British home-moving platform Homemove has launched AI beta valuation tool “Homer,” which has revealed that the average online property valuations are off by 13% — an average of £70,000 in London and £38,000 in the rest of the UK.
Homer harnesses visual AI, using image recognition and optical analysis to create accurate online property valuations, unlike websites that can share property estimates which are off. It uses over 20 data points and then visually analyses key property features from user uploaded images. It scans for structural damage, clutter, untidy rooms and outdated interiors. Beyond simple data-driven estimates, this visual inspection captures the nuances of a home’s value, giving sellers a more comprehensive understanding of its real worth. Added to a data-driven price estimate, this visual inspection offers sellers the opportunity to find out what their home is really worth, the company said.
“Homer serves as a bridge, linking customers with human estate agents. Visual inspections remain vital in real estate, revealing clues that shape property value,” James Freestone, COO and co-founder of Homemove said. “With Homer, this takes these insights to a new level, analysing visual features in photographs to make even more accurate predictions.”
Considerations that the Homer AI tool works to when looking at property images includes: poor maintenance including overgrown landscaping, chipped or peeling paint, broken windows, rusted gutters, and visible mold/mildew; outdated interiors including dated wallpaper, old cabinetry, ancient appliances, worn-out carpets and outdated bathroom fixtures and clutter/messiness, which can make a home feel smaller.
The research went on to show that 18% of online valuations were overestimated, 7% were underestimated and a quarter of prices were off by more than 10%.