AI for Real Estate: What Agents Need to Know (2026)

I'm Chris from Studio Slate. We built the website for Ashworth & Co, a family homebuilder that needed to own its brand outside the portal ecosystem, and we build custom AI systems for real estate agencies that integrate with Rex, Reapit, and the major Australian CRMs. This guide is for principals and senior agents who are hearing about AI from every direction and want to know what is actually worth their time and money.

The adoption gap: everyone is using AI, almost nobody is using it well

The headline numbers sound like AI has already taken over real estate. 82% of agents have integrated AI tools into their business according to Realtors Property Resource (RPR), a NAR subsidiary. 97% of brokerages report their agents are using AI, up from 80% two years ago according to Delta Media Group's 2026 Real Estate AI & Leadership Survey.

But the impact numbers tell a different story.

The NAR 2025 Technology Survey of 1,241 active agents found that 46% said AI has had no noticeable impact on their business. Only 17% reported a significant positive impact. Another 33% saw moderate improvement. That means nearly half the industry is using AI and getting nothing measurable from it.

The reason is straightforward. Look at where agents are actually using AI:

The top three use cases are all content generation. Agents are using AI to write faster, not to think differently about their pipeline. The tools making the biggest dent (ChatGPT at 58%, Google Gemini at 20%, Microsoft Copilot at 15%) are the same general-purpose tools every industry uses. Only 39% use AI for market analysis or pricing, which is where the actual competitive advantage sits.

Property description generators: the commodity

Writing listing descriptions is the gateway drug of real estate AI. It is useful, easy to adopt, and produces immediately visible results. It is also the part of AI that is already a commodity.

Every major Australian CRM now has description generation built in. Rex AI generates ad copy, SMS, and email from listing data in a single click. Reapit (formerly Agentbox) is building its RAI platform with similar capabilities. Soho AI auto-drafts descriptions that get reviewed by an editorial team. Proply has a ChatGPT-powered content assistant embedded in its proposal builder. ActivePipe generates property alerts and nurture emails automatically.

And none of them are doing anything you cannot do with ChatGPT and a good prompt.

A listing description that used to take 30 to 60 minutes now takes under two minutes. That is genuine time savings. If an agent writes 10 listings a week, reclaiming 5 hours is meaningful. But the output quality is essentially identical across all tools because they use the same underlying language models. The "real estate AI" wrapper adds CRM integration and portal formatting, not better writing.

The 17% of agents reporting significant AI impact are not the ones who found a better description generator. They are the ones who moved past content generation into workflow automation and data-driven prospecting.

Lead response and nurture: where automation makes money

The real estate industry has a response time problem. According to REA Group's own 2024 data, 48% of portal enquiries go completely unanswered. The average agent response time to a new lead is 3 to 4 hours. Research from MIT shows that 50% of leads are lost if not contacted within 5 minutes. 78% of buyers go with the first agent to respond per NAR data.

This is where AI automation delivers measurable ROI, and a growing number of Australian-built tools target this exact gap.

PropertyPilot, a Brisbane-founded platform, connects to Rex, Reapit, AgentBox, VaultRE, and PropertyMe to automate lead response in under 90 seconds. It runs 18 automation products across sales, listings, and property management: missed call text-back, lead follow-up sequences, after-hours responders, OFI smart check-in, vendor communication, and dormant lead remarketing.

Agent Alpha takes a similar approach. It sits between your CRM and your team, watching for events and triggering AI-crafted responses, inspection reminders, and weekly vendor reports. It claims to go live in under five minutes by connecting via API to Rex, Reapit, or Agentbox.

Rex CRM has its own AI-powered prospecting tools built into the platform, and Reapit is building agentic AI capabilities into its roadmap under the RAI brand, where the system does not just respond to triggers but makes decisions about what to do next.

