23% use generative AI when looking for a home, but 76% intend to — more than three times the current rate | a survey of 507 people

Itera, Inc. (head office: Shinjuku-ku, Tokyo; Representative Director: Takayuki Muto), which provides the AI search optimisation tool LLMOcheki, surveyed 507 people who have looked for somewhere to live — to rent or to buy — about how they gathered information.

23.47% had “asked a generative AI such as ChatGPT, Gemini or Copilot” and 24.26% had “read the AI summary at the top of Google search results (AI Overview)” — roughly one in four in each case. The most used source was property portal sites (55.23%): AI is not the main act today. Yet when asked whether they would want to use generative AI when looking for a home in future, 75.93% said they would.

The most researched topic via AI was “how safe and liveable an area is”, at 47.23%. Comparing specific properties came to only 20.05%, revealing a clear division of labour between property information and local information.

Background

As generative AI spreads, how consumers gather information is changing. But for a decision as expensive and as drawn out as finding somewhere to live, there has been little data on how much AI is actually used.

We therefore ran an online survey of 507 people who have looked for a home.

Key findings

• 23.47% asked a generative AI; 24.26% read an AI Overview
• The largest source was property portal sites (55.23%) — AI is not the main act
• Yet intent to use it in future stands at 75.93%, more than three times the current rate
• The most researched topic via AI was “how safe and liveable an area is” (47.23%)
• “Comparing and evaluating specific properties” came last at 20.05%
• 86.19% have some complaint about property portal sites
• On trust, portals score 49.70% against AI’s 9.86%

Finding 1: Sources used when looking for a home (n=507, multiple answers)

Property portal sites came first at 55.23%, a majority, making them the largest source.

Asking a generative AI reached 23.47% — more than ten points above social media such as Instagram, TikTok and YouTube (13.21%). Social media is widely discussed in the housing and property industry, yet as an actual means of gathering information, generative AI is used more.

Finding 2: What was researched with generative AI (n=245, multiple answers)

The most common was “how safe and liveable an area is” (47.23%). “Comparing and evaluating specific properties”, by contrast, came last at 20.05%.

The pattern suggests people search for the properties themselves on portal sites and turn to AI for the surrounding judgement calls.

“Recommended agencies and services” also reached 30.87%, showing that three in ten people consult AI about which company to go with.

Finding 3: Intent to use generative AI in future (n=507)

Combining “definitely want to use it” and “would use it given the chance”, 75.93% expressed intent to use.

That is more than three times the current usage rate (the 23.47% who asked a generative AI). Only 24.06% were clearly averse, suggesting that most of those not using AI today are not unwilling — they simply have not started.

Finding 4: Complaints about property portal sites (n=507, multiple answers)

Most of the complaints raised are of a kind that gathering information via AI could resolve. “Too many sales calls after an enquiry” (28.01%) in particular may become a reason to use AI, which lets people get information without handing over their contact details.

Finding 5: Trust in the information (n=507)

On trust, portal sites (49.70%) were far ahead of AI (9.86%).

However, 29.78% said “about the same for both”, so those rating AI at least equal come to 39.64% in total. For a new source of information, it has already earned a degree of trust.

What the results suggest

More than a threefold gap between current use and future intent

The most striking feature of this survey is the gap between current usage (23.47%) and intent to use in future (75.93%).

Portals also lead on trust, so generative AI is not the main act in home hunting today. But the strength of the intent points to that structure changing.

What people research with AI is the area, not the property

It is also notable that the most researched topic via AI was “how safe and liveable an area is” (47.23%) while “comparing and evaluating specific properties” came last at 20.05%.

Property search is completed on portal sites, while the judgement calls — what an area is actually like, what the going rate is — are increasingly put to AI.

What property businesses should do

We suggest the following for property businesses:

1. Put area information — safety, living environment, transport, school catchments — into words, down to what the area is actually like
2. Publish going-rate data by area and layout, stating the period covered and the number of cases
3. Explain contracts, upfront costs and mortgages, including the criteria for judging them
4. State concretely which areas you cover and what conditions you can handle
5. Check regularly how AI describes your company

Area information in particular is hard for a nationwide portal to provide — which makes it a field where a property business rooted in its area can hold an advantage.

About the survey

Percentages are rounded to two decimal places, so totals may not come to 100%.
Multiple-answer questions total more than 100%.

About LLMOcheki

LLMOcheki is a SaaS product that continuously measures how a brand appears and is cited in AI search across eight engines: ChatGPT, Gemini, Perplexity, Claude, Google AI Overview, AI Mode, SearchGPT and Copilot. Its “n-shot statistical measurement” runs the same question multiple times and expresses appearance as a probability with a 95% confidence interval, making the highly variable output of AI answers visible in a statistically dependable form. It covers measurement, competitor comparison, and content generation and improvement informed by the analysis — all in one place.

LLMOcheki https://llmocheki.com/