Skip to content

Does AI spell the end for the Valuation Surveyor?

Phil Winckles

In recent months, there has been a lot of talk about how AI might affect the valuations process, and whether this could spell the end for the valuations professional. Phil Winckles, partner in Fisher German’s national valuation team, explains the state of play in the industry and what the future might hold.

In the ‘good old days’ before high-speed internet, valuing property and portfolios in particular could be a laborious task.

Having to drive out to view all comparables, cross reference recent sales with other professionals in the area, and having to sift through reams of transactional data to get a feel for the market meant the process was anything but simple.

Over the last 15 to 20 years, there have been dramatic technological advances which have made this easier.

The introduction of Google Street View gives the ability to gain a basic understanding of a property without visiting it, and the development of online databases has provided an immediate source of transactional data.

In addition, online services including Land Registry data and mapping provide notable time savings.

But recently, there has been a lot of talk about how AI might affect the valuations process, and whether it could spell the end for valuation surveyors.

Firstly, it is worth highlighting that true AI -computer-based systems that have the ability to ‘learn’ – is still in its infancy.  The property profession as a whole is one which is typically slow to adopt new technologies and development.

Automated valuation models (AVMs) have been in place for a number of years now, although these are not currently AI-based systems.  Instead, they draw upon data within the market and apply this based upon general property trends.

Many of us will be familiar with some of these systems, and may even have tried more basic systems such as Zoopla’s house price calculator.  Whilst these can assist non-professionals in general trends, I have not met clients who would not rely solely on this tool to make key valuation decisions.  The intervention of a professional, with profound market knowledge is critical.

Ultimately, property valuation is a heavily data-driven process, and the role of the valuer is to scrutinise the data, verify it, and produce a valuation for a client based upon this evidence and their wider market intelligence.

Some of this intelligence is gleaned from conversations with other professionals in the market based upon immediate market interactions.  However, that data is often stored in different ways and can sometimes be wildly inaccurate.

In truth, the property industry is not there yet when it comes to fully adopting AI for valuations.

What AI is currently very good at is using existing data and forming fast conclusions from that data. For general AI programs like ChatGPT, a lot of its output is based on scraping information from the Internet.

In valuations, this is only one part of the job. Indeed, AVMs do this job very well already, and a generalist AI program is unlikely to ‘beat’ specialist software in this regard.  For simplistic valuations, such as residential housing within a stable market and with plentiful transactions, there could well be an early adoption for AI in valuation.

However, predicting the future, especially in something like the state of a property market, is much harder.  The impact of market shifts, economic fluctuations, occupier usage habits, and incoming legislation is notoriously difficult to apply to values.  For example, the current green premium/two-tier split between offices with high ESG credentials and older buildings, has already thrown off previous valuations at a greater pace than was anticipated.

Major global events and how the market is affected by them, such as the economic crash of 2007 and the pandemic, are things that AI modelling would find very difficult to predict.

The human element is vital in these cases to re-evaluate the market and come up with accurate valuations. Once the market settles, automated software and AI may then become more useful.

However, the property market can hit major peaks and troughs after a period of relative calm, which would also frustrate an AI’s ability to predict values effectively.

Ultimately, like a lot of industries, we view AI similarly to previous technological developments. They are tools for us to use, rather than replacing us.

For our part at Fisher German, we are always looking to embrace new technology, and explain to our clients about what we are using and how it affects our valuations.  We have been an early adopter of VALOS to improve valuation reporting, data collation and compliance and have also invested heavily in our data warehousing to improve the efficiency and quality of data which we will hold.

As a firm, we remain of the view that transparency is key, and addressing ethical challenges within a machine world remains a key focus.

At the moment, true AI that ‘learns’ is not there in valuations terms. In a few years’ time, it may well be that AI can speed up the data verification process, but we’re not at that stage yet.

The important thing for us to monitor advances in technology closely, learn to work with it and use it to produce and even better service for our clients. AI may have a lot of applications for valuations in the near future, but it will be a long time before it can surpass trained professionals.

Topics

Related