Aligning workwear innovation with customer-focused insights

Overview

Carhartt used the Databricks Data + AI Platform with Delta Lake and Lakeflow Jobs to analyze over 180,000 customer reviews with a Llama-3.1-70B model for sentiment analysis and emotion detection, achieving a 100% reduction in pipeline failures, 50% less code review time and 6x faster delivery of new customer features.

The challenge

Harnessing customer data at scale proved challenging due to its sheer volume and complexity: data was fragmented across multiple systems and formats, making it difficult to create a unified customer view, and machine learning workflows were slow and inefficient, requiring manual intervention to train and update models.

The solution

Carhartt implemented the Databricks Data + AI Platform with Delta Lake at its core to centralize customer review data for analytics and machine learning, layered Databricks Lakeflow Jobs to streamline orchestration of data preparation, ingestion and processing, and used the Llama-3.1-70B model via the Databricks Foundation Model API to perform sentiment analysis, emotion detection and customer feedback classification, surfaced through self-service dashboards.

Reported business value

Carhartt achieved a 50% reduction in code review time, an 80% decrease in setup and maintenance effort compared to prior orchestration tools, and the complete elimination of pipeline failures, reducing time-to-market for new features from months to just two weeks.