Artificial intelligence (AI) is currently disrupting the Swiss property market, particularly in the area of property valuation. Thanks to automated valuation models (AVMs), it is possible to obtain a quick and reliable valuation in just a few minutes.
Major players such as UBS, as well as specialised fintech companies such as PriceHubble and MoneyPark, use these digital tools to provide instant valuations, which are particularly suited to standard properties.
Traditionally, property valuation relied on human expertise, with site visits and traditional statistical methods. Today, AI brings speed, accuracy and enhanced analytical capabilities, transforming the way professionals and individuals approach this key stage of the property market.
In Switzerland, it is essential to distinguish between property valuation, which is often quick and sufficient to set a price, and property appraisal, which is more in-depth, carried out by a certified and licensed expert (member of USPI, CSEI or SIV) and includes a physical visit, a legal analysis and the professional liability of the latter.
AI primarily facilitates valuation by analysing the market, comparing properties and mapping trends, while real estate expertise remains essential when complex or legal aspects need to be taken into account.
This article explores how AI is revolutionising property valuation in Switzerland, highlighting its advantages, limitations and future prospects.
Traditional property valuation in Switzerland
Historically, property valuation has been based on three main methods:
1. Traditional methods
- The hedonic method (comparable values)
It breaks down a property into its key characteristics (location, surface area, number of rooms, condition, etc.) and uses statistical analysis to calculate the average price of similar properties. Widely used by banks for mortgage loans, it also serves as the basis for online valuation platforms, offering quick and reliable valuations.
However, this method requires a sufficient volume of comparable transactions, which limits it to standard properties. For atypical or complex properties, a more qualitative and in-depth analysis remains necessary. (source: UBS, Hedonic estimation: how it works, 2024). - Income approach
It is based on the net rental income that the property can generate and updates it using a capitalisation rate to determine its value. This is the preferred method for valuing rental or commercial properties, with a focus on profitability. (source: BCBE, Selling a property at the right price: how to determine its market value, 2024). - Cost method (actual or intrinsic value)
It estimates the cost of rebuilding the property, deducts depreciation, and then adds the value of the land. Suitable for atypical or new properties, it is used when there are no comparable properties available. (source: realadvisor.ch, how much is my house worth?).
2. Limitations of traditional methods
Although effective in many cases, these methods quickly reveal their weaknesses when faced with certain situations or types of property.
- Long and costly process
Traditional methods can be time-consuming and costly, especially for complex or unusual properties. Physical visits, detailed analyses and the collection of multiple pieces of information significantly increase the time and budget required. - Results sometimes subjective
They are often based on subjective judgements regarding criteria such as obsolescence or capitalisation rates, leading to discrepancies between different valuers. - Unreliable data in low-transaction areas
Finally, the limited availability of data in certain regions or segments limits the accuracy of estimates.
These constraints paved the way for artificial intelligence, promising faster and more reliable assessments.
AI: a paradigm shift
1. AVMs and Machine Learning
Artificial intelligence and machine learning have given rise to automated valuation models (AVMs), which are capable of analysing millions of data points in real time, such as past transactions and market trends.
Using algorithms, these AVMs exploit historical data: transactions, previous valuations, advertised prices, etc., employing statistical methods such as hedonic models or repeat sales indices to estimate the value of a property. (source: Investopedia, Modèle d’évaluation automatisé (AVM) : définition et fonctionnement, 2025).
Some advanced models incorporate key contextual data, such as crime rates, access to transport and environmental quality, to assess the attractiveness of neighbourhoods.
In Switzerland, sophisticated technologies combining computer vision (image analysis), natural language processing (textual information extraction) and enriched geospatial data are used.
The start-up Avendo illustrates this approach by generating complete sales files in a matter of minutes using AI and local databases. PriceHubble, a pioneering Zurich-based start-up, uses big data and machine learning to provide automated real estate analyses and forecasts with detailed scoring.
