Photograph the real product yourself on a phone, in daylight, against a plain background, then use AI to remove the background, build a styled scene around it and resize it for each channel. Shopify's built-in editor does much of this at no extra cost; standalone apps cost roughly $10-$20 a month. Never let AI redraw the product itself.
That last rule is the one that keeps a cheap workflow from becoming an expensive one. AI image tools are brilliant at backgrounds, shadows and props, and unreliable at the product: they soften logos, rewrite printed text, change proportions and shift colours. A photo that doesn't match what arrives produces returns, complaints and, on marketplaces, removed listings. So the AI works around your real photograph, never instead of it.
A home setup for under $60
You don't need a studio. You need soft, even light, a plain surface and a phone held still. Typical costs, which vary by shop and brand:
| Item | Rough cost | Why it matters |
|---|---|---|
| Your phone | $0 | Recent phone cameras are more than good enough for web images |
| Two sheets of white foam board | $5-$15 | One as the background sweep, one to bounce light into shadows |
| Small tripod with a phone clamp | $15-$25 | Same angle and distance for every product, so the range looks consistent |
| Bulldog clips or tape | $3 | Holding the sweep in a curve so there's no horizon line |
| Optional: pop-up light tent | $25-$60 | Helps with shiny items and dark days |
Set up beside a large window, out of direct sunlight, with the window to one side of the product and the second foam board on the other side to lift the shadows. Overcast days are your friend. Tape a small mark on the table so every product sits in the same spot.
Shooting the real product: six shots per item
AI can do a lot with a decent photo and very little with a poor one. Background removal struggles with blur and heavy shadows, and scene generation copies whatever lighting it sees. For each product, take the same six shots:
- Front, straight on, product centred and filling most of the frame.
- Three-quarter angle, which shows depth and usually becomes the lifestyle base.
- Back or underside, if it carries information (labels, sizes, fittings).
- A close-up of the texture, finish or print.
- Scale: the product next to a hand or a familiar object.
- Anything that opens, folds or comes apart, shown open.
Phone settings that help: switch on gridlines, don't use digital zoom (move the phone instead), tap and hold on the product to lock focus and exposure, and turn off any "beauty" or scene-enhancement mode that alters colours. Take a quick test shot of a white sheet of paper; if it looks blue or yellow on screen, fix the white balance before shooting the whole range.
Clean white-background images for shops and marketplaces
The first AI job is the simplest and most valuable: cutting the product out and placing it on pure white or a flat brand colour. Options that work on a small budget:
- Shopify's media editor. Shopify Magic can remove a background or replace it with a solid colour, detecting the product automatically with manual correction if it clips an edge. Background removal needs a PNG if you want transparency; other formats can't keep a transparent background.
- Photoroom. A dedicated product-photo app with background removal, AI backgrounds and batch editing on paid plans. Its help pages say the free plan can't be used for commercial purposes, so budget for a paid plan if the images will sell anything. Check the current price in your app store; the app and web pricing pages show different tiers.
- General design tools. Canva and similar tools include background removal on paid plans and are useful if you'll also make social graphics from the same photos.
If you sell on Amazon, the main image rules are strict: a pure white background (RGB 255, 255, 255), the product filling about 85% of the frame, nothing but the product, no text, logos or borders, and a real photograph rather than an illustration or mock-up. Check Amazon's current image requirements in Seller Central before you upload, because image rules are enforced by suppressing listings. AI background removal is a quick way to meet the white-background rule; generating the product is not.
Building lifestyle scenes around the real product
Lifestyle images, showing the product in use or in a setting, help buyers imagine owning it. This is where generative AI saves the most money, because a styled shoot with props and locations is expensive. The trick is to keep the cut-out product untouched and have the AI generate only the surroundings.
In Shopify's media editor, the Generate option transforms an image with a plain-language prompt, and Extend (outpainting) widens the background around the subject. Photoroom and similar apps have equivalent tools. A prompt pattern that works across tools:
Place the product on [surface: e.g. a pale oak packing bench] in
[setting: a bright small-business stockroom], soft daylight from
the left, matching the product's existing shadow. Keep the product
exactly as it is: do not change its shape, colour, text or logo.
