Keller Williams Unlocks Agent Efficiency with AI

The leading real estate franchise company embraces the DataRobot AI Platform to boost efficiency across the business, from agent training and retention to valuing homes.

Related products:

AI Experimentation, AI Platform, AI Production

Partnered with:

Google Cloud

Challenge

As Keller Williams embraces predictive and generative AI for applications across its business, it wants to ensure it does so safely, accurately, and meaningfully.

Solution

Keller Williams turned to the DataRobot AI Platform to automate the machine learning process while integrating with the Google Cloud Platform. DataRobot enables fast development and testing without sacrificing quality, and governance for predictive and generative AI.

Result

The company moves predictive models into production at least 3X sooner. With DataRobot assisting with generative AI governance, Keller Williams expects to save at least 3 to 4 weeks in bringing new products to market.

Models to Production 3X Sooner

When looking for an AI partner, Keller Williams turned to the DataRobot AI Platform to automate the machine learning process and for its integration with the Google Cloud Platform.

DataRobot and Google Cloud Platform fit seamlessly into the Keller Williams AI ecosystem. We have predictive models in production and are looking to build generative models that add value and benefit our business.

Frederique De Letter
Senior Director Business Insights & Analytics, Keller Williams

For the data analytics team, DataRobot brings a low-code interface that means fast development and testing without sacrificing quality. Behind that, De Letter and the team value DataRobot’s feature engineering capabilities, easy MLOps, and rapid deployment into production.

Every month, the company derives more than four million inferences, or learnings, and finds answers to support decisions much sooner.

“Before DataRobot, it used to take a data scientist at least 3 to 6 months to prepare the data, iteratively engineer features, and train a model that we could take to market,” De Letter said. “DataRobot allows us to cut that time down to less than a month for a first iteration. We can review output much quicker with the business and they, in turn, feel much closer to the process.”