100 AI product ad prompt codes for ecommerce
Turn one approved product photo into a controlled set of static ads and short vertical videos with 100 reusable prompt codes, QA rules, and a full-funnel test plan.

Lamina Team
Product Team @ Lamina

Do not spray 100 unrelated prompts at one product photo. Build a 10×10 matrix instead: hold the approved product steady in every render, give each variation one creative job, then turn the strongest stills into short 9:16 motion sequences.
Use the source image as the product anchor, not a loose visual cue. AdLibrary says to audit the input before generation: details that are missing, obscured, or pinned against a cluttered edge are hard to rebuild faithfully; Inkfox likewise recommends image-to-image transformation where the existing item must stay recognizable. That rule governs every code below. Keep the shape, material, color, logo, label, proportions, and packaging unchanged; alter the scene, framing, or message.
These 100 codes are production prompts. They do not promise that all 100 variants will win. A useful creative test needs controlled variation: Gestel advises changing one decision at a time, while Oakgen’s concept matrix covers clean hero, lifestyle, problem, feature, comparison, bundle, seasonal, and premium-minimal routes. Treat the codes as deliberate hypotheses, not a stack of decorative backdrops.
| Metric | Value | Source |
|---|---|---|
| AI assets generated on Lamina (last 30 days) | 304 | Lamina platform telemetryas of 2026-08-18 |
| Median time to generate an asset | 225s | Lamina platform telemetryas of 2026-08-18 |
| 90th-percentile generation time | 472s | Lamina platform telemetryas of 2026-08-18 |
| Recommended video hook window | 0–3 seconds | wavespeed.aias of 2026-02-26 |
| Recommended video proof window | 3–10 seconds | wavespeed.aias of 2026-02-26 |
| Recommended motion limit | 1 motion per beat | wavespeed.aias of 2026-02-26 |
What must every AI product ad prompt include?
Every AI product ad prompt needs a product-preservation rule, scene direction, output job, format, lighting, and explicit exclusions. Brandgene’s AI photoshoot brief adds campaign goal, brand context, channel or crop, shot type, and quality rules; Gestel’s product-still guidance calls for subject, camera or lens, lighting, setting, and mood. Leave one of those fields open and you risk a handsome asset that cannot run on a PDP, Meta placement, TikTok, or YouTube Shorts.
Put this master wrapper ahead of every code: “Use the uploaded product photo as the exact product reference. Preserve [PRODUCT SHAPE], [MATERIAL], [COLOR], [LOGO], [LABEL TEXT], [PACKAGING], and [PROPORTIONS]. Do not add, remove, recolor, or redesign any product element. Create [OUTPUT JOB] for [AUDIENCE] on [CHANNEL], aspect ratio [RATIO]. Brand world: [PALETTE], [STYLE], [TONE]. Show [ONE BENEFIT OR PROOF]. Lighting: [LIGHT]. Composition: [FRAMING] with [COPY SPACE]. Exclude extra products, invented text, distorted labels, altered hands, and unsupported claims.”
Fill the bracketed fields from an approved brief. Do not cram three benefits, two offers, a creator script, and a seasonal set into one request. One prompt gets one sellable visual decision.
How do you turn one product photo into a 100-ad test system?
Audit the anchor image
Begin with a high-resolution, uncropped master where you can inspect the product silhouette, material, color, logo, and label. Mark reflective edges, translucent packaging, tiny legal copy, and every detail that cannot drift. AdLibrary calls the source photo the quality ceiling. Review starts before generation.

Write one campaign card
Set the audience, one approved benefit or proof point, offer, brand palette, required ratio, and QA rules. Brandgene recommends these inputs for an AI photoshoot brief because a product reference by itself does not create a usable campaign direction.

Choose one cohort variable
Build a cohort around one choice: setting, use case, visual angle, offer framing, or format. Hold everything else fixed. Gestel recommends one-decision-at-a-time testing so you can read the result instead of guessing at it.

Generate stills before motion
Make the static concept first. Approve product truth, then use that still as the anchor for a Reel or Short. Videotok recommends varying hooks, scenes, proof points, and formats around a recognizable source image that leaves room for motion.

Review and publish only approved variants
Check the logo, label, color, material, dimensions, claims, crop, and safe space. Cliprise warns that label text is especially fragile in product-photo-to-video workflows, particularly as motion complexity climbs.