Add [1-2 props: a roll of brown tape, tissue paper] behind it, out
of focus. Realistic photo, no text, no people.An illustrative first result for a printed mailer box, and what was wrong with it: the bench and props looked convincing, but the box appeared about the size of a shoebox next to a tape roll the size of a dinner plate, so the scale was wrong. The generated shadow fell to the right while the product's own shadow fell to the left. And the AI had "tidied" the printed logo on the box lid into a slightly different font. The fixes were to regenerate with "a standard 50 mm roll of packing tape" as the prop so scale had a reference, to add "shadow falling to the right, matching light from the left", and to switch to a tool mode that locks the cut-out product so the logo stayed pixel-for-pixel original.
One technical limit to know: Shopify's generated images are produced at about one megapixel, so a 3,000-pixel photo comes back much smaller. Use generated scenes for secondary images and social posts, and keep your full-resolution white-background shot as the main image.
A packaging supplier's 60-product catalogue on a $15 budget
An illustrative packaging supplier sells about 60 products (mailer boxes, shipping cartons, tissue, tape and void fill) through its own online shop. Its product images came from three different manufacturers' catalogues: some renders, some photos, different angles and background colours. The range looked like a marketplace of strangers, and two customers had complained that a box was "a different brown" from the photo.
The owner's plan, over two weekends:
- Shoot (about 6 hours): all 60 products with the foam-board setup by the warehouse window, six shots each where relevant. Boxes shot flat-packed and assembled.
- Clean up (about 3 hours): background removal in the shop's built-in editor, then a batch colour check against the real boxes held next to the screen in daylight.
- Scenes (about 4 hours): lifestyle scenes for the 15 bestsellers only, using the prompt pattern above with the brand's usual props.
- Cost: $20 of foam board and clips, a tripod the owner already had, and one month of a $15-a-month editing app for batch exports, cancelled afterwards.
What changed, in this illustration: every product page now shows the same angle, light and background, so the range looks like one company's. The "different brown" complaints stopped, because the photos are now of the actual stock under daylight rather than a manufacturer's render. And the owner had a reusable set: when a customer asks for an image for their own website, the supplier sends the white-background shot.
One request the owner turned down: AI mock-ups of customers' logos printed on boxes, generated without a real printed sample. They look impressive, but AI renders print more crisply than a two-colour print on kraft board ever comes out. The compromise was to show real printed samples from past orders (with those customers' permission) and label any AI mock-up clearly as a "digital preview, not a print proof".
Reflective and engraved items: a funeral director's urns
Shiny products are the hardest to photograph and the easiest for AI to spoil. An illustrative funeral director selling brass and ceramic urns online found two problems. First, the brass reflected the room: a window, a door and the phone itself appeared in the curve of each urn. Second, when an AI "enhance" feature was used to clean up the reflections, it also redrew the engraved sample text on the front, turning a neat serif name into slightly different letters.
The fixes were mostly photographic. Shooting inside the pop-up light tent surrounded the urn with white, so the reflections became soft white gradients instead of a room. Moving the phone further back and zooming by cropping afterwards kept the phone out of the reflection. After that, the AI was used only for background removal and a plain, soft grey backdrop, with the engraving untouched. The general rule for anything with text, engraving or fine detail: mask it, lock it, or leave it out of the AI's reach entirely, and compare it with the original photo at full size before publishing.
Small print on packaging: a laboratory's test kit
Products whose packaging carries text are a particular trap. An illustrative laboratory selling a home water testing kit tried a fully generated "kit on a kitchen worktop" image. The box looked right at a glance, but the printed text on it was nonsense: plausible-looking words that weren't words, and a contaminant list that didn't match the kit. At thumbnail size nobody would notice; at full size on a product page it looked careless, and for a lab selling accuracy, careless is fatal.
