The AI tool flagged the property three weeks before the owners contacted any agent.

How did it know? Combination of signals: recent council searches, power usage changes, kids graduating nearby schools, mortgage age matching typical refinancing patterns. Individually meaningless. Combined with machine learning, surprisingly predictive.

The agent who got the listing was thrilled. The vendor, when she learned how she’d been identified, was distinctly less comfortable.

This tension is becoming common across Australian real estate—and we’re not talking about it enough.

What’s Actually Happening

Predictive prospecting tools have improved dramatically. Platforms like PropTrack and various CRM add-ons now offer some form of seller prediction. Third-party data enrichment services compile surprisingly detailed profiles.

The inputs vary by provider but commonly include:

• Property ownership duration
• Demographic life stage indicators
• Financial behaviour patterns (aggregated)
• Property search activity
• Development application history
• Utility connection changes

Some of this data is public record. Some comes from data aggregators with murky collection practices. The models themselves are black boxes—vendors don’t explain exactly what signals they weight.

The Legal Reality

Australian privacy law offers limited protection here.

The Privacy Act 1988 applies primarily to organisations with over $3 million revenue—exempting most small agencies. Real estate practices handling personal information have some obligations under Australian Privacy Principles, but enforcement is rare.

The OAIC’s guidance on direct marketing requires ability to opt-out, but doesn’t prevent initial collection of public-record data or predictive modelling on aggregated information.

In practice: much of what these tools do is technically legal, even if ethically questionable.

Where It Gets Uncomfortable

Consider the signals that might indicate someone is likely to sell:

• Divorce filings (public record in some jurisdictions)
• Probate notifications
• Health-related property modifications
• Changes in vehicle registrations suggesting household changes

An AI doesn’t know it’s processing sensitive life events. It just sees patterns that correlate with listing behaviour. But the humans using these tools do understand—or should.

The question isn’t whether we can identify pre-listing households. The question is whether we should, and how.

What Good Practice Looks Like

Some principles are emerging among agencies taking this seriously:

Transparency about methods. If you’re using predictive tools, be honest when asked how you identified potential vendors. “Our system suggested you might be considering a change” is better than pretending coincidence.

Opt-out mechanisms. Maintain suppression lists for people who request not to receive prospecting contact. This goes beyond legal requirements but builds trust.

Human review of sensitive signals. Don’t let automation decide which recently-divorced households to contact. A human should evaluate whether outreach is appropriate given apparent circumstances.

Clear data retention policies. Don’t keep enriched prospect profiles indefinitely. If someone doesn’t respond to outreach, delete the analysis.

