realestate.co.nz has launched an AI image search tool for property listings in New Zealand and says it is the first website in the country to bring the technology to market.
The feature lets users search for homes by visual elements in listing photos, rather than relying only on written descriptions. Searches can include terms such as weatherboard exterior, wooden flooring, home gym, or pool, with results drawn from images on the platform.
The system works in two stages. It first analyses each uploaded image and tags features such as room type, interior style, flooring, and details like custom shelving. A second layer then interprets search terms and related concepts to return results.
Users may therefore see listings that match both exact requests and closely related ideas. The tool can also identify broader or more subjective features, including a strong view or shade trees in a garden.
Sarah Wood, Chief Executive Officer of realestate.co.nz, said the product reflects a change in how people search for homes online. "In our 30-year history, we've developed a deep understanding of how people connect with property," Wood said.
"Property searches have traditionally depended on how accurately a home is described and whether all the key features a buyer is looking for are included in the listing description."
"But buying a home is an emotional decision. People respond to how a space looks and feels. This tool connects people with homes based on what they're genuinely drawn to, whether that's a swimming pool, home gym, wine cellar, or walk-in wardrobe!"
The launch adds a new search layer to a property portal that has operated since 1996 and describes itself as New Zealand's longest-running property website. It also points to a broader shift among online marketplaces towards using AI to make large image libraries searchable in more detailed ways.
Images uploaded to the platform sit within Snowflake's AI data cloud, which serves as the underlying data environment for the search system. Snowflake said the tool shows how AI can be applied to a large property dataset that includes photos, maps, addresses, and listings.
"realestate.co.nz is the leader in property intelligence, managing vast amounts of data including photos, maps, addresses, and listings," said Tony Shaw, Country Manager New Zealand at Snowflake.
"By leveraging Snowflake's Cortex AI, realestate.co.nz is using this wealth of data to personalise the customer experience with high accuracy and at significant scale, and this use of AI benefits all New Zealanders buying, selling, or leasing property," Shaw said.
Beta phase
The image search tool is labelled beta, although it is already live on the site. The label is intended to signal that AI image interpretation may not be fully accurate in every case while the product is monitored and refined.
Simon Hargraves, Chief Product and Technology Officer of realestate.co.nz, said the company is treating the feature as a live product while continuing to study how people use it.
"The beta label is mainly about setting the right expectations. With new AI-powered features, especially those involving image interpretation, it is important to acknowledge that accuracy may not be perfect in every instance. Labelling it as beta allows us to be transparent with users while we continue to monitor, learn, and improve the experience," Hargraves said.
He said the company is collecting usage data to understand which searches are most common and which property attributes matter most to users. That information is expected to shape how the tool evolves over time.
"We will be looking closely at the most commonly searched features and making sure the tool performs strongly where user demand is highest," Hargraves said.
Wood said the tool may also affect how agents and sellers approach listing photography, because images now play a more direct role in discovery. Stronger images that show a home's distinctive features will carry more weight when users search visually.
"This tool places greater importance on high-quality images that showcase a property's best or most unique features, because those are the images directly powering search results," Wood said.
The company has previously introduced other AI-driven search and listing features, including AI floor plan surfacing and a seller-focused assistant called Tāhū. "Thirty years ago, we brought property searching online. Now we're redefining it again," Wood said.