Case Study: Square enhances customer communication with AI tools from Weights & Biases

Square Brings Conversational AI to Businesses of All Sizes With W&B

Square, a financial services and mobile payment company, needed to enhance its customer communication tools for business owners. To help entrepreneurs manage conversations efficiently, its Machine Learning team required a robust system to handle automated model retraining, track model lineage, and manage the deployment of its large language models. They turned to Weights & Biases for its MLOps platform to address these challenges.

Using Weights & Biases, Square implemented a centralized system for tracking experiments and managing model artifacts. This provided a complete lineage for all models, allowing for an automated A/B testing framework where new models progress from shadow to challenger to champion status. The solution enabled Square to quickly debug models, ensure performance metrics were met, and confidently deploy the highest quality models into production, significantly improving their AI-powered Square Assistant tool.

Square

Ethan Rosenthal

Staff AI Engineer