An AI photoshoot uses an image model instead of a camera to produce a set of product photos. You upload one clean source photo, then generate the angles, backgrounds and scenes a listing needs. Shopify puts a camera shoot at $50 to $350 an image. Every result still has to match the product that ships.
Search for the term and the results show two different products. Some results generate a portrait of a person from selfies. Others generate a product on a backdrop from a packshot. As of September 2026, Google’s AI overview for the query covers portraits and product images in one answer.
They are two different jobs. The input differs, the cost differs, and since July 2026 the disclosure rules differ too. This guide covers the product side: what the workflow is, what it costs, what Amazon and Shopify require, and where it still fails.
Key Takeaways
One source photo, many results. The model starts from the product you upload and re-lights, re-angles and re-stages it. That is why the source shot decides the ceiling.
The model sets the cost. Each model costs a different number of credits, so the model you pick for each shot decides the cost of the set.
Amazon’s image guide has no rule against AI product images. Its AI rule covers photorealistic AI-generated people. The rule that matters is older and simpler: images must accurately represent the product as a realistic, professional-quality image, and the product should fill 85% of the frame (sellercentral.amazon.com, September 2026). For clothing, Amazon says only photos are allowed.
The July 2026 disclosure duty covers people. Amazon’s synthetic-performer tag applies to photorealistic people generated entirely by AI. A bottle on a backdrop does not trigger it.
We found no image-accuracy rule in Shopify’s Acceptable Use Policy, as of September 2026. Its published image requirements are dimensions and file size, and its terms make you responsible for your store content. What constrains you there is consumer law and your return rate.
Images are the first thing shoppers look at. 56% of test subjects began by exploring product images before reading titles or descriptions (baymard.com, October 2026).
What is an AI photoshoot?
An AI photoshoot is a session where an image model produces the photographs, working from a reference you supply rather than from a camera and a set. You upload one clean shot of the product. The model generates the additional angles, backgrounds, lighting setups and lifestyle scenes that a listing or a campaign needs. The output is a set of files, not a single image. For a first attempt, a beginner’s guide to AI for product photography starts with one product and one edit.
The word “photoshoot” is doing real work in that sentence. A single generated image is not a photoshoot. What brands need is a coherent set: a white-background main image, several detail angles, something in scale, and a styled scene or two. All of them look like the same product, photographed on the same day.
The two things people mean by “AI photoshoot”
These two jobs share a name and almost nothing else. Sorting which one you are doing decides your tool, your budget and your legal exposure.
| Photoshoot of a person | Photoshoot of a product | |
|---|---|---|
| Input | Selfies or a face reference | One clean packshot |
| Typical use | Headshots, portraits, profile pictures | Listings, PDPs, ads, catalogs |
| Buyer | Mostly consumer | Brands, sellers, agencies |
| Disclosure duty | Yes, if photorealistic and synthetic | No, a product is not a performer |
| Accuracy risk | Low, it is your own likeness | High, it has to match what ships |
The disclosure row is the one that changed. In July 2026, Amazon told sellers to tag some images before upload. The tag applies to images that show a photorealistic person generated entirely by AI, and the keyword is “contains-synthetic-performer” (Amazon Seller Central, product image guide, September 2026). CNBC reports the change followed New York’s synthetic performer law (cnbc.com, September 2026). That law has applied since 9 June 2026. It defines a synthetic performer as a digital asset made to look like a human performing, where the person is not recognizable as any real performer. It sets civil penalties of $1,000 for a first violation and $5,000 for each later one (nysenate.gov, September 2026). This is general information, not legal advice.
Read the trigger carefully, because it is easy to misread. The duty attaches to a synthetic human. Generating a marble backdrop behind a candle does not create a performer. If your shoot has no people in it, this rule does not reach you.
How to do an AI photoshoot, step by step
The sequence below is the product workflow. It assumes you have the physical item and a phone. For the words to type at each step, see the six-line brief for AI product images.
- Shoot one clean source photo. Even lighting, plain background, product sharp and fully in frame. This single file sets the quality ceiling for everything downstream, which is why it is worth the twenty minutes the source shot deserves. How to take product photos without a studio gives a pass or fail check for each step.
- Remove the background. You want the product isolated before you start staging it, so the model composes around it rather than fighting the original room.
- Generate the main image first. Pure white, product filling the frame, no props. This is the most constrained shot and the one with the least creative latitude.
