Case Study: Technomic achieves 97.8% ingredient categorization accuracy with Lexalytics AI

Lexalytics Creates AI for Technomic to Automatically Categorize Thousands of Ingredients

Technomic, a food and beverage market intelligence company, needed a better way to categorize hundreds of thousands of ingredient records coming from multiple vendors with inconsistent labeling. Its internal rules-based approach was too slow and inaccurate, so it turned to Lexalytics for an AI- and NLP-based solution to automate ingredient categorization.

Lexalytics built a proof of concept in two weeks and trained a machine learning model to assign ingredients to the correct subcategories, including a confidence score for each result. The system achieved 97.8% accuracy, saved up to 40 people-hours of labor per category, improved consistency, and led Technomic to expand the Lexalytics solution across its entire 444,000+ ingredient database.

Technomic

Adam Roberts

Director, Product Management