BACK TO PROJECTS

BACK TO PROJECTS

AI STRATEGY

E-COMMERCE

Product Classification System

AI system using Llama 3 for embeddings and matching.

CLASS ACCURACY

99.8%

DELIVERY TIMELINE

8 Weeks

REGION

Saudi Arabia

CLIENT PROFILE

Industry

E-commerce Strategy

Location

Saudi Arabia

Year

2024

TECHNOLOGY STACK

  • Llama 3 3.2
  • Python

The Challenge

A significant challenge was managing a large volume of unstructured product data in Arabic, which urgently needed standardization to UNSPSC codes for consistent reporting.

"Unstructured Arabic product data needing UNSPSC standardization."

The Solution

We implemented an advanced AI-driven product classification system leveraging Llama 3 3.2 for sophisticated text processing.

Text Embeddings

Converting translated Arabic descriptions into machine-readable embeddings.

Cosine Similarity

Matching embeddings with UNSPSC code database for automated classification.

Llama 3 3.2 Model

Sophisticated text processing of Arabic product descriptions.

Full Impact Analysis

1

Successfully integrated legacy product data with current systems, breaking down data silos.

2

Enabled continuous reporting across historical and current data.

3

Improved data standardization and accessibility, streamlining operations.

FUTURE ROADMAP

Further refinement of models for new product categories.

Ready to build your success story?

Initiate Consultation

Partner. Build. Scale

QUICK LINKS

Insights

Services

READY TO START ?

Let's discuss how we can engineer it into reality. We are ready to partner with you.

CONTACT

contact@wbos.co

All rights reserved. WBOS.Powered by Opmiz