Which prompt codes create clean hero ads?
Use codes 01–10 when the product is the message. They serve as the campaign control group: a clear item, disciplined light, and enough negative space for a platform-native headline.
01 Clean pedestal: “Create a premium clean hero of [PRODUCT] on a [BRAND COLOR] pedestal, soft directional studio light, front three-quarter angle, 4:5, headline space above.” 02 White catalog: “Place [PRODUCT] on pure white with a natural contact shadow, straight-on packshot, 1:1, no props.” 03 Color sweep: “Show [PRODUCT] on a curved [BRAND COLOR] paper sweep, crisp side light, 4:5, copy space left.” 04 Shadow graphic: “Create a minimal hero of [PRODUCT] with a hard-edged window shadow, neutral ground, 1:1.” 05 Floating product: “Suspend [PRODUCT] against [PALETTE], realistic shadow below, centered 9:16.” 06 Premium black: “Stage [PRODUCT] on matte black with controlled rim light and restrained reflections, 4:5.” 07 Monochrome set: “Set [PRODUCT] in a tonal [BRAND COLOR] scene, preserve all pack colors, 4:5.” 08 Macro hero: “Frame the [MATERIAL OR FEATURE] of [PRODUCT] in a macro detail, exact texture, 1:1.” 09 Editorial crop: “Create an asymmetrical editorial hero of [PRODUCT], soft studio light, generous right-side copy space, 4:5.” 10 Shelf-ready: “Show [PRODUCT] upright, full pack visible, isolated against a pale background, 1:1 PDP-compatible.”
Which prompt codes show the product in use?
Use codes 11–20 to make the use case tangible without changing the SKU. Lifestyle earns its slot when it shows scale, routine, context, or a buyer-recognizable moment.
11 Morning routine: “Place [PRODUCT] in a realistic [AUDIENCE] morning routine at [LOCATION], product fully visible, 4:5.” 12 Desk solve: “Show [PRODUCT] solving [USE CASE] on a tidy desk, natural window light, 4:5.” 13 Travel moment: “Stage [PRODUCT] in a carry-on travel context, accurate scale, overhead composition, 1:1.” 14 Gym bag: “Show [PRODUCT] beside a gym bag after [USE CASE], energetic daylight, 9:16.” 15 Kitchen counter: “Place [PRODUCT] on a lived-in kitchen counter during [USE CASE], clean focal plane, 4:5.” 16 Bathroom shelf: “Show [PRODUCT] on a bright bathroom shelf, label facing camera, soft reflection, 4:5.” 17 Commute: “Create a city-commute scene with [PRODUCT] held naturally, product remains sharp, 9:16.” 18 Nightstand: “Place [PRODUCT] beside a nightstand for [ROUTINE], low warm light, 4:5.” 19 Outdoor utility: “Show [PRODUCT] used during [OUTDOOR ACTIVITY], exact packaging protected from distortion, 9:16.” 20 Gift moment: “Show [PRODUCT] being presented as a gift, no altered packaging, premium home setting, 4:5.”
Which prompt codes make a problem and proof visible?
Use codes 21–30 for awareness and consideration creative. Stage the buyer’s friction or one approved proof point. Do not fabricate a medical, performance, or comparison claim.
21 Problem scene: “Depict [AUDIENCE] encountering [PAIN POINT], with [PRODUCT] visible as the approved solution, 4:5.” 22 Friction close-up: “Create a close-up of [USE CASE FRICTION], then place [PRODUCT] cleanly in foreground, 9:16.” 23 Before-and-after layout: “Create a split visual illustrating [APPROVED BEFORE] and [APPROVED AFTER], preserve product exactly, 4:5.” 24 Ingredient proof: “Build a macro still life around [APPROVED INGREDIENT OR MATERIAL] and [PRODUCT], clinical-clean light, 1:1.” 25 How-it-works: “Show [PRODUCT] beside a simple visual demonstration of [MECHANISM], no invented text, 4:5.” 26 Scale proof: “Place [PRODUCT] beside [COMMON OBJECT] for accurate scale, neutral set, 1:1.” 27 Feature callout: “Frame [PRODUCT] with visual emphasis on [FEATURE], clean background, left copy space, 4:5.” 28 Durability moment: “Show [PRODUCT] in [APPROVED CONDITIONS] to illustrate [APPROVED PROOF], 9:16.” 29 Routine proof: “Show the exact moment [PRODUCT] enters [ROUTINE], one action only, 4:5.” 30 Objection answer: “Stage [PRODUCT] beside a visual response to [BUYER OBJECTION], premium minimal composition, 4:5.”
How do you prompt material, detail, comparison, and bundles?
