Lemaitre Sécurité: AI Chatbot for B2B E-commerce Catalog

Context

Lemaitre Sécurité is a French manufacturer of professional safety footwear, present on the market for over 70 years. Their B2B catalog includes more than 2000 references, each meeting specific standards (S1P, S3, ESD, etc.) and adapted to particular work environments.

The challenge: professional buyers (purchasing managers, PPE fleet managers) spent an average of 12 minutes searching for the right reference on the site, often without finding the optimal product for their need.

The solution

We deployed a RAG chatbot directly integrated into their PrestaShop store. The assistant allows visitors to describe their need in natural language:

"I'm looking for safety shoes for a warehouse worker in a cold room, with ESD protection"

The system:

1. Analyzes the query and extracts constraints (cold, ESD, handling)
2. Queries the vector database containing the entire catalog
3. Returns the 3-5 most suitable products with their technical sheets
4. Suggests alternatives if the need is not 100% covered

Technical architecture

• Indexing: Automatic PrestaShop catalog sync every 4h
• Embeddings: Multilingual FR/EN model for product sheets
• Retrieval: Hybrid search (semantic + filters on standards)
• Generation: Contextualized responses with product sheet citations

Results

After 3 months of deployment:

The chatbot handles an average of 450 conversations per week, with a user satisfaction rate of 4.2/5.

Testimonial

"Our B2B customers have precise technical constraints. Before, they called us for guidance. Now, the chatbot responds 24/7 with the same level of expertise as our sales team."

— E-commerce Manager, Lemaitre Sécurité

Tech stack

• Python, FastAPI
• PrestaShop (custom module)
• Qdrant (vector database)
• OpenAI GPT-4o-mini

---

Project completed in 4 weeks. Contact us for a similar project.