AI-based eCommerce products recommendations and metadata enrichment

Client

U.S. eCommerce SaaS platform serving both SMB and enterprise distributors.

Challenge

1. Slow quoting: Sales reps spent up to 30 minutes per customer compiling quotes because product data lived in silos.
2. Weak competitive positioning: Identifying equivalent or superior alternatives to rival SKUs required manual research.
3. Poor searchability: Over 80% of catalogue items lacked complete metadata, crippling search and cross-sell efforts.

Solution

1. AI-based process allows sales managers to upload any invoice format to the Platform and receive an AI-generated quote for alternative products. AI uses a complex recommendation algorithm to propose the right product, based on the agreement with manufacturers, margin, and specifications match.
2. AI-based enrichment process is used to populate the product’s metadata.
3. AI-based search provides users with more accurate results and allows them to find alternative products for any input information (SKU, name, etc.).

Results

• 70% reduction in quote preparation time.
• 95% of the catalogue now has complete, standardised metadata.
• Zero ramp-up gap: New reps quote as fast as tenured staff within one week.

Technologies

• Google Gemini
• Perplexity API
• Azure OpenAI
• Google API
• Pinecone
• Mongo