The published numbers from agencies deploying automation in Australia:

Whether these numbers hold across all agency types and markets is a separate question. But the principle is sound: speed of response is a competitive moat in real estate, and AI automation makes it possible to respond to every enquiry instantly, including the ones that arrive at 10pm on a Friday.

CMA and market analysis: where agents want AI to go

When asked what they want AI to do better, agents consistently point to the same things: clean and accurate CMAs, better pricing narratives, client-ready reports with improved market analysis, and hyper-local content generation.

This is the gap between what AI does today (writes descriptions) and what agents need it to do (wins listings). A well-prepared comparative market analysis is one of the most important tools in a listing presentation. It takes 2 to 3 hours to prepare manually: pulling comparable sales, adjusting for property differences, analysing days on market, and building a vendor-ready document. AI can compress that to minutes by pulling data from MLS systems and public records, but confidence in AI for pricing and market data remains low. Only 39% of agents use AI for market analysis, and agents specifically flag misinterpretation of market data (47% concern) and compliance risks (49%) as barriers.

In the Australian context, automated CMA quality depends entirely on the data source feeding the model. Which brings us to the part of AI in real estate that actually matters.

The Australian data landscape: CoreLogic, PropTrack, and Domain

AI tools are only as good as the data they run on. In Australian real estate, three platforms control the data layer that every AI application depends on.

Cotality (formerly CoreLogic) is the institutional standard. It covers over 14 million Australian properties with 200 million property attribute records and 950 million digital assets. Its Automated Valuation Models (AVMs) are used by over 80% of Australia's top 50 residential lenders. Nearly 90% of its AVM valuations fall within 15% of actual sale price, with accuracy improving 8% since 2023 through machine learning. Professional subscriptions start at $500+ per month.

Cotality's most interesting AI product for agents is Propensity to List: a predictive model that identifies properties likely to be listed for sale within the next three months. It analyses platform activity, transfers, listings, and market trends. The model delivers a customer segment that is 15 times more accurate than a random selection. This is the kind of AI that wins listings before competitors even know the vendor is thinking about selling.

PropTrack is owned by REA Group and integrated directly with realestate.com.au. It offers real-time APIs covering over 12 million properties with data on transactions, property attributes, sale history, listing history, and suburb-level market statistics (supply and demand, rent insights, sale insights). Its AVM performs at close to 90% accuracy within a 15% margin. The tight integration with REA gives PropTrack a unique advantage: it can correlate listing browsing behaviour with buyer intent in ways standalone data providers cannot.

Domain (owned by CoStar Group) provides its own property data and listing platform. It is the second-largest portal in Australia and offers data feeds used by CRMs and proptech platforms.

Archistar is a specialist platform for developers and investors. It covers over 130,000 users and 1,000 property firms, offering AI-powered site analysis, feasibility calculators, and 3D generative design. It integrates with Domain and CoreLogic data. For agents servicing developer clients, Archistar is relevant. For residential sales agents, it is not.

A newer entrant, Microburbs, has published a detailed methodology whitepaper arguing that the major providers have structural conflicts of interest (PropTrack is owned by REA, Cotality serves lenders and agents) that shape how they present data. Microburbs claims a 6% AVM error rate on validated transactions versus 13% for CoreLogic. Whether that claim holds across all property types is worth scrutiny, but the argument about incentive structures is worth considering when evaluating any data provider.

What is overhyped

Not all AI in real estate is created equal. Some categories get disproportionate attention relative to the value they deliver.

"AI-powered" CRMs that are just ChatGPT wrappers. As Ashley Stinton, managing partner of NAR REACH (NAR's technology accelerator), told RISMedia: "I think AI as a buzzword has become a little bit overhyped, and that can be noisy. Everything is AI now." Multiple industry leaders describe a pattern of vendors adding a ChatGPT layer on top of existing products and calling it "AI-powered." The listing description generator that comes with your CRM is useful. It is not a competitive advantage when every agent on the same platform has access to the same feature.

