AI + human valuation: A powerful team of real data and insights

Technology has transformed financial management—and commercial real estate valuation is no exception. Today, with the touch of an app, anyone can see their entire net worth and digitally manage everything from their investment portfolio to their checking and savings accounts. Automation has made our financial lives so convenient that more people currently use financial technologies than social media. Engagement increased 52% from 2020 to 2021 alone.

Likewise, automated valuation models (AVMs) are providing this same convenient financial insight and accessibility to commercial real estate owners and investors. Through an AVM, owners can access individual property or portfolio valuation data as easily as any individual can check a bank account balance.

“In the past it has been next to impossible to do this with commercial real estate because it is not as liquid as other investment assets,” says Tyrone Hodge, Global Head of Risk Advisory at JLL. “However, we have successfully created a suite of risk analytics tools by combining statistical models with automation. We now have the ability to provide timely insights on changes in property values and market insights to owners and lenders.”

Automated valuation is already ubiquitous in the residential market. Homebuying marketplaces like REA Group (Australia), Rightmove (UK) or Zillow (US) give an estimated property value or predict the sales price of a listing through an AVM by using comparable sales and listing data, just as a real estate appraiser does in a property evaluation. In the commercial sector, AVMs are more complex versions of the same system. They process a long list of data, like market changes, risk assessments, net operating income, occupancy and debt coverage. 

“We are using data and analytics, artificial intelligence and alternative statistical methods to estimate value and market risk, and we have a high level of accuracy,” says Hodge. Some investors have been apprehensive about AVMs because the process diverges from traditional valuation methodology, but that process is highly subjective. Asset appraisers consider a narrow field of data, relying largely on comparable sales, occupancy and net operating income.

AVMs pursue a scientific process, removing as much subjectivity from the process as possible. Computers and AI are capable of analysing large amounts of data that collectively can have a measurable impact on value. They weigh a wide field of data insights and property nuances and are updated automatically as data changes. This continuous cycle of updates and recalculations has become an important tool for investors, lenders and appraisers to gain a complete and accurate picture of value.

Access real-time insights on asset values and mitigate climate and market risk in your portfolio to optimise future investment and lending opportunity in real estate.

In 2022, a large study was published in The Lancet Digital Health analysing the use of AI to detect breast cancer. The study looked at doctor diagnosis, AI diagnosis and a combination of the two. Alone, doctors were better at detecting breast cancer in patients than the technology solution, but when used together—both doctor and AI—breast cancer detection improved 2.6% with fewer false positives. This research is an insightful depiction of success when human and artificial intelligence are combined. AVMs benefit from the same collective knowledge. Automated analysis is important, but it doesn’t replace the human element. When the two work together, investors will gain a richer and more precise estimation of value.

“AVMs are meant to complement traditional valuations, not eclipse them. It is really meant to expand our reach,” says Charles Fisher, Risk Analytics Director, Value and Risk Advisory at JLL. “The idea is to become faster, more accurate and more efficient but not remove the human component.” By using automation as a tool for human analysis, AVMs marry the traditional qualitative approach with a more scientific quantitative approach. This combined approach carries the potential to improve the accuracy of valuation estimates and give owners regular unrestricted access to the data.

Humans also give important perspective to the data that the AVM cannot. AI can’t tell you how it arrived at a specific value or what it means in the context of the broader market, the owner’s total holdings or the business model. “The computer can’t tell you what to do,” says Hodge. “When an owner sees the information on their dashboard, they immediately want to call an advisor to find out the next logical actionable steps. Do they sell or hold or change the strategy?”

In a traditional valuation model, investors and lenders don’t have regular access to asset pricing estimates. To estimate market value, many piece together information from past reports and internal analysis. An AVM, however, gives investors an accurate and current estimation of value without a lengthy formal appraisal. “Investors have an opportunity to understand the value of their portfolio all of the time, whenever they want,” says Hodge.

Through this routine self-assessment, investors can track nuanced changes in property value and uncover emerging trends across a portfolio. When there is a concern in pricing trends, investors can connect with an expert to provide deeper insights into the data or perform a risk assessment to reveal the outside factors influencing value.

Investors have every reason to expect as much 24/7 data about their professional real estate holdings as they can access when managing their personal finances. AVMs deliver on that expectation, providing continuous accurate and precise valuation data. For property owners and lenders, this means the freedom to access and absorb valuation information easily when, where and how they need it.