REA Group brings conversational AI search realestate.com.au

REA Group has started offering conversational search capability on its flagship website realestate.com.au after expanding natural language search to half of the site’s visitors.

The digital property advertising specialist revealed at its financial results briefing that it quietly launched conversational search capability in beta form last month and had since made it available to 10 percent of the site’s visitors.

REA Group chief executive Cameron McIntyre said that the conversational search beta is the next step in the company’s plan to evolve beyond filter-based search following a successful 12-month trial of natural language querying.

McIntrye said that while natural language search was a useful extension to keyword-based methods, it still limited the consumer experience to “search-by-search” input.

“Conversational search is quite a different experience because [it] involves an AI agent that's responding and it generates a conversation which takes you down Interesting more interesting pathways. It's more engaging from a consumer perspective,” McIntyre said.

He said, however, that the company is trialling natural language input alongside it first to give consumers time to adapt.

“We want to trial both because search in particular is an evolution, not a revolution and you've got to allow consumers to evolve [with] it over time, you don't want to rush it. If you rush it, you can come unstuck.

“We're taking a very responsible approach to this to ensure that we're still maintaining and building on our traffic and engagement with consumers which is why you're seeing multiple variations that we're testing,” he said.

McIntrye said conversational search is expected to increase buyer level of intent indicators.

During the briefing he provided the example of a buyer seeking a property with a tennis court but finding none available.

Ordinarily, that might bring the consumer’s engagement to an end, but the conversational search feature might prompt the buyer to enter a discussion about properties that can accommodate a court, increasing the chance that they will move toward a sale.

The AI-powered search features are the most visible part of REA Group’s optimism for the technology more broadly.

McIntyre said that the company has been using AI to accelerate its ability to bring new products to market and grow its capability.

Pointing to the example of search, he said that products that might have taken years to build were now being turned out in a matter of days.

He said that he saw acceleration of product deployment into market as being a prime metric for the success of the AI strategy.

“As we get product to market faster and address the needs of our customers and consumers faster, I think that sets us up for great outcomes,” he said.

However, the company also reported that AI technology costs were placing upward pressure on its operating expenses in Australia, causing it to break slightly higher than revenue growth in Australia at nine percent compared to eight percent for income.

REA Group’s chief financial officer Janelle Hopkins said that the company expected the increase to be offset by savings from its AI investments.

“AI cost and investment is increasing, but on the flip side, there is additional productivity that is starting to come through. We’re already evidencing it and it will continue to come through into the future,” Hopkins said.

The company has entered partnerships with Google and OpenAI to use their Gemini and ChatGPT agentic AI offerings respectively.

REA Group entered its partnership with OpenAI late last year.

The partnership allows the company to use technology underpinning ChatGPT to provide home a value tracking tool called realEstimate as part of its realestate.com.au platform.

The company uses Google’s Gemini in its financial services arm, Mortgage Choice, with brokers using the agentic AI technology as productivity tool.

REA has separately made its own investments in the technology in a bid to become what it calls an “AI prime” operation.

The investments include a 61.5 percent stake in Palantir, the owner of Candian AI and machine learning-driven 3D floorplan visualisation specialist, iGuide.

It also has a stake in Jitty, a UK-based AI property search company.

To support its strategy, it has also invested in specialist data providers, such as commercial real estate data provider Arealytics and neighbourlytics, an Australian firm that uses data to provide information about the social and lifestyle dimensions geographic areas.

Additionally, it uses AI across its internal operations with 90 percent of its Australian staff trained to use the technology at a foundational level.

REA said that 85 percent of those used the company’s internal AI assistant regularly and that 90 percent of employees across its global operations also used the technology.