Codes 31–40 are for shoppers who know the category already and need a reason to inspect the product. Keep the frame literal. Product detail beats an invented badge.
31 Texture macro: “Create a macro crop of [PRODUCT MATERIAL], exact grain and finish, studio side light, 1:1.” 32 Closure detail: “Show the [CAP, ZIP, BUTTON, APPLICATOR] of [PRODUCT] in close-up, exact construction, 4:5.” 33 Label detail: “Create a sharp close-up of [PRODUCT] label, preserve every visible character without generating new copy, 4:5.” 34 Color range: “Arrange approved [PRODUCT] color variants in an even row, preserve each exact colorway, 4:5.” 35 Size range: “Show approved [PRODUCT] sizes together with accurate proportions, white background, 1:1.” 36 Bundle flat lay: “Arrange the approved [BUNDLE ITEMS] in a clean flat lay, every SKU identifiable, 4:5.” 37 Build-a-kit: “Show [PRODUCT] with approved complementary items in a [ROUTINE] kit, 4:5.” 38 What’s included: “Create a top-down layout of only included items, clean separation, no additions, 1:1.” 39 Comparison frame: “Show [PRODUCT] next to a neutral representation of [OLD METHOD], no competitor marks, 4:5.” 40 Premium minimal: “Place [PRODUCT] with one relevant material prop, deep shadow, restrained luxury styling, 4:5.”
Notice how it says the help of AI, not with AI.
Which prompt codes cover social proof, offers, and seasonal ads?
Alex Cooper, Adcrate co-founder, makes a useful distinction: use AI as an assistant; do not let it dictate the creative. Codes 41–50 keep the approved product and approved message in charge, using visual variation to produce conversion-ready assets.
41 UGC frame: “Create a creator-style vertical frame of [AUDIENCE] holding [PRODUCT], natural phone-camera feel, product label clear, 9:16.” 42 Review card: “Show [PRODUCT] beside a blank testimonial card area, brand palette, 4:5; add no fabricated review.” 43 Social-proof scene: “Place [PRODUCT] in a believable [AUDIENCE] routine, candid composition, 9:16.” 44 Founder-style setup: “Create an authentic desk or workshop scene featuring [PRODUCT], vertical framing, 9:16.” 45 Offer hero: “Create a clear offer-led hero for [PRODUCT], preserve copy space for approved offer text, 4:5.” 46 Limited-drop: “Stage [PRODUCT] in a restrained limited-release visual, brand color background, 4:5.” 47 Gift guide: “Show [PRODUCT] in a gift-guide composition, seasonal accents, exact pack intact, 4:5.” 48 Summer use: “Place [PRODUCT] in an approved summer context, bright natural light, 9:16.” 49 Winter ritual: “Stage [PRODUCT] in a winter routine with subtle seasonal materials, 4:5.” 50 Event edit: “Create an [EVENT] campaign image for [PRODUCT], use only approved event language, 4:5.”
Which prompt codes fit awareness, consideration, and conversion?
Funnel stages need different ads, not just fresh backgrounds. David Ishag’s guidance separates pain points and belief challenges for awareness, differentiation for consideration, and offers or social proof for conversion; codes 51–60 turn that order into a usable brief.
51 Awareness pain: “Lead with a visual of [PAIN POINT], introduce [PRODUCT] as the approved answer, 9:16.” 52 Awareness belief: “Challenge [COMMON BELIEF] through a visual contrast, product visible, no unsupported claim, 4:5.” 53 Awareness category: “Show [PRODUCT] in the moment [CATEGORY NEED] arises, 9:16.” 54 Consideration differentiator: “Focus entirely on [ONE DIFFERENTIATOR] of [PRODUCT], macro-plus-pack composition, 4:5.” 55 Consideration routine: “Show how [PRODUCT] fits step [NUMBER] of [ROUTINE], 4:5.” 56 Consideration comparison: “Visualize [APPROVED DIFFERENCE] between [PRODUCT] and [OLD METHOD], 4:5.” 57 Conversion offer: “Create a direct response product hero with reserved space for [APPROVED OFFER], 4:5.” 58 Conversion bundle: “Feature the approved [BUNDLE] with clear product hierarchy, 4:5.” 59 Conversion proof: “Place [PRODUCT] beside approved proof asset space, no invented numbers or ratings, 4:5.” 60 Conversion reminder: “Create a minimal retargeting frame with [PRODUCT] large in view and CTA safe space, 9:16.”
What you actually need is different ads for different stages of the customer journey.