The working approach: photograph the real kit (box, sample bottle, instructions, prepaid envelope) laid out flat from above, remove the background, and generate only the worktop around it. The real label stays legible because it's a photograph. For the thumbnail, a white-background shot of the box front was clearer than any scene. It's a useful habit for any product with small print: look at your thumbnail on a phone, and if the text on the product is unreadable, crop tighter or choose a different shot.
What AI can rescue from a poor photo, and what it can't
Sometimes the only photo you have is a bad one: a supplier's snapshot, a shot taken in the warehouse under strip lights, or an old image of a product you no longer stock in the same colour. It helps to know which faults AI editing fixes reliably and which it only disguises.
- Cluttered background: fixes well. Background removal handles a messy bench as easily as a white sheet, provided the product's edges are sharp.
- Colour cast from artificial light: fixes well if you correct against the real product in daylight, not by eye on the screen alone.
- Slightly dark image: partly. Brightening recovers some detail, but dark areas turn grainy, and AI "enhancement" may invent texture that isn't there.
- Blur or camera shake: doesn't fix honestly. Sharpening tools redraw edges and fine print, which is exactly the product detail you mustn't change. Reshoot.
- Product cut off at the edge of the frame: don't use AI to extend it. Extending the background is fine; extending the product means the AI is guessing what the missing part looks like.
An illustrative case: a car repair garage selling a few refurbished parts online had one blurry photo of an alternator taken on the workshop floor. Background removal gave it a clean white backdrop in seconds, but the part number stamped on the casing was unreadable, and the "sharpen" tool turned it into different digits. The garage reshot it on a sheet of card by the workshop door in five minutes, and the new photo needed only the background removed.
Three budgets, three workflows
| Budget | Tools | Suits | Limits |
|---|---|---|---|
| $0 | Phone, window, foam board you already have; your shop platform's built-in editor or a free design tool | Under 30 products on Shopify or similar | Batch work is slow; free tiers may bar commercial use, so check each licence |
| About $15 a month | A paid product-photo or design app for batch background removal and scenes | 30-300 products, regular new lines | Credits or export caps on some plans |
| About $60 once, then $15 a month | Add a light tent, tripod and reflectors | Shiny, dark or detailed products | Still needs an hour a week of someone's time |
For comparison with paying a professional, including when a photographer is still the better choice, see AI product photos versus a photographer. For many small shops the answer ends up mixed: a photographer for the hero range once a year, AI-assisted phone shots for everything else.
Honesty rules for AI-edited product photos
The line is simple to state: edits that change the setting are fine; edits that change what the customer will receive are not. A quick guide:
| Usually fine | Crosses the line |
|---|---|
| Removing or replacing the background | Changing the product's colour, shape or size |
| Generated props and surfaces around the product | Adding features, accessories or finishes it doesn't have |
| Correcting white balance to match the real item | "Enhancing" the product's texture or finish beyond reality |
| Extending the frame to fit a banner | Showing custom print quality you can't produce |
| Clearly labelled mock-ups and previews | Mock-ups presented as photos of real products |
Shopify adds an invisible watermark to images generated in its editor, which doesn't restrict commercial use and can't be removed, and some social platforms detect or ask you to label AI-generated content. The practical detail of labelling on social media is covered in how to label AI-generated images and video, and the legal side of using AI images in marketing in whether you can legally use AI-generated images.
A pre-publish check for every product image
Run this on each image before it goes live. It takes about a minute per product and catches most of the failures above.
- Colour: hold the real product next to the screen in daylight. Does it match?
- Shape and details: compare with your original photo at full size. Logos, text, engraving and edges unchanged?
- Scale: is there a size cue somewhere in the set, and are generated props realistically sized?
- Nothing extra: no props that look like they're included in the sale, unless they are.
- Lighting logic: shadows and highlights fall the same way on the product and the scene.
- Thumbnail test: on a phone, is the product recognisable and any key text readable?
- Channel rules: main images meet marketplace requirements; file size and dimensions are right.
- Alt text: a short, plain description for screen readers and search, such as "kraft mailer box, 250 × 180 × 80 mm, assembled".
