Case Study: PropertyGuru achieves faster model development and deployment with H2O.ai Driverless AI

AI Helps Property Management Company Maximize Their Business

PropertyGuru, a leading property management company in Singapore, faced the challenge of scaling machine learning across high-volume listing workflows—image moderation, churn prediction, credit forecasting and listing performance—so their mobile-first app could stay responsive and accurate. To avoid manual, non-scalable processes they adopted H2O.ai’s Driverless AI to accelerate model building and productionization.

Using H2O.ai Driverless AI, PropertyGuru ingested images and tabular data, used AutoViz for outlier detection, leveraged automated feature engineering, pre-built recipes and hyperparameter tuning, and deployed models directly to AWS Lambda. The outcome: much faster model development and deployment (the team estimates Driverless AI would have saved about three months from development to production), quicker iteration and experimentation, easier access for non-technical teams, faster churn predictions for marketing, and new capabilities like Google Lens experiments—delivering clear improvements in agility and time-to-model.

PropertyGuru

Gautam Borgohain

Data Scientist