Will AI Replace Real Estate Appraisers?

With the triumphant march of artificial intelligence, we have reached a situation where AI offers only an estimated market value or a range thereof based on the Land Board's transaction data. However, the answer that comes in seconds raises the question of whether the machine can really evaluate real estate and whether this information is adequate.

What does a real estate appraiser do?

A real estate appraiser is a specialist in their field whose job is to determine the market value of a property and justify it in an appraisal report. Market value means the price that real estate could be expected to fetch under typical market conditions. Appraisals are needed, for example, as collateral for bank loans, for determining real estate selling prices, in court disputes, in financial reporting, or for making investment decisions.

However, the appraiser's work begins long before writing the final valuation figure. They gather and analyze information about the property and review documents. When appraising an apartment, it may be necessary to review documents in the land register, the building register, or the Land Board's geoportal. For newer objects, project materials must also be reviewed, as well as information related to parking and storage spaces and how these are included in the property's composition.

An important part of the work is also an on-site inspection. The appraiser examines the actual condition of the real estate and photographs it. For an apartment, all rooms are inspected, and for a house, for example, the basement, attic, and outbuildings are also checked. The appraiser also verifies whether the actual condition matches what is stated in the documents and registers.

This is important because the area, layout, or building description in the register may not always correspond to reality. Apartments may have undergone modifications or extensions, the documentation of which needs to be checked, and such discrepancies can entail legal, technical, or financing-related risks.

The appraiser then selects an appropriate valuation method. For residential real estate, the comparison approach is often used. This means comparing the property being valued with similar properties that have already been sold, and the factors affecting the value of the property being valued are adjusted according to the standard of the sold properties.

When appraising real estate, consideration is given to, for example, the location of the property (such as distance from an attraction center), area, condition of the building, condition of the interior finish, parking and storage space, as well as the presence of a balcony or terrace, heating system, and for houses also the size of the land and outbuildings, among other factors. In this way, the market value of the real estate is ultimately determined.

How does AI appraise real estate?

When placing an AI-based program that appraises real estate next to an appraiser with professional credentials and so-called "real" real estate appraisal experience, it inevitably raises the question of how the latter is possible at all. The answer is data processing. For example, based on the Land Board's official transaction statistics, it is possible to analyze at what prices apartments and houses have actually been sold in different regions.

The program compares the property being valued with actual transactions in the same region and age and size class. The analysis also considers the condition of the real estate, floor, layout, and for houses the size of the land. One of AI's greatest advantages here is speed, as artificial intelligence processes a large amount of existing data very quickly, compares various transactions, and provides an initial valuation based on them.

One automatic valuation solution uses more than 700 price records. These include median prices for apartments and houses in all Estonian counties and municipalities and Tallinn city districts. The data is also broken down by construction year and apartment size and is updated once per quarter.

As a result of automatic valuation, a single fixed number is not provided, but rather a range of market values. However, this reflects a major drawback of AI valuation – namely, that not all properties that affect the value of real estate are taken into account, because they are not reflected in the database. These include, for example, the view of the property, interior finish, and planning peculiarities, which are very important when drawing up a valuation report.

What does AI see and what is overlooked?

So we come to the next point – what can be detected when appraising real estate with the help of artificial intelligence, and what cannot. As mentioned above, AI's greatest strength in real estate appraisal is processing large amounts of data. This makes it possible to compare many transactions, find patterns, and provide data-based preliminary assessment. But real estate itself is not just a data entry.

As a result of automatic valuation, the view from the windows, smart solutions, interior finish, integrated home appliances, and so on are not visible. All these properties make the living space more valuable or sometimes less attractive.

Another problem is data quality. If the data is incomplete, outdated, or contradictory, the valuation based on it becomes unreliable.

A similar problem arises when the information in the register does not correspond to the actual situation. Only with the help of an appraiser is it possible to clarify whether the property has, for example, been rebuilt or if the layout in the documents differs from the actual one. An automatic system can only analyze the information available to it, but it cannot automatically conclude that it can detect differences in the actual situation.

Does AI help with time-consuming work?

Work related to real estate appraisal includes many activities that can be automated when data is available, and here AI has clear potential. For example, gathering and standardizing transaction data, finding similar objects, comparing price levels, and analyzing market data are activities that, when automated, can significantly reduce the workload of a real estate appraiser.

AI processes data many times faster than a human and can identify patterns or outliers in price statistics that an appraiser can then review. So the question is not whether AI can do the appraiser's job, but rather what part of the appraiser's work can and should be automated.

Where does artificial intelligence fall short of real humans?

AI-based appraisal is suitable for situations where there are enough transactions to compare and the objects being compared are similar to each other. For a typical apartment, there may be many similar transactions in the selection, based on which to develop a valuation and market value range.

The situation becomes more complicated with unique, historical, or special-purpose objects. When there are few comparable transactions or the property's characteristics are very different from the usual, it is difficult to compile an adequate statistical comparison.

Artificial intelligence also cannot conduct on-site observation and document verification compared to the actual situation. An appraiser, however, can compare the actual situation with register and project data and identify potential discrepancies. This plays a particularly important role when the value of the property is affected by some circumstance or factor that becomes clear during an on-site inspection.

A professional appraiser's expert opinion is not the same as an AI-generated valuation

An important difference in this entire issue is what each type of assessment can be used for. AI-based real estate valuation is indicative. It is usually free and no one is responsible for the result, but it gives a person a quick overview of what range their real estate market value might be in.

Meanwhile, the work of a professional appraiser requires appropriate qualifications. In Estonia, the professional levels related to real estate appraisal are 5, 6, and 7. A professional appraiser must know the standards regulating the appraisal field, be able to analyze property and the market, justify their valuation, and is also responsible for their work.

Will AI and the appraiser work together?

AI's strength is processing, comparing, and finding patterns in large amounts of data. An appraiser's strength is professional analysis, on-site observation, document verification, and interpretation of complex circumstances.

AI's strengths can be used intelligently to simplify real estate appraiser's work and workload. For example, AI can provide an appraiser with a preliminary valuation and direct attention to objects or circumstances that require additional verification. The appraiser can then focus on the highlighted points and other factors where automatic valuation is insufficient.

Such collaboration does not eliminate the human role or take away the appraiser's work, but rather changes it. AI deals with data processing and repetitive tasks, while the appraiser focuses more on expert knowledge, on-site inspection of the property, assessment of its particularities, and making final conclusions.

A similar trend is also evident more broadly in the real estate sector. AI automates simpler and repetitive activities also in the work of, for example, real estate brokers, but it is important to understand that human experience, knowledge, and decision-making ability cannot be replaced by artificial intelligence.

Therefore, AI cannot in any way replace the expert opinion of professional appraisers, especially for more complex and special properties. Artificial intelligence processes and analyzes data, while the appraiser applies their expert knowledge and makes final conclusions based on it.