2. How artificial intelligence is changing things
- Speed
AI makes it possible to carry out property valuations in a matter of seconds, whereas traditional methods can take several days or even weeks. This acceleration is based on the ability of AI models to process large amounts of data quickly and simultaneously. This speed facilitates decision-making for brokers and banks, particularly in a context where time is a critical factor. - Increased accuracy
In data-rich urban areas, AI provides more accurate valuations than traditional methods by analysing numerous factors: location, size, age, specific characteristics, comparable sales, market trends and macroeconomic context. These more accurate estimates help sellers and buyers set or recognise a balanced price, thereby reducing the risk of over- or undervaluation for successful real estate transactions. - Predictive vision
Some models use machine learning to predict the future value of properties by analysing local trends, urban projects and economic indicators, providing a useful dynamic estimate for investors and professionals. (source: McKinsey, Generative AI can transform the real estate sector, but it must evolve to take full advantage of it, 2021). - High regulatory compliance
Some solutions, such as those offered by PriceHubble, comply with the strictest standards, including the guidelines of the European Banking Authority (EBA). This enables professionals to ensure regulatory compliance during property valuation, particularly in the context of granting and monitoring loans.
Result: less risk of overvaluation or undervaluation, which makes transactions safer for buyers and sellers.
The limits and prospects of artificial intelligence in property valuation
Despite its many advantages, AI in property valuation still has certain limitations that are important to be aware of in order to fully understand its scope of application.
1. Current limitations
- Data quality and availability
The effectiveness of AI models depends on the quality and diversity of the available data. In rural areas, in neighbourhoods with few transactions, or for atypical properties, comparable data is often scarce or insufficient. This limits the ability of AVMs to provide reliable estimates, making human intervention necessary. - Difficulty in taking qualitative specificities into accounts
Certain qualitative characteristics of a property, such as the actual condition of the finishes, the quality of the materials, recent renovations, or even the emotional and subjective aspects associated with a neighbourhood, are difficult to quantify and integrate into AI models. Physical visits and expert judgement remain essential for these aspects. - Risks associated with reliance on automated models
Relying solely on automated estimates can lead to overconfidence, especially when users are unaware of the limitations or opaque functioning (‘black box’) of algorithms. Biases in the data or incorrect settings can then lead to costly estimation errors. - Regulatory limits and legal responsibilities
AI-based assessment tools do not replace regulatory expertise in certain contexts, particularly for legal or tax assessments. Legal responsibility remains with certified human experts, and AI should be considered a decision-making tool, not an end in itself. - Data protection and transparency
The widespread use of personal and property data raises questions about privacy, security and ethics. It is crucial that solutions comply with applicable regulations (GDPR, Swiss data protection legislation) and that users have access to transparent information about how the models work.
In practice, professional intervention remains essential for high-stakes transactions or for investment properties.
2. Future prospects
The evolution of AI models suggests new practices:
- Hybrid AI + expert estimation
The future seems to be shaping up around enhanced collaboration between AI and human experts: AI performs a quick and comprehensive initial analysis of the data, while the expert steps in to validate, refine and contextualise the estimate. This hybrid model combines speed and reliability, ensuring better quality assessments. - Real-time data integration
AI models will gradually incorporate dynamic, real-time data such as rent trends, air quality and smart city indicators. Entire neighbourhoods could thus be optimised for quality of life thanks to intelligent systems combining energy management, connected mobility and adaptive infrastructure. - Explainable AI
To boost user confidence, valuation models will become <<explainable>>, meaning they’ll be able to precisely justify each estimate by detailing the factors and data that led to that valuation. This transparency will make it easier for professionals and individuals to accept the results. - Advanced Customisation
AI will offer increasingly refined personalisation, taking into account not only traditional property criteria, but also the lifestyle, interests and specific needs of buyers or tenants. This will enable tailored recommendations to be made, going beyond simple price or location.
Ultimately, these innovations could streamline the Swiss market, speed up the granting of loans and secure investments.
Artificial intelligence is transforming property valuation in Switzerland by offering faster and more accurate estimates thanks to the integration of a wide range of data. However, limitations related to data and atypical properties confirm the importance of human expertise.
The future is shaping up to be one of hybrid collaboration, where AI performs an initial analysis that is then validated by experts. Future advances will focus on real-time data, explainable AI and increased personalisation, combining technological performance with human judgement. AI complements humans, ushering in a new era for property valuation in Switzerland.
Search