J M .
Jaydeep Mayani
Ahmedabad, IN
Available for
Consultation
Back to the eCommerce Case Study

How to Generate Product Photography Without a Photoshoot

Technical Deep-Dive AI Automation Jaydeep Mayani

Every eCommerce catalog assumes the same starting point: a physical photoshoot. A model or mannequin, a studio, a photographer, lighting - for every single product. In my earlier case study, I described how I automated Shopify uploads and catalog data. This piece goes one level deeper: the pipeline that removes the photoshoot itself.

I'm Jaydeep Mayani, a Project Manager and AI Automation Builder based in Ahmedabad, India. This is a technical walkthrough of the Make.com-orchestrated pipeline I built that takes a single fabric swatch photo and generates studio-quality, publish-ready product photography - no garment stitched, no model booked, no studio needed for the first pass.

Why Skip the Photoshoot?

For high-SKU sellers - sarees, kurtas, blouses, and similar apparel - a full photoshoot per SKU is one of the slowest, most expensive steps in getting a product live. It requires:

  • A finished, stitched sample of the garment before any photography can happen
  • A studio, lighting setup, and either a model or mannequin
  • Post-production editing before the image is publish-ready

If the only thing that's actually ready at the start is a fabric swatch, all of the above becomes a bottleneck standing between raw material and a live listing.

The Pipeline, Step by Step

1. Detecting New Work Automatically

Make.com watches a structured Google Drive folder (organized by batch and date) for new fabric reference photos. Before processing anything, it cross-references a Google Sheet tracker to confirm the item hasn't already been handled - so the pipeline only ever touches genuinely new work, and re-runs are always safe.

2. Generating the Garment from a Fabric Reference

The fabric photo gets sent to Google's Gemini image-generation model with a precisely engineered prompt - not a vague "make this look nice," but exact construction specifications: sleeve style, neckline shape, closure placement, and an explicit instruction to preserve the fabric's real color and pattern exactly as photographed. This precision is what separates a usable product photo from a generic AI image.

3. Generating a Styled Lifestyle Shot

A second prompt takes that result and generates a complementary lifestyle flat-lay - styled props, natural lighting cues, a cohesive scene - giving the listing both a clean product shot and a styled marketing image from the same original reference.

4. Writing the Product Copy Automatically

A third call analyzes the finished image and returns structured JSON - title, color, fabric, occasion, and a ready-to-use product description - using a defined response schema so the output drops straight into the catalog system without manual reformatting.

{ "Title": "...", "Product_Description": "...", "Color": "...", "Fabric": "...", "Occasion": "..." }

See how this connects to the full Shopify upload pipeline

Where Precision Actually Matters

The riskiest part of any AI image generation workflow is fidelity - a generated photo that doesn't accurately represent the real product is worse than no photo at all. The prompt engineering here is deliberately strict: exact fabric color and pattern preservation, explicit exclusions ("strictly NOT sleeveless," "no V-neck"), and a fixed aspect ratio and background so every output is consistent enough to drop straight into a catalog template.

Who This Is For

  • Automation builders and agencies designing similar AI-driven content pipelines
  • Apparel and fashion sellers evaluating whether photography can be partially automated
  • Project managers scoping AI automation projects who want a real, working example rather than a theoretical one

Want a similar pipeline built for your catalog?

Get in Touch

FAQ

Can AI generate product photography without a real photoshoot?

Yes - using an AI image-generation model prompted with a single fabric or material reference photo, it's possible to generate studio-quality product photography and lifestyle shots without a physical shoot.

What AI model powers this image generation pipeline?

This pipeline uses Google's Gemini image-generation models, called through Make.com's HTTP module with a structured prompt describing exact garment construction details.

How does Make.com orchestrate the AI image generation process?

Make.com watches a Google Drive folder for new product references, cross-checks a Google Sheet to find unprocessed items, sends each reference to the AI model with a detailed prompt, and writes the generated results back to a tracked document.

Is AI-generated product photography accurate to the real product?

Accuracy depends on prompt precision - the pipeline explicitly instructs the model to preserve the reference fabric's real color and pattern while only generating the garment construction and styling around it.

Who is Jaydeep Mayani?

Jaydeep Mayani is a Project Manager and AI Automation Builder based in Ahmedabad, India, with 7+ years of experience across project management, digital marketing, and AI-powered workflow automation for eCommerce and content operations.