Case Study: homes.co.nz boosts AVM accuracy with Restb.ai

Leading New Zealand Real Estate Portal Levels Up Their AVM

homes.co.nz, a leading New Zealand real estate portal, wanted to improve the accuracy of its automated valuation model (AVM) and better account for differences in home quality within the same area. The company looked to Restb.ai to help solve a key blind spot in AVMs: the interior condition of homes, especially kitchens and bathrooms.

Restb.ai provided its room condition analysis technology, using image-based computer vision models to assess room quality from property photos. After testing more than 10,000 images, homes.co.nz saw significant gains, including up to a 9.2% reduction in median error rate in certain regions, improving AVM accuracy and helping buyers and sellers see more accurate property values.

homes.co.nz

Tom Lintern

Chief Data Scientist