- Generate the angles. Front, back, side, top, detail, in scale. A listing needs several specific angles before it is complete, and generating them together keeps the lighting consistent across the set.
- Generate the styled scenes last. Lifestyle and context shots have the widest tolerance, so they are the cheapest place to experiment once the required shots are done. AI lifestyle photo generators compares the tools for this shot.
- Check every file against the physical product. Color, logo, texture, proportions, count. This step is not optional, and a person has to do it. A product photography checklist lists the checks stage by stage.
Step six is where a set fails review. The model makes a plausible image. It does not check that image against your product, so a label can drift, a seam can move, and a strap can gain a buckle it never had. The AI image retry rate shows how many generations one usable image takes.
What an AI photoshoot costs
On DesignerBox each model costs a different number of credits, so the model you pick sets the cost of the run. The cost is shown before the run, so you check a six-shot set against your plan first. Route the plain matte goods through a lower-cost model. Save the higher-cost one for the shots that carry a label.
Two plan gates matter more than the credit math. Uploading your own photos and the commercial license start on the Pro plan. AI video, virtual try-on, upscaling, the image editor and the video editor start on the Premium plan. Plans and credits are on the pricing page. The free plan runs on sample products, and its results carry a watermark. It is still useful for judging the look before you plan a live listing.
Set that against a camera. Shopify puts product photographers at $500 to $3,000 a day, or $50 to $350 for each product shot (shopify.com, September 2026). A six-shot set at those rates is $300 to $2,100 for one product. The comparison is only rough, because a studio day covers many products and delivers files you never have to fact-check. Generation moves the cost from the shoot to the review. For a fuller breakdown of the studio side, including the reshoot cycle that inflates it, see what a product photoshoot actually costs.
What the marketplaces require
Amazon’s AI rule covers images of AI-generated people, and Shopify’s image requirements do not mention AI. The difference between them is worth knowing before you plan a set.
Amazon’s general constraints are about the image. All images must accurately represent the product as a realistic, professional-quality image. The product should fill 85% of the frame. Main images need a pure white background at RGB 255,255,255, with no text, logos, borders or watermarks, and the product shown outside its packaging. Files run from 500 to 10,000 pixels on the longest side, and 1,000 pixels or more turns on zoom. Amazon recommends a main image plus at least six more images and one video (sellercentral.amazon.com, September 2026). For clothing, Amazon’s clothing guide says only photos are allowed, and it does not say whether a photorealistic AI image counts as a photo.
Shopify publishes dimensions and nothing about truthfulness. Square images display best at 2048 by 2048 pixels, up to 5000 by 5000 or 25 megapixels, under 20MB, with PNG preferred and JPEG next (help.shopify.com, September 2026). We found no accuracy clause in its Acceptable Use Policy, as of September 2026, and its Terms of Service make you responsible for your store and your content (shopify.com/legal/terms, September 2026). On your own store the limits are consumer protection law and the returns your images cause. We cover what that means on a Shopify store separately.
The accuracy rule is the one to design around. It applies whether the image came from a camera or a model, and it is the reason step six above exists.
Where generation still breaks
Failure is not random across a catalog. It tracks the material, because the properties that are hard to render are physical ones: transparency, reflectivity, fine repeating texture, and precise text on a label.
Clear glass, mirrored metal, knitted fabric and anything carrying dense small type are the categories that need the most correction, and sometimes need a camera. Matte, opaque goods with simple shapes give the cleanest results. The practical move is to sort the catalog by material rather than by product category and route only the reliable half through generation on the first pass.
Model choice helps at the margins. Nano Banana Pro generates at 1K, 2K and 4K and accepts up to six high-fidelity object images (ai.google.dev, September 2026). Nano Banana 2 takes up to ten. OpenAI says GPT Image models “can still struggle with precise text placement and clarity” (developers.openai.com, September 2026), which is a useful warning for label work. What each model is best for sits on the model list.
When to book a photographer instead
Generation is the wrong call in three situations, and saying so is cheaper than a suppressed listing. Our guide on whether AI will replace photographers splits the work into capture and variation.
Book a camera when the product is the hard-material case above and the shot is the main image, because that is the slot with the least tolerance for drift. Book one when you are launching a hero product whose photography is the brand, where the point is a specific art direction rather than coverage. And book one when you do not yet have a usable source photo, since generation cannot rescue a bad original. It can only repeat its flaws.