Which prompt codes convert static product ads into Reels and Shorts?
Codes 61–70 follow a tight hook, proof, CTA structure. Wavespeed recommends the hook at 0–3 seconds, proof from 3–10 seconds, and a final 2–3-second CTA; use one motion per beat to limit jitter and label deformation.
61 Pedestal reveal: “9:16, 8 seconds: 0–2s camera rises to [PRODUCT]; 2–6s slow orbit showing [FEATURE]; 6–8s hold with CTA safe space.” 62 Macro push-in: “9:16, 7 seconds: hook macro texture; proof slow push to full [PRODUCT]; final hold for CTA.” 63 Routine handoff: “9:16, 8 seconds: show [PAIN POINT] for 2s; one hand introduces [PRODUCT] for 4s; final product hold.” 64 Ingredient transition: “9:16, 8 seconds: [INGREDIENT] detail; single match transition to [PRODUCT]; CTA hold.” 65 Before-to-after: “9:16, 8 seconds: approved before visual; one wipe to approved after; product end card.” 66 Unboxing: “9:16, 7 seconds: hand opens approved packaging; reveal [PRODUCT]; stationary CTA end frame.” 67 Scale reveal: “9:16, 7 seconds: close crop; pull back once to show [PRODUCT] scale; CTA hold.” 68 Color sweep: “9:16, 8 seconds: [PRODUCT] against brand-color set; one lateral camera slide; feature detail; CTA.” 69 Bundle build: “9:16, 8 seconds: add each approved bundle item with one simple placement motion; hold full bundle.” 70 Creator hold: “9:16, 8 seconds: creator lifts [PRODUCT] into frame; one turn to label; final direct-to-camera CTA space.”
Which prompt codes help you localize formats and recover weak outputs?
Codes 71–100 finish the system with placement variants and controlled repairs. Cliprise warns that fine label text breaks easily in motion, so write a regeneration brief that names the defect. Never ask the model to generally improve an almost-correct asset.
71 Feed crop: “Adapt approved hero to 4:5; keep full [PRODUCT] visible and copy space top.” 72 Square crop: “Adapt approved hero to 1:1; center [PRODUCT], retain contact shadow.” 73 Story crop: “Adapt approved hero to 9:16; product in central safe zone, reserve top and bottom UI space.” 74 Thumbnail: “Create a high-contrast 16:9 thumbnail with [PRODUCT] large and left text space.” 75 PDP support: “Create a 1:1 secondary PDP image showing [FEATURE], no lifestyle props.” 76 Marketplace: “Create a marketplace-safe white-background image, full product visible, no extra text.” 77 Email banner: “Create a 3:2 email hero with [PRODUCT] right and clean copy space left.” 78 Carousel card: “Create carousel card one: [PRODUCT] hero with minimal background, 4:5.” 79 Carousel detail: “Create carousel card two: [FEATURE] macro, 4:5.” 80 Carousel use: “Create carousel card three: [PRODUCT] in [USE CASE], 4:5.” 81 Logo repair: “Regenerate only the logo area from the source reference; retain scene, crop, light, and all other pixels.” 82 Label repair: “Regenerate only the label from the source reference; preserve visible text exactly.” 83 Color repair: “Correct [PRODUCT] to approved [COLOR]; do not alter lighting, material, or background.” 84 Material repair: “Restore exact [MATERIAL] texture from source; keep shape and scene unchanged.” 85 Proportion repair: “Correct [PRODUCT] proportions to source reference; keep camera angle unchanged.” 86 Hand repair: “Correct hand anatomy while retaining product grip, pack scale, and label visibility.” 87 Reflection repair: “Reduce distorted reflection; preserve product silhouette and original studio light.” 88 Edge repair: “Clean the [CAP OR HANDLE] edge from source reference; retain background.” 89 Copy-space repair: “Move no product details; create clean [LEFT OR RIGHT] negative space for approved copy.” 90 Crop repair: “Reframe to [RATIO] with the entire [PRODUCT] visible and safe margins.” 91 Claim-safe: “Remove all generated claim text; preserve only product and blank copy area.” 92 Prop-safe: “Remove unapproved props; retain [PRODUCT], lighting, and set geometry.” 93 SKU-safe: “Remove extra or incorrect SKU; retain only approved [PRODUCT].” 94 Background-safe: “Replace cluttered background with [BRAND COLOR] studio sweep; preserve product pixels.” 95 Motion-safe: “Reduce motion to one slow camera move; product and label remain stable.” 96 End-card-safe: “Create a static final 2-second end card with [PRODUCT] and CTA safe space.” 97 Lighting-safe: “Match approved [SOFT OR HARD] lighting; preserve pack color.” 98 Shadow-safe: “Add a natural contact shadow beneath [PRODUCT]; no floating effect.” 99 Scale-safe: “Correct product scale relative to [OBJECT]; preserve the source proportions.” 100 Final master: “Create the approved master ad: exact [PRODUCT], [CONCEPT], [RATIO], [ONE BENEFIT], [COPY SPACE], and all QA exclusions.”
