Back to Portfolio
Case Study

AI ProductEngine

An automated product creation pipeline for a 5,000+ item Magento catalog.

TopSupplies, AI Product Creation Engine

The Challenge

TopSupplies works with thousands of promotional products across several supplier catalogs. Every item needs Danish naming, clean product text, categories, Magento attributes, color and material filters, image checks, and review control. Doing that manually slows down catalog expansion and makes consistency hard to maintain.

Our Solution

We built a product creation engine on Supabase Edge Functions. Supplier SKUs are imported, normalized, matched against existing Magento templates, checked with AI vision, enriched with attributes and Danish copy, validated before upload, and then created in Magento with images, product data, and audit logs ready for review.

Key Features

Supplier feeds in one flow

Multiple supplier catalogs are pulled into one repeatable creation process instead of being handled SKU by SKU.

AI product agent

The agent reads product data and images, writes Danish product content, suggests attributes, and prepares each item for Magento.

Validation before upload

Image quality, duplicates, category matches, and required Magento fields are checked before a product reaches the store.

Review-ready product pages

Products arrive with images, variants, filters, and audit logs so the team can review instead of rebuild.

Technologies Used

Supabase
Edge Functions
AI Agents
pgvector
Magento 2

How It Works

01

Supplier databecomes structured

Product feeds from suppliers are imported into Supabase and normalized into one format. Images, SKUs, descriptions, variants, and supplier metadata become searchable and ready for automation instead of scattered across files and portals.

Supplier data becomes structured
02

AI preparesthe product page

The product agent compares the item with existing Magento products, chooses a useful template, checks images, writes Danish product text, and resolves categories, filters, and attribute values before anything is pushed live.

AI prepares the product page
03

Magento receivesa reviewed product

The system creates the product with media gallery, product attributes, category placement, and validation logs. The TopSupplies team gets a controlled review flow instead of manual copy, paste, and cleanup work.

Magento receives a reviewed product

Product engine in production

From supplier datato review-ready products

The system turns product creation into one controlled flow: import, enrichment, validation, Magento creation, and review.

TopSupplies product engine workflow
5.024

Products indexed

505

Magento products created

23.046

Duplicate candidates checked

1.534

Attribute options mapped

Numbers are based on the active product system: product index, Magento creation, duplicate checks, and attribute mapping.

Perfect

Let's talk

15 min intro call
Available for new projects

Let's work together

Book a free demo and see how we can help your business with tailor-made AI solutions.