The realistic pattern for most catalogs is both. Photograph the hero products and the awkward materials. Generate the long tail, the angle variants, the seasonal restages and the ad crops. To size the split, count the images a catalog photoshoot needs before you book anything.
A workflow for the long tail
A DesignerBox product photography workflow fits that split. Build the pass once on one product and check it against the physical item. A saved workflow runs the same way on the next SKU, so you do not start from a blank prompt every time. The source photo, the brand rules, the run, the edit and the finished files stay in one place. An approved shot never leaves the workspace to be rebuilt somewhere else.
Anyone can make an AI picture. Making hundreds that still look like your brand is the hard part. You set the brand once, and the workflow reads it on every run.
There is a free plan, and it runs on sample products. Start from a template before you plan a catalog around it.
FAQ
How do you do an AI photoshoot?
Upload one clean, evenly lit photo of the product, remove its background, then generate the main image, the required angles, and the styled scenes in that order. Check every output against the physical item for color, logo and proportion before you publish. The source photo sets the quality ceiling for the whole set.
Can I do an AI photoshoot for free?
Partly. DesignerBox has a free plan. It runs on sample products, which is enough to judge the look, and its results carry a watermark. Before you rely on any tool that advertises a free session, check two points. The file must be clean enough for a live listing. The license must let you sell with it.
Are AI product photos allowed on Amazon?
Amazon’s general image guide does not ban AI-generated product images, as of September 2026. Its images must accurately represent the product, and main images must meet the white-background, 85% frame fill and no-text rules, however the image was produced. Amazon allows only photos for clothing. Since July 2026, images showing a photorealistic person generated entirely by AI must carry the contains-synthetic-performer tag.
What does an AI photoshoot cost?
The model you pick sets the cost. On DesignerBox each model costs a different number of credits, and the cost is shown before the run. Shopify puts traditional product photography at $50 to $350 an image, or $500 to $3,000 a day (shopify.com, September 2026).
How many photos does a product listing need?
Amazon’s image guide recommends a main image, at least six more images and one video (sellercentral.amazon.com, September 2026). Practically, a complete set is a white-background main image, three or four angles, one detail shot, and something that shows scale. Baymard found 42% of test subjects tried to judge a product’s size from its images, so the in-scale shot earns its slot (baymard.com, October 2026).
Will the images look fake?
The material matters more than the model. Matte, opaque products with simple shapes look convincing. Clear glass, polished metal, fine knits and dense label text are where artifacts appear. Sorting the catalog by material before you start is what keeps the visible failures out of the set.
Is an AI photoshoot the same as an AI headshot?
No. They share a name and little else. A headshot generates a person from selfies and, when photorealistic, can trigger disclosure duties on some platforms. A product photoshoot generates objects from a packshot and carries an accuracy obligation instead.
Sources
- Amazon image requirements, including the accuracy rule, 85% frame fill, main-image prohibitions, the 500 to 10,000 pixel range, the 1,000 pixel zoom threshold, the image recommendation and the contains-synthetic-performer tag: Amazon Seller Central, product image guide, September 2026
- Amazon’s July 2026 notice to sellers and its link to New York law: cnbc.com, September 2026
- New York S.8420-A, the synthetic performer definition and the $1,000 and $5,000 penalties: nysenate.gov, September 2026
- Shopify merchant responsibility for store content: Shopify Terms of Service, September 2026
- Shopify product image dimensions, the 25 megapixel and 20MB limits and accepted formats: help.shopify.com, September 2026
- Product photography day rates and per-image rates: shopify.com, September 2026
- Shoppers exploring product images before titles or descriptions, and gauging scale from images: baymard.com, October 2026
- Nano Banana Pro and Nano Banana 2 resolutions and object reference limits: ai.google.dev, September 2026
- GPT Image text placement limitation, quoted verbatim: developers.openai.com, September 2026
- DesignerBox plan gates and the watermark on the free plan: DesignerBox pricing page (designerbox.ai/pricing), September 2026
Marketplace requirements, model specifications and photography rates verified from platform, vendor and research documentation as of September 2026. The Baymard figures were re-checked on 2 October 2026. Platform image policies change without notice. Verify current requirements on each marketplace before publishing a listing. This is general information, not legal advice. Individual results vary.
