Feature article

AI in NZ real estate: where we are, and where we are going

What impact will Artificial Intelligence have on the local real estate industry today, and in the future?

8 September 2023

What is AI, and how will it impact real estate?

Without a doubt, Artificial Intelligence (AI) is the technology trend of 2023. But what does that mean for Aotearoa’s real estate industry today, and in the future?

Alan Clark, Head of Trade Me Classifieds says short term, AI has created a multitude of new possibilities to help real estate agents operate more efficiently. And long term?

“If it continues its current track, the way we interact with the technology is very likely to  impact industry sentiment and create some unique opportunities for agents,” he predicts.  

The rise of generative AI

While the term ‘AI’ has suddenly become incredibly buzzy, we’ve all been interacting with it for years.

“It’s the technology behind chatbots, and at Trade Me we use it in various ways, including to power search and recommendation algorithms, and generate our automated property valuations,” says Mr Clark.  

However, there’s been a rise in a new type of AI, known as generative AI, that has created meaningful change in capabilities. 

Generative AI has the ability to create text, images, music, or other media based on data that it’s scanned and synthesised from the internet. 

The most commonly recognised model is ChatGPT. ChatGPT’s algorithms crawl through the internet, including sources such as Wikipedia and all open books, training themselves how to recognise and respond to questions. Its learnings grow exponentially - hence the sudden explosion in capabilities we’ve seen in the last couple of years.

How is the real estate industry currently using AI?

From an agent perspective, generative AI can make today’s activities more efficient. 

You can use tools to automate building comparative market appraisals, create listing descriptions, or schedule viewings. It can also assist with jobs such as responding to queries about some documented feature of the house, suburb, or local planning laws. 

Mr Clark says regulatory concerns around the tools are valid though. 

“Fair trading rules will apply to AI-created marketing materials, and agents are accountable for providing accurate information to vendors, whether or not they use AI to collate that information. However, we are already at a stage where ChatGPT can pass the American Bar Examination with a 90th percentile score so accuracy should continue to increase in all fields.”

Real estate remains a highly emotional experience for most Kiwi, and we believe human connection will remain critical, says Alan Clark, Head of Trade Me Classifieds

The irreplaceable value of agents

Despite its utility AI is limited by, well, being artificial. 

“Real estate remains a highly emotional experience for most Kiwi, and we believe human connection will remain critical,” explains Mr Clark. 

“Phrases like ‘falling in love with a house’ and ‘a house is not a home’ are common. When it comes to finding an agent, most Kiwi rate trust and understanding as the top reasons for their choice^.” 

While AI may replicate suburb knowledge and efficiency, it’s unlikely to build trust the way that top agents do. A study by the American Psychological Association showed that artificial intelligence based preachers in Japan had less trust and credibility, and received fewer donations from church goers, than human preachers. 

Furthermore, as there is a rise of deep fakes (using AI to create imagery, audio or video that impersonates someone) online, researchers predict the public will have an increased suspicion around online interactions^^. This in turn may lead to an increased value placed on offline connections.

Mr Clark points out that AI can also only pull from data available, and agents often have information not recorded in any database. “Knowing and interpreting information like the sentiment of open home visitors, or how much the vendors want a sale to finalise a divorce will remain highly valuable.”

International trends to track

Globally, leaders like US real estate portal Zillow and Google are experimenting with better ways for buyers to find their perfect property. Interfaces that can help buyers find options that meet their criteria, such as character houses perfect for teenagers and remote working aren’t far from being ubiquitous. 

For vendors, potential applications include being able to present their house with different decor, furnishing or landscaping (Chinese property app Anjuke has announced a similar tool) or automatically create floor plans.

What’s next for AI in real estate?

Trade Me Chief Technology Officer Paolo Ragone says for agents, data will become richer, and processes smoother as companies continue to invest and innovate in the space. 

“Here at Trade Me we’re already experimenting with using generative AI to create listing descriptions, improve advice given to buyers and sellers, and identify potential vendors amongst the millions of Kiwi that browse the site each day,” he says.  

“Our dedicated data science team is evaluating new products and technologies and building our own proprietary models - all with the aim of improving the experience for buyers, sellers, and agents.” 

As models become increasingly more sophisticated, generative AI is a space that continues to offer new opportunities for innovation. 

“We are betting on (and working towards) AI continuing to evolve to support - not replace - the real estate industry; and that’s an exciting path to be on,” says Mr Clark.  

^ Property buyer and seller survey, March 2023

^^ Suspicious Minds: the Problem of Trust and Conversational Agents, April 2023