Artificial intelligence is beginning to challenge long-standing assumptions about how residential property is valued, with new research demonstrating accuracy levels well beyond traditional valuation models.
A research team led by Dr Yishuang Xu at The University of Manchester has developed an AI-driven property valuation system capable of predicting house prices with accuracy exceeding 96%. This compares with the 70% to 85% accuracy typically achieved by conventional valuation approaches used across the housing market.
Property valuation remains one of the most consequential yet opaque processes for buyers, sellers and lenders. Price assessments can vary significantly, often leaving stakeholders uncertain about how figures are derived. The newly developed system aims to address this by combining predictive performance with transparency.
Unlike many existing automated valuation models, the system produces confidence intervals alongside each estimate, offering insight into how reliable a given valuation is. As Dr Xu explains: “our system doesn’t just give you a number, it tells you how confident to be in that number, and which features are driving the valuation.”
The model has been trained using millions of historical property transactions across England and Wales, supplemented with data on energy performance, local economic conditions and broader housing market trends.
Advanced machine learning techniques and explainable AI methods allow the system to identify which variables have the greatest influence on individual valuations, from sustainability metrics to regional economic change.
This approach opens up a range of potential applications across the property ecosystem. Buyers and sellers could benefit from more realistic pricing bands during negotiations, while lenders and insurers may gain a more robust basis for assessing exposure and risk. For policymakers, improved visibility into pricing drivers could support better-informed housing and planning decisions.
As Dr Xu notes: “this transparency could transform how people buy homes.”
