible angles and cutaways: inside the mechanism, exploded views, scale comparisons
- A catalogue that must look uniform: same light, same angle, same floor across hundreds of items
- Changes later: a new colourway next quarter is a re-render, not a reshoot
- Environments you cannot reach or build: the product on a glacier, in a server hall, at sea
02When photography still wins
- Materials where authenticity is the sale: leather grain, food, fabric drape, hand-finished surfaces can be rendered, but a camera is faster to truth
- People using the product, where a real hand and a real expression carry the message
- Lifestyle context that has to feel found rather than built
- Very low product counts with simple needs, where a half-day shoot is cheaper than modelling
03The hybrid that most brands end up with
A photographed hero image for the emotional moment; rendered variants, angles and technical views for everything else; and generative tools to extend backgrounds, create settings and produce campaign variations around both. The question is rarely either/or; it is which part of the catalogue gets which treatment.
04What you need to provide
- For CGI: CAD files (STEP, IGES or native formats) if they exist; otherwise good reference photographs from several angles and physical samples for materials
- Material references: swatches, finish specifications, or photographs of the real surface under neutral light
- Brand guidelines: backgrounds, lighting mood, angle conventions
- For photography: the products themselves, props, and a shot list agreed in advance
- For both: where the images will run and at what sizes, so they are made at the right resolution the first time
05Questions to ask before you choose
- How many variants will exist over the product's life?
- Will the product change after launch?
- Does the material or the human moment carry the sale?
- Do we have CAD, or would the model have to be built from photographs?
- Where will these images run — a catalogue grid, a hero banner, print?
In short
- CGI wins on variants, consistency, impossible views and products that don't exist yet.
- Photography wins where material authenticity or real people carry the sale.
- Most brands end with a hybrid: a photographed hero, rendered everything else, generative extensions around both.
- CAD files and material references are what make renders fast and faithful.
Questions
Can you render from photographs if we have no CAD?
Often, yes, for products with simple geometry. Complex products are better modelled from CAD; we will say which before quoting.
Will renders look fake?
Badly lit ones do. Well-made renders with real material references are routinely indistinguishable from photographs, which is why so much product imagery you already see is rendered.
Can generative AI replace both?
For backgrounds, settings and campaign variation it already does much of the work. For exact product fidelity, a render or a photograph remains the anchor that generative tools work around.
Sounds like your problem?
Tell us about it. We'll say honestly whether the swarm can help, and what it would take.
More guides