Un Amour de Tapis × Rastro AI
Using AI to automate catalog operations for France's leading multi-marketplace rug retailer
Rastro didn't just save us time. Rastro gave us a new scale.
The Challenge
Un Amour de Tapis manages one of France's most extensive rug catalogs, distributed across Amazon, La Redoute, TikTok Shop, and Mirakl marketplaces—and in several languages.
The internal team had already mastered multilingual commerce—French, English, German, Spanish—but all product data lived in sprawling, unstructured spreadsheets. Each marketplace imposed its own constraints:
Amazon
200 character limit + A9 SEO compliance
La Redoute
Short, descriptive, inspiring titles
Maisons du Monde
70 character strict limit with model, color, and dimensions
The result: hundreds of hours spent copying, pasting, reformatting, and manually correcting data. Even a simple update—changing a color name or dimension—required hours of work and risked breaking something elsewhere.
The problem wasn't translation, it was chaos. Every modification broke another cell—our data was never really clean.
The Rastro Solution
Rastro AI designed a custom automation flow for Un Amour de Tapis that replaced fragile spreadsheets with intelligent, centralized catalog operations.
1. AI-Powered Attribute Extraction
Rastro's vision agents detect colors, materials, and patterns directly from product photos, then cross-reference results with supplier specifications for validation.
2. Centralized Data Logic
All title generation, descriptions, and format rules consolidated in a single flow—no more multiple Excel sheets or version conflicts.
3. Channel-Adapted Content
For each product reference, Rastro automatically generates optimized titles and descriptions per marketplace—Amazon, La Redoute, TikTok Shop—respecting character limits and expected tone.
4. Automated Validation
Every attribute—color, size, material—is verified, normalized, and enriched before export. Source traceability built in for audit compliance.
Today, everything is centralized: every title, every description, every language follows the same logic. It's fast, consistent, and much more reliable.
Results
In just a few weeks, Un Amour de Tapis went from slow manual processing to large-scale production across multiple countries and channels.
| Metric | Before Rastro | After Rastro |
|---|---|---|
| References added per week | ~300 | 1,500+ |
| Time for 10,000 references | 3+ months | < 1 week |
| Manual hours per week | 30-40 hrs | < 5 hrs |
| Error rate | High | < 1% |
| Channels managed | 5 (manually) | 5 (automated) |
Before Rastro, publishing 10,000 references was a three-month project. Now, it's a matter of days—and we're confident in the data quality.
Why This Matters
Un Amour de Tapis didn't need more translators—they needed structure.
By replacing fragile spreadsheets with intelligent automation, the company transformed catalog management into a controlled and scalable process.
Today, they can launch thousands of new references, in multiple languages and across multiple marketplaces, in just days—with consistent quality and without work overload.
Ready to scale your catalog operations across multiple channels?
Let's walk through your marketplace requirements and show you how Rastro can automate your catalog workflows.