Once the photos are right, the product pages usually need the same treatment in words. Writing product descriptions with AI that sell uses the same principle as this tutorial, facts first and AI second, and the images and copy work best when they're done together. For social posts built from the same product shots, creating on-brand social graphics with AI picks up where this leaves off.
Product photo questions from small sellers
Should I use AI-generated models to show clothing or accessories?
They can help show how an item is worn when you can't afford a model shoot, and some product-photo apps now offer them. The risk is that the garment's fit, drape and colour are partly generated, so the image may not match what arrives. Keep at least one real photo of the actual item, check the AI version against it, and say clearly that the model image is AI-generated.
Which file size and shape should I export?
Export square images of at least 2,000 pixels on the longest side for your shop and marketplaces, then let each platform resize. Check each platform's current minimum and zoom requirements. Keep a master copy at full resolution, because AI background tools sometimes output smaller files; Shopify's generated images, for example, come out at about one megapixel.
Do I own AI-edited product photos?
You own your original photograph, and most tools' terms let you use edited outputs commercially on paid plans. Free tiers can differ: Photoroom's help pages say its free plan can't be used for commercial purposes. Read the licence of each tool you use, and keep your original photos so you can re-edit elsewhere if a tool changes its terms.
Further reads
- Is Midjourney Worth It for Small Business Marketing Images? — When a paid image generator is worth it for marketing visuals.
- Taking a Local Shop Online With AI: Product Listings in a Weekend — Photos and listings for a whole range in one weekend.
- How to Write Shopify Product Descriptions With AI That Convert — Descriptions that match the new photos on Shopify.
- Can a Small Online Shop Use AI to Compete With Bigger Brands? — Where AI helps a small shop compete with bigger brands.
- Online Shop AI Mistakes That Hurt Trust and Conversions — Other AI shortcuts that quietly cost online shops trust.
- AI Virtual Staging for Property Listings: Costs and Disclosure — How another sector handles disclosure of AI-edited images.
- How Much Does It Cost to Add AI to a Shopify Store? — Adding AI to Shopify can cost $0 extra, or $35 to $500+ a month with apps. Itemised prices, three store budgets and the costs listings don't show.
- How to Use AI for Etsy Titles, Tags and Descriptions — Search phrases first, then prompts for each field: a title under 15 words, 13 non-repeating tags of 20 characters or fewer, and a description buyers can act on.
- Best AI Tools for Etsy Sellers, Ranked by Time Saved — Seven kinds of AI tool ranked by the hours they save a small Etsy shop each week, with prices, a worked example for each, and the tools not worth paying for.
- How to Write eBay Listings With AI That Sell Faster — Sold-price research, a title formula, item specifics and flaw-honest descriptions for eBay listings drafted with AI, plus the policy traps to avoid.
- How Resellers Use AI to List Second-Hand Items Faster — A second-hand listing workflow that uses marketplace AI for the drafting and keeps measurements, flaws and photos honest.
- AI Model Shots for Boutiques: Try-On Images Without a Photoshoot — How AI on-model images work, what tools like FASHN and Photoroom cost, which garments come out wrong, and how to check and label every image honestly.
- How to Design Flyers and Printed Marketing With AI — Let AI draft copy and layout options, keep it away from small text, and do the print checks yourself: size, bleed, resolution, colour and QR code.
- How to Use Shopify Sidekick to Run Your Store Faster — What Shopify Sidekick can and can't do, reorder suggestions after Stocky, ten requests that save store time, and the checks before you approve a change.
- How to Check AI-Generated Images for Errors Before Posting — An eight-point checklist and four viewing tricks for catching AI image mistakes, with a filled-in check and a campaign of eight images put through it.
- Brand Guidelines for AI Images: Colours, Style and Limits — Write the image page of your brand guidelines: colours in hex and words, one style with reference images, and the limits AI images must never cross.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Shopify help pages (media generation in the file editor); Photoroom pricing page and help pages (plans; free plan commercial-use restriction); marketplace image guidance for main images (white background, product filling the frame).