| Tier | Price | Included | Best for |
|---|---|---|---|
| Starter cohort | Model-dependent | 10 still concepts | One angle and one placement |
| Funnel set | Model-dependent | 30 still concepts across awareness, consideration, and conversion | Testing three message roles |
| Launch library | Model-dependent | 100 prompt codes plus selected 9:16 motion versions | A structured multi-format campaign |
First controlled still test
Model-dependent10 prompt codes × your model’s per-render price × chosen render count per code
Three-stage funnel creative set
Model-dependent30 still concepts × your model’s per-render price, plus selected video renders
Full 100-code library
Model-dependent100 concepts × your model’s per-render price; exclude human review, revisions, and media spend
How should ecommerce teams approve AI product ads before launch?
Approve an AI product ad only after a human has checked product truth and campaign truth. Product truth covers shape, color, material, logo, label, SKU, included items, and dimensions against the approved reference. Campaign truth covers the claim, offer, audience, crop, copy space, and CTA against the brief.
For short video, inspect the first, middle, and final frames. Do not trust the thumbnail. Cliprise identifies source quality, model choice, prompt clarity, and motion complexity as product-preservation factors; a stable opening frame does not ensure a stable label during an orbit or handoff. Keep motion simple, regenerate the named defect with codes 81–100, then log the version and cohort variable.
Use a launch sheet with five fields: prompt code, unchanged control elements, changed test variable, output format, and approval status. It keeps the record of every decision, including variants you choose not to run.
What is the practical way to use all 100 prompt codes?
Run 10 codes in the first cohort, not all 100 at once. Start with codes 01, 03, 11, 21, 27, 36, 41, 54, 57, and 61: hero, brand world, use case, pain point, feature, bundle, creator framing, differentiation, conversion, and motion. Keep the approved product reference, benefit, and audience fixed, then select the next cohort from the winning visual job.
A prompt library earns its keep when it narrows the next decision. Initial still renders show a median 225 seconds in Lamina telemetry, while the 90th percentile reaches 472 seconds; batch an intentional cohort and leave time for review rather than treating generation time as publication time. Human art direction and approval remain part of published-asset cost.
The operating discipline is simple: keep the product photo fixed, move one variable, preserve an audit trail, and turn only approved still concepts into short motion. Ecommerce teams get breadth without losing the exact SKU the customer expects to receive.
Continue reading

AI product-video ads benchmark for ecommerce
A reproducible ecommerce benchmark for AI product-video ads: measure first usable draft, SKU fidelity, edit burden, and approved export—not render time alone.

Lamina Team
Product Team @ Lamina

AI video generation for ecommerce ads: a hands-on benchmark for creating on-brand product reels without waitlists
A practical benchmark for turning product images into on-brand ecommerce Reels, with a cost-tested image workflow, QA gates, and tool-selection guidance.

Lamina Team
Product Team @ Lamina

AI product-video benchmark for ecommerce reels
A fixed-input benchmark for Lamina, HeyGen, Tagshop AI, and Pippit, with observed Lamina render timing and a publish-ready scoring protocol.

Lamina Team
Product Team @ Lamina