Case Study: G5 boosts lead scoring accuracy with H2O.ai Driverless AI

G5 boosts lead scoring accuracy to 95% with H2O.ai

G5, a real estate digital marketing platform, faced the challenge that only 14% of its call-based leads were productive, which drove low agent job satisfaction and weak conversion. Their small product team had no dedicated data science resources to build the complex machine learning models needed to identify higher-quality leads. They turned to vendor H2O.ai and used its H2O Driverless AI and H2O Word2Vec products to address this marketing optimization problem.

H2O.ai's solution used Word2Vec to process call transcripts and Driverless AI to automatically engineer features and train a model. This was deployed via low-latency MOJO pipelines on AWS Lambda, providing agents with a simple ranked list of high-value leads. The results were transformative: lead scoring accuracy climbed to over 95%, and the qualified-lead connect rate soared to 85%, up from 14%. This solution also reduced feature engineering and tuning time by 80%, and one customer saved over $500,000 a year.

G5