Real Estate Monitor: ChatGPT and the metaverse in the real estate sector

ChatGPT, the metaverse, and more: How AI and other technologies are changing the real estate market

ChatGPT and the metaverse are new topics that are affecting the real estate sector. But what advantages do AI tools and digitalization offer? Although the benefits of AI are still limited today, combined with other technologies it offers enormous potential. Despite the opportunities that chatbots and the metaverse offer for the real estate sector, there are also various risks that need to be considered when using them.

Chatbots in the real estate sector

Thanks to the speed of technological development, chatbots such as ChatGPT and Bard have recently seen rapid improvements, and this has opened up new opportunities and fields of application in the real estate sector. It has become possible to produce well-written texts and answer complex questions in next to no time. Additionally, tools such as DALL-E, Midjourney, and Stable Diffusion allow users to generate images from text.

Chatbots are already being used to provide customer services and product advice: Especially in the case of repeated inquiries, the output is of such a high standard that it is practically impossible to distinguish it from natural language. But how do such tools perform in the real estate sector? We put the tools to the test and looked at the following specific issues:

  • Real estate marketing and management with regard to writing real estate ads
  • Real estate construction and planning with regard to creating floor plans
  • Market analysis and house hunting with regard to finding the right neighborhood
  • Real estate valuation with regard to real estate price appraisal

In our assessment of these use cases, we found that the tools are already very useful in some areas. However, they also produce useless or incorrect results. In other words, it's always necessary to take a critical look at what these tools have produced. What should also be emphasized is how easy it is to interact with the tools: They were all ready to be used in less than ten minutes.

Interaction with AI using natural language

ChatGPT and Bard are chatbots that users can interact with using natural language. Users ask questions in the form of a text or an image plus the task that they want the tools to perform. The input is then processed by artificial intelligence (AI) based on large language models (LLM). The output is provided in text form or as an image.
The questions asked are then saved and incorporated into the answers to subsequent questions. This enables natural dialogue.

How large language models (LLM) work

Architecture of large language models (LLM)

Simplified illustration of the functioning of a large language model.

Source: OpenAI:; Google: LaMDA: Language models for dialogue applications; Credit Suisse

The AI revolution in the real estate sector has only just begun

The main difficulties currently lie in the risk of incorrect answers and in validating the output, as, by default, no sources are provided. However, solutions for the latter problem are already available.

At the moment, it's not clear what exactly happens to the input data. It is therefore not advisable to enter confidential data, since all the data is stored and used to improve the model.

Good datasets are crucial for obtaining high-quality results, but these are often not yet available. Nevertheless, chatbots will significantly improve day-to-day office work in many places in the future.

If language models are linked to databases or specialized algorithms and tools, we are likely to see game-changing improvements in quality. It will also be interesting to see how regulators react to the speed of progress, as there is also huge potential for misuse of the technology. The question also arises as to whether output generated by AI must be accordingly flagged as such, as it is practically impossible to distinguish whether the content was written by a human or a machine.

All in all, the recent progress made by chatbots has been impressive. Users already have enormous opportunities for speeding up and automating processes. And progress is likely to continue over the next few years.

The number of VR and AR glasses sold will increase

More widespread use of VR and AR technology expected

Forecast number of globally sold units of VR and AR glasses, in millions.

Source: International Data Corporation

Digital real estate in the metaverse

Real estate developers and companies from other sectors are already investing millions in the acquisition of virtual properties. At first glance, the acquisition of such a plot is similar to the purchase of land in the physical world. In particular, it's also the property's location that determines the asking price in the metaverse: Attractive locations cost more.

Metaverse users can view the plots of land on sale on a plan similar to a cadastral plan and submit a bid to the seller at the click of the mouse. Payments in the metaverse are carried out using cryptocurrencies. The digital goods purchased are secured using non-fungible tokens (NFTs) – digital certificates that verify the ownership of virtual assets, such as real estate and buildings, on a blockchain. The latter functions as digital and decentralized land records. After the virtual piece of land has been acquired, a property is typically built on it that is used for commercial purposes.

Hurdles for the virtual real estate market

Owning real estate in the metaverse offers retailers the opportunity to promote or sell their products. Companies can also strengthen their brand by having additional visibility in the metaverse. And there are other promising areas worth exploring for the real estate market. For example, real estate service providers can create a virtual twin of a physical apartment in the metaverse, which interested parties can view or furnish conveniently from home using VR glasses.

The use cases for the real estate industry appear diverse, but they have one thing in common: Although such offerings are already possible using existing technology, there still seems to be little demand for expansion into the metaverse.

The success of the metaverse will depend, firstly, on whether people choose to spend more time in the virtual space in the future. Secondly, it will depend on how companies view the trade-off between the opportunities and risks of being present in the metaverse.

Companies face a number of hurdles: the high price volatility of real estate trading in the metaverse and the cryptocurrencies required; uncertainty regarding the penetration of virtual platforms; and legal uncertainties. Finally, the availability of technologies on which the metaverse is built will also play a role.

It is important to note that the transition to a new technology involves considerable costs and therefore takes time. For this reason, it's far too early to talk about a metaverse revolution. Nevertheless, it would be premature to write off the metaverse completely.

Google trends: Relative frequency of searches on the metaverse

Interest in the metaverse has recently declined

Google Trends: Relative frequency of “metaverse” search queries, Switzerland and worldwide, maximum = 100

Source: Google, Credit Suisse

The Swiss real estate market remains on track

Unlike the virtual real estate market, the analog Swiss real estate market remains in the grip of inflationary and interest rate trends. The most recent figures show that inflation has not yet been tamed. Second-round effects, such as the upcoming rent increases due to the rise in the reference interest rate, are preventing a rapid drop in inflation. In this challenging market environment, the Swiss real estate market is nevertheless performing significantly better than many of its counterparts abroad.

Although higher interest rates are significantly dampening demand, the supply of residential property is increasing only slowly as a result of the current shortage. At 1.8% for condominiums and 1.6% for single-family dwellings, supply rates remain low. As a result of falling demand, however, price growth weakened significantly in the first quarter of 2023. Asking prices for condominiums rose by 3.5% over the course of the year, while those for single-family dwellings rose by 3.6%. We anticipate only a slight increase in prices by the end of the year. However, from 2024, prices are expected to fall due to the persistent decline in demand.

House hunting is becoming more challenging

Investors in multi-family dwellings with rental apartments, by contrast, should already expect to see falling prices in the current year. However, improving earnings prospects will mitigate the potential price correction.

The Swiss rental housing market continues to move rapidly toward a shortage. The term "housing crisis" overstates the current situation, but house hunting has become markedly more challenging. As a result, a significant decline in vacancy rates and higher rental prices can be expected both for new rentals and for existing rental agreements.

Real Estate Monitor Q2: Annual growth rates in advertised rents

Rental price development as a reflection of the supply situation

Annual growth rates in advertised rents by market region (Wüest Partner).

Source: Wuest Partner, Credit Suisse
Last data point: April 2022

What does the current situation on the real estate market mean for you personally? We would be happy to help.

Schedule a consultation This link target opens in a new window Download the Real Estate Monitor Q2 2023 This link target opens in a new window