19 November 2026
Beyond the Listing:

What AI doesn't know about finding a home

New OLX Group research across Poland, Portugal and Romania explores what people really expect from AI when searching for a home.

 

A woman holding a smartphone looks out of a window during a property viewing, while a man examines the apartment behind her.

Finding a home takes more than spotting a property you like. There’s the neighbourhood to consider, prices to understand, options to compare and, eventually, a big decision to make.

AI can already help with many of these steps, from recommending relevant listings to providing property information and automated valuations. But what do people actually want from these tools, and where do they still prefer a human perspective?

New research among users of OLX Group’s real estate platforms in Poland, Portugal and Romania offers an interesting look at those expectations. The findings suggest that while AI has plenty to offer, home seekers place particular value on practical support, transparency and the ability to make informed decisions.

What home seekers really want from AI

When asked which AI-powered features would be most useful during their property search, respondents weren’t necessarily looking for anything complicated. Some of the most popular options focused on making everyday tasks easier and helping them find relevant information.

Alerts for new listings matching their criteria came first, selected by 30% of respondents. Comparing properties side by side, analysing neighbourhood information and explaining contracts or legal terms followed closely, each selected by 27%.

But preferences aren’t exactly the same everywhere. In Poland, new-listing alerts are particularly popular, while people in Portugal show greater interest in neighbourhood information and help understanding contracts and legal terminology.

These findings suggest that the value of AI in real estate isn’t necessarily about offering the most advanced technology. Often, it’s about helping people find the information they’re looking for more easily, save time, compare options and feel better equipped to make decisions throughout their property search.

When technology changes what we see

Property photos are often the first thing that catches someone’s attention when browsing listings. And while AI can help enhance images and show a property’s potential, people have different opinions about which modifications are acceptable.

Across the three markets, 42% of respondents consider altering room proportions or camera angles misleading or unacceptable. Another 40% say the same about digitally removing structural problems, such as cracks or water stains.

These concerns aren’t entirely surprising when we look at people’s experiences with property viewings. In Romania, 41% of respondents reported multiple occasions when a property looked significantly worse in person than it did in the listing. Similar experiences were reported in other markets, although the frequency varies.

Not all digital modifications are viewed in the same way, however. Changes to lighting or virtual staging, where furniture and décor are digitally added to property photos, tend to raise fewer objections than alterations that could affect how people perceive a property’s physical characteristics.

The findings point to an important distinction: technology can help people visualise a property’s potential, but they also want to understand what the property is actually like.

Transparency makes a difference

One of the clearest findings from the research is that people want to know when AI has been involved in creating or significantly altering listing content.

Nearly four in five respondents who answered the disclosure question believe platforms should indicate at least some forms of AI-generated or significantly altered content. The largest group, 63%, wants this information disclosed across photos, descriptions and videos.

There are differences between markets, too. Respondents in Poland are generally more cautious about AI-generated listing content, while those in Portugal show greater openness towards edited photographs and virtual staging. In Romania, AI-written property descriptions and neighbourhood summaries receive comparatively higher levels of acceptance.

This doesn’t mean people are rejecting AI-generated content. Rather, the results suggest that being clear about how technology has been used can help people understand the information they’re seeing and make their own decisions about it.

When AI and estate agents disagree

AI can also help people understand property prices through automated valuation tools, giving home seekers another source of information when exploring the market.

But when asked whose estimate they would trust if an AI tool and a human estate agent gave different valuations, respondents showed a clear preference for human expertise. Across the three markets, 36% would trust the agent’s estimate, compared with just 5% who would favour the AI tool.

In Portugal, the preference for human expertise was particularly strong, with 44% choosing the agent’s valuation.

Interestingly, another 21% of all respondents said they would trust neither estimate and would prefer to do their own research. 

The findings show that people don’t necessarily want technology to make every decision for them. AI can provide useful information and additional perspectives, while professional knowledge and personal research continue to play an important role, particularly when decisions involve significant financial commitments.

Building AI around what people actually need

Taken together, the research offers a useful perspective on how AI can support the property search experience. People value technology that makes information easier to access, helps them compare options and simplifies tasks, but they also want clarity about the content they see and the freedom to make their own decisions.

At OLX Group, we’re constantly exploring how AI can make finding and selling a home easier and more intuitive. Our Real Estate apps on ChatGPT and Toqi bring property search into more conversational experiences, allowing people to explore their options in new ways.  Deal Finder, our property search agent on WhatsApp, is another example of how we’re making it easier for home seekers to find properties through tools they already use.

On the selling side, Agentic Selling helps property sellers navigate the process of creating a listing, from understanding property values to preparing descriptions. The tool provides AI-powered guidance and suggestions while keeping important decisions, including the final asking price, in the seller’s hands. And it’s already gaining traction: 20% of private sellers on Otodom are using Agentic Selling. 

These are examples of how we’re building AI-native marketplaces around real customer needs, making complex tasks easier while keeping people informed and in control.

After all, finding a home is about more than discovering the right listing. It’s about having the information, confidence and support to make one of life’s biggest decisions.


About the research

The research surveyed users of OLX Group real estate platforms in Poland, Portugal and Romania, exploring their experiences and attitudes towards AI in the property search process.

The analysis covers 2,819 respondents across the three markets: 1,197 in Poland, 891 in Portugal and 731 in Romania. The number of responses varies by question, and the findings represent users of the surveyed platforms rather than the wider population.

The study also includes additional findings from Ukraine, which were analysed separately due to the smaller sample size.

AI in Real Estate: Your questions answered