Chatbot-style lead capture on websites. These tools promise to qualify leads through automated conversation. In practice, most produce rigid, scripted interactions that sophisticated buyers see through immediately. The 47% of agents using chatbots or AI assistants per RPR data have not abandoned human follow-up. A chatbot that responds instantly and hands off to a human within minutes is useful. A chatbot that tries to replace the human conversation is not.

Virtual staging as a market differentiator. AI virtual staging is a real and useful tool, but it is commoditising rapidly. When every listing has AI-staged photos, none of them stand out. The technology has also raised compliance and misrepresentation concerns under Australian Consumer Law, where property photos must accurately represent the property.

Suburb prediction tools for agents. Tools like TUDI and BuyersBuyers offer AI-powered suburb analysis for investors, processing demographics, employment data, infrastructure approvals, and development pipelines. Early adopters report 12-18% gains versus 6-9% from traditional methods. These tools are useful for investor-focused buyer's agents. For a principal running a sales agency in a defined territory, suburb prediction does not change the day-to-day business of winning and selling listings.

The pattern across all overhyped categories is the same: the AI solves a problem that was not the bottleneck. Writing a listing description 28 minutes faster does not win you the listing. Responding to an enquiry 3 hours faster does.

Privacy: the part most agents are ignoring

The OAIC launched its first privacy compliance sweep in January 2026, targeting approximately 60 businesses across six sectors. Rental and property is one of the six, specifically because agents routinely collect driver's licences, passports, payslips, and bank details during property inspections and tenancy applications.

This directly affects how agencies use AI.

The OAIC's guidance on commercially available AI products makes the position clear:

1. Privacy obligations apply to any personal information input into an AI system, and to the output data generated by AI where it contains personal information.
2. Businesses should update their privacy policies with clear information about how they use AI.
3. If AI systems generate or infer personal information, this counts as a "collection" under the Privacy Act and must comply with APP 3.
4. The OAIC recommends organisations do not enter personal information into publicly available AI tools.

For a real estate agency, this means:

• Do not paste client details into ChatGPT, Claude, or Gemini. These are public tools operated by overseas companies. Entering a vendor's name, address, financial situation, or personal circumstances into a public AI model is a privacy risk the OAIC has specifically flagged.
• CRM-embedded AI is safer. Tools like Rex AI and Reapit AI operate within the CRM's existing data handling framework. The data does not leave the platform in the same way it does when you paste it into a public chatbot.
• Audit your data flows. Map what information is collected at each stage of your sales or rental process and confirm it aligns with your privacy policy. If AI is part of the workflow, the privacy policy must say so.
• Non-compliance penalties are real. Up to $66,000 per infringement for privacy policy breaches, with higher civil penalties for serious or persistent non-compliance.

From 10 December 2026, new APP 1.7 to 1.9 requirements come into force requiring transparency about automated decision-making. If your AI system makes decisions that substantially affect individuals (like scoring buyer intent or ranking vendor prospects), your privacy policy must disclose that.

A practical adoption path for Australian agencies

Based on where the technology is, where the ROI sits, and what the compliance landscape looks like, here is a sequence that makes sense for a principal-level agency in 2026.

Month 1: Use ChatGPT for content, properly. Get every agent on your team using a structured prompt for listing descriptions, vendor reports, and social media posts. Do not pay for a specialist tool yet. The $25/month ChatGPT Plus subscription handles 90% of content generation needs. Establish a rule: no client names, addresses, or financial details go into public AI tools.

Month 2: Audit your CRM's existing AI features. Rex AI, Reapit AI, and most modern CRMs already include AI features you may not be using. Description generation, communication drafts, AI-powered filters, and lead scoring are often included in existing subscriptions. Use what you already pay for before adding new tools.

Month 3: Automate lead response. If your agency is not responding to portal enquiries within 5 minutes, this is the highest-ROI AI investment you can make. PropertyPilot, Agent Alpha, or a custom automation built on Make.com or n8n that connects your CRM to an instant-response workflow. The cost is $200 to $500 per month. The return is every lead your competitors lose by responding 3 hours late.

