AI and AVM’s can assist a valuer but also put them at risk due to future computerisation. The main question is whether these developments could help the valuer deliver a more accurate and efficient valuation, or alternatively, completely or partially, replace the role of the valuer.
For instance, automated valuation models automated valuation models (AVMs) will impact the role of the valuer by making it easier and faster to arrive at conclusions based on data. The philosophy that ‘valuation is an art as much as it is a science’ derives from the human touch in the process and, to an extent the, margin for variation this creates in valuation. Some may argue that AI can’t fully impersonate the human element of valuation. However, AI can learn from previous situations, like the AI chess computer which improved its performance with each additional game until it performed better without human aid. Arguably the more AI is used in the valuation process the better it will become at performing and producing accurate results.
Most AI and AVM’s work on a basic comparable method and as a result, comparable transactions need to be reasonably high, with the differences between properties identified through straightforward attributes. In a survey conducted by Northumbria University, the 90% of property professionals questioned agreed that a ‘valuers’ ability to evaluate comparables is a major advantage over AVMs.
The system also can’t perform a physical inspection of the property and instead must rely on the average condition which may not wholly reflect the reality. When asked whether they agreed that ‘inspection by a valuer reduces the risk of fraud’ 87% of respondents agreed, 10% were neutral; and 3% disagreed.
“Agents will not be replaced by technology …. they will be replaced by agents with technology” – Helm, 2018.
AI also doesn’t have capacity for the ‘valuers local knowledge’ element to help interpret this data. The Northumbria University study found that 87% of respondents felt because of this, valuations are more accurate than AVMs. It also may prove very complicated for a valuer to explain the methodology that an AI-equipped AVM has employed should the client ask or if there is a legal requirement to do so if a valuation figure is legally challenged. This therefore begs the question to what degree AI can assist the valuer before it is at the expense of the client’s satisfaction. As property expert Declan King MRICS states, it’s undoubtable that AI and AVM’s pose great opportunities for the profession in assisting the valuer and ‘allowing them to focus more on their core skill of valuing’. However, because of its inherent limitations, ‘the valuer is central to the process’ and will continue to be in the future.
So it would seem we’re safe from the robots . . . for now.