How to Build an AI Product Photography Shot List Before You Prompt

Generating one attractive product image is easy to mistake for finishing a product shoot. A storefront, marketplace listing, launch email, and social post need different evidence. The buyer must understand what the object is, how it looks up close, how large it is, and where it fits into daily life. One dramatic hero frame rarely answers all of those questions.

A shot list solves this before prompt writing begins. It assigns each image a job, defines what must stay consistent, and prevents ten variations of the same three-quarter angle. This guide focuses on planning synthetic or AI-assisted product imagery. It does not replace truthful photography when exact dimensions, color, fit, safety features, or included accessories are material to a purchase.

Start with the claims the images must support

Write down the facts a customer needs to verify visually. For a desk lamp, that might include the shape of the shade, the footprint of the base, the controls, the cable, the direction of light, and its scale beside common desk objects. For a bag, the list may include closures, interior compartments, strap options, hardware, and how it sits on a body.

Separate three kinds of information:

1. Identity: silhouette, colorway, label, packaging, and signature details.
2. Function: how the product opens, connects, pours, folds, fits, or operates.
3. Context: the environment, user, season, or problem it belongs with.

AI imagery is safest when identity information comes from an approved product reference and the generated parts add context. If the system invents a vent, button, ingredient, port, or included item, the picture has become a false product claim. Keep a reference board visible during review.

The six-shot minimum set

Not every product needs exactly six images, but these roles expose the gaps in a typical gallery.

1. The clean hero

The hero gives an immediate, unobstructed read of the product. Use a simple background, a clear three-quarter or front view, and enough contrast to separate the silhouette. Decide whether the hero must work on white for a marketplace or can use a branded environment on your own site.

Prompt the product first, then composition and surface:

Approved amber glass soap bottle with black pump, centered three-quarter view on warm off-white seamless paper, soft shadow falling right, label facing camera, generous space around the object, realistic studio photography.

Do not overload this frame with ingredients, towels, hands, and decorative plants. Those may support later shots; the hero's job is recognition.

2. The material or detail shot

Choose one feature that cannot be understood at thumbnail size: stitching, surface grain, connector layout, fabric weave, embossed lettering, or a closure. The frame should make the texture legible without implying a level of precision the source does not contain.

Use macro framing carefully. If the output invents microscopic detail, it may look persuasive while being wrong. When accuracy matters, composite or crop a real detail photograph instead of generating it.

3. The scale shot

Dimensions written in a specification table are important, but people also judge size relationally. Place the product beside a familiar object or in a recognizable setting: a mug under a coffee machine, headphones beside a laptop, or a carry-on inside an overhead-bin context.

The comparison object must have a stable, familiar scale. Avoid objects with wide size variation unless you state the intended version. A “notebook” could be pocket-sized or larger than a tablet.

4. The in-use shot

Show the product doing its main job. A hand may operate a control, a person may wear the item, or the object may affect the environment. Keep the action simple and physically plausible.

Review anatomy, grip, contact points, shadows, and reflections. A hand hovering through a handle or a cable terminating nowhere undermines trust even if the overall image is polished.

5. The context or lifestyle shot

This frame answers “Is this for someone like me, in a place like mine?” Choose a specific but not over-scripted situation. Instead of “luxury lifestyle,” define a compact morning kitchen, a shared studio desk, or a rainy commuter entryway.

Context should not quietly change the product. Re-check the number of seams, buttons, label placement, proportions, and color against the approved reference. If those drift, treat the scene as a concept, not product photography.

6. The campaign crop

Plan at least one frame for design use rather than catalogue completeness. It might put the product on the right with quiet space on the left, or keep all essential detail inside a central crop-safe area. Define the final placements: website banner, vertical story, square post, or email header.

Do not ask a single generated canvas to solve every ratio. Generate or extend intentional compositions for the most important formats. The structured prompt method helps isolate the composition clause so the subject can remain stable.

Build a consistency block once

Every shot should repeat a short product identity block. Store it separately from scene instructions:

Identity block: cylindrical 250 ml amber glass bottle; matte black ribbed pump; cream rectangular label with a thin charcoal border; label centered on the front; no cap; proportions and colors must match the supplied reference.

Then add a shot-specific block for camera, action, setting, and light. This separation makes drift easier to diagnose. If the bottle changes, revise the identity or strengthen the reference. If the scene is weak, revise only the shot block.

For a multi-SKU range, create one identity block per product and one shared brand block for palette, backdrop materials, shadow softness, and retouching. Do not rely on a vague phrase such as “same style.” List the repeatable visible rules.

Make a contact sheet, not a folder of favorites

Place candidates in a grid with the shot role written below each frame. A contact sheet reveals repeated angles, inconsistent scale, and color drift faster than opening images one at a time.

Review in three passes:

• Truth pass: product geometry, components, labels, proportions, and claims.
• Set pass: palette, light direction, shadow density, camera height, and crop.
• Delivery pass: resolution, file format, background requirements, and safe area.

Reject an image that fails the truth pass even if it is the most beautiful. Retouching can remove a dust spot; it should not be used to rationalize an invented feature.

Match files to placements

Name exports by product, role, ratio, and version:

amber-soap-hero-4x5-v03.webp

Keep the accepted prompt and reference IDs in the asset record. Record whether the image is fully generated, composited with a real product photograph, or photographic. That provenance helps future editors understand what can be safely revised.

For the final storefront, test the image at the actual rendered size. Fine details that look excellent on a large monitor may disappear in a 160-pixel thumbnail. Conversely, excessive sharpening and texture can look harsh after a marketplace resizes the file.

A reusable planning table

Before generating, fill one row per deliverable:

Role	Customer question	Must show	Must not change	Final ratio
Hero	What is it?	full silhouette, front label	shape, color, label	1:1
Detail	How is it made?	one material or mechanism	real construction	4:5
Scale	How big is it?	familiar comparison	dimensions	4:5
In use	How does it work?	contact and action	controls, anatomy	3:2
Context	Where does it fit?	audience and environment	product identity	4:5
Campaign	Where does copy go?	negative space	crop-safe product	16:9

The table turns prompting into production. Each image has an acceptance test, and missing coverage becomes obvious before launch day.

Disclosure and commercial accuracy

Consumer-protection rules vary, but the durable principle is simple: advertising should not materially mislead. The U.S. Federal Trade Commission explains that advertising claims must be truthful, not deceptive, and supported when appropriate. An AI-generated scene does not relax that obligation. If an image could create a false impression about size, performance, ingredients, included items, or real-world results, change it or disclose its illustrative nature.

Use AI for ideation, backgrounds, lighting exploration, and campaign variants where it adds speed. Use verified source photography or accurate 3D assets for details that buyers rely on. You can create early scene studies with BananaLite, then route the accepted concept through the level of product verification the claim requires.

Sources and further reading

• FTC Advertising and Marketing Basics — primary guidance on truthful, non-deceptive advertising.
• Google Merchant Center image requirements — practical requirements for product data images.
• GS1 product image specification resources — standardized thinking about product-image views and file delivery.

A planned set beats an accidental collection. Define the buyer question behind each frame, keep product truth separate from scene invention, and review the gallery as one system.