Case Study: Zappi achieves a 43.9% NLP accuracy gain and production deployment in 5 weeks with Kortical's AutoML

Customer feedback automation with Natural Language Processing (NLP)

Zappi, a marketing-research company automating surveys, had spent two years trying to classify complex free-text feedback (180 possible labels) and needed an NLP solution that would beat both their in-house model and a manual outsourcing team. They engaged Kortical and tested the Kortical AutoML platform on this tricky problem.

Kortical applied its end-to-end AutoML workflow—feature engineering, model training/selection/validation and deployment—producing a model in a few hours that improved accuracy from 0.41 to 0.59 (a 43.9% gain) and enabling immediate API testing; the full solution was live in production within five weeks. Kortical’s platform gave Zappi rapid, production-ready NLP at scale.

Zappi