If you're shooting product photos over weeks or months, you'll notice they don't match. Different cameras capture color differently. Lighting shifts. One session has warm tones, another is blue. A phone camera from Tuesday doesn't match a DSLR from last month. Your catalog looks fragmented.
Generative Fill in Photoshop fixes this without reshooting. You can standardize backgrounds, adjust lighting direction, add consistent shadows, and correct white balance across all your product images—even if they were taken weeks apart on completely different equipment. The key is using the right prompt structure and selection method for each product type.
Here's how to actually do it.
Pay once. Keep forever.
Turn raw product photos into studio-ready catalog images at 15–20 per hour using Photoshop's Generative Fill — with prompts actually written for how Adobe Firefly parses text, not generic AI advice. This guide delivers 12 copy-paste prompt
What's included
Or get free updates & new releases:
Follow for updatesYour first product shoot might be on a phone in natural window light. Your second might be with a basic camera on a gray day. A third uses a ring light. Without a controlled studio setup, consistency is nearly impossible—and reshooting everything defeats the purpose of batch processing. Each camera has its own color science. Sunlight at 2 p.m. on Tuesday isn't the same as 10 a.m. on Thursday. Backgrounds vary. Shadows fall at different angles. When customers browse your catalog, these inconsistencies are visible. It looks unprofessional, even if individual photos are sharp and well-composed. Traditional retouching is expensive for a 50-image catalog. But Generative Fill lets you normalize these differences—background replacement, shadow standardization, lighting direction correction, and white balance alignment—all from a single Photoshop template you reuse across every image.
First, pick one photo from your best shoot as your style reference. This is your visual standard—the white balance, shadow direction, and background you want the rest to match. Second, use Generative Fill with a specific prompt to standardize each remaining photo: background replacement to match your reference, shadow added in the same direction, and color temperature adjusted. The prompt syntax matters here—vague instructions produce inconsistent results, but structured prompts using the four-part Firefly grammar formula yield repeatable outcomes. Third, run a quick 8-point QA check on each batch. You're looking for consistent shadows, matched background tone, and aligned lighting direction. If one slips, the specific repair instructions for that failure type take 2–3 minutes to apply. This workflow takes 8–12 minutes per image for precision, or 5 minutes if you're okay with faster, looser consistency. For a 30-image catalog, that's an afternoon—not weeks of back-and-forth with a retoucher.
Jewelry, apparel, and flat-lays see the biggest gains because they tolerate background standardization well. A watch looks professional on white, cream, or subtle texture—the product reads clearly regardless. Glass and shiny products require more care. Reflections matter. But even here, Generative Fill can even out lighting direction and add consistent catchlights without destroying the product's dimension. Phone-shot photos of hand-held items (cosmetics, small electronics, tools) respond fastest. You're often correcting perspective, fixing a cluttered background, and adding a studio shadow. Three generations of prompts usually lands the right result. Products already shot in a studio with controlled lighting? You probably don't need this. Consistency workflows save time when your source photos are already decent but visually mismatched.
Generic instructions like 'make it look professional' fail. Generative Fill needs specific grammar. The four-part Firefly formula: (1) subject identification, (2) action or change, (3) context or style reference, (4) technical constraint. Example: 'white ceramic mug, shadow cast to the left as if lit from upper right, on a seamless white background, no blur or distortion.' That's different from 'nice mug on white background.' The second fails often. The first works repeatedly—across different product categories—because it's explicit about shadow direction, lighting source, and what you're protecting (sharpness). Word count matters too. 15–25 words per fill is the sweet spot. Under 10, results are vague. Over 30, Firefly splits focus and produces inconsistent outputs. The exact cheat sheet—with templates for jewelry, glass, apparel, and flats—is the difference between a 2-minute fix and a 20-minute iteration loop.
The most common failure: inconsistent shadow direction across your batch. You standardized background, but the shadow points right on image 3 and left on image 17. The QA checklist catches this in seconds. The fix: re-run Generative Fill with a more explicit shadow direction phrase—'shadow cast to the left as if light source is upper right at 45 degrees.' Second failure: white balance drift. One image is warm (cream-toned white), another is cool (blue-white). Customers notice. The repair is a secondary Generative Fill step that adjusts the base layer's temperature before you fill. The checklist tells you exactly when this is needed. Third: loss of product detail. You're filling the background, but the edge of the product softens. This happens with poor selection. The documented case studies show how different product types need different selection methods—jewelry needs a tighter mask, apparel can tolerate looser edges. The 8-point checklist turns these from 'I'm not sure what went wrong' into 'shadow direction is off—rerun with this prompt.' It cuts troubleshooting time dramatically.
Once you've standardized your first 3–4 images, you have a reusable workflow. Create a Photoshop action or template that includes your selection method and baseline prompt. Batch processing means you're running the same fill operation across 10 images with minimal variation—just tweaking for product size or specific material (glass vs. fabric). The Batch Consistency workflow in the guide walks through setting up your template, organizing files, and running fills without manual intervention between each image. You're aiming for 12+ images in a 2–3 hour session. QA still matters. But if your template is solid, 80% of your images pass on the first run. The remaining 20% need one re-run or a targeted fix. That's dramatically faster than processing each image individually and hoping they match.
Generative Fill consistency processing saves time if: you have 8+ images to standardize, they're shot on different cameras or days, backgrounds need replacement or unification, and you can tolerate a 5–10 minute per-image investment. It doesn't help if: every image is already studio-perfect, you need extreme precision (e.g., jewelry with complex reflections), or you're only processing 1–2 images (manual retouching is faster). For most e-commerce sellers with rotating product launches—12 items every 2 weeks, shot whenever light is decent—this workflow is a game-changer. It replaces 'hire a retoucher' with 'run templates and QA-check.' Cost and turnaround both drop.