Month 6: Evaluate data-layer tools. If your agency sells at volume and you need better prospecting intelligence, Cotality's Propensity to List or similar predictive products become relevant. These are not cheap ($500+ per month) and only justify the cost if you convert the intelligence into listings. A single additional listing per year pays for the subscription.

Month 12: Consider custom AI. Once you understand what generic tools cannot do for your specific workflows, vendor reporting cadence, buyer matching criteria, and listing presentation process, a custom build starts to make sense. This is where the AI is trained on your data, written in your principal's voice, and integrated with your exact CRM and portal feeds.

Real Estate Marketing Online: What Actually Works in Australia

Portal costs are rising 7-8% per year. How agents build owned digital assets that generate leads without portal dependency.

Read more

Where this goes next

The Colliers 2026 Outlook Report states that 92% of real estate professionals are running AI pilots. Australia's PropTech market was valued at $1.83 billion in 2025 and is forecast to reach $6.9 billion by 2035. Morgan Stanley projects AI will automate 37% of real estate tasks by 2030.

The direction is clear. The tools that win will not be the ones that write the best listing description. They will be the ones that identify the vendor three months before they list, respond to the buyer 90 seconds after enquiry, and produce a CMA that is both data-accurate and presentation-ready in the time it takes a competitor to open a spreadsheet.

The 17% of agents seeing significant results from AI are already operating this way. The remaining 83% are using ChatGPT to write social media captions. Both are technically "using AI." The difference in outcomes is everything.

Sources

• NAR 2025 Technology Survey - 1,241 agents surveyed, AI adoption at 68%, ChatGPT dominant tool, 46% report no noticeable impact
• RPR/NAR AI Adoption Survey 2026 - 225 agents, 82% using AI, 68% for listing descriptions, 71% cite time-saving as top benefit
• Delta Media Group 2026 Real Estate AI & Leadership Survey - 97% of brokerages report agents using AI, up from 80% in 2024
• RPR Agent Confidence Survey - 52% not confident in AI-generated content, 63% cite accuracy concerns, agents want better CMAs
• RISMedia: Real Estate's AI Hype - NAR REACH managing partner on "AI as buzzword" overhype, chatbot wrapper phenomenon
• Cotality (CoreLogic) Data - 14M+ AU properties, 200M attribute records, 950M digital assets
• Cotality Propensity to List - Predictive model, 15x accuracy versus random selection
• PropTrack Products - 12M+ properties, real-time APIs, AVM, market statistics
• PropertyGo: How AI and PropTech Are Changing AU Property - Cotality and PropTrack AVM accuracy (~90% within 15%), $1.83B PropTech market
• Rex AI Features - AI Ads, AI SMS & Email, AI Filters launched June 2024
• Reapit AI Roadmap - RAI platform, automation to agentic AI progression
• PropertyPilot - AU lead response automation, 18 products, REA Group and MIT response time data
• Agent Alpha - CRM-agnostic automation layer for AU real estate
• Microburbs Methodology Whitepaper - Independent AVM validation, data provider conflict-of-interest analysis
• Property Investment Professionals: AI Suburb Prediction 2026 - TUDI, CoreLogic, 12-18% investor gains, tool pricing
• OAIC: Privacy and Commercial AI Products - AI privacy obligations, APP compliance
• Lexology: OAIC 2026 Privacy Sweep - Rental and property sector targeted, $66,000 per infringement
• Real Estate Business: OAIC Privacy Compliance for RE Agencies - Practical implications, data flow audit recommendations
• MinterEllison: OAIC Privacy Enforcement - Bunnings precedent, ADM transparency from December 2026
• Datagrid: AI Automates CMA - Morgan Stanley 37% task automation projection, CMA workflow
• Flowtivity: AI for RE Agents Australia - Listing creation time benchmarks (30-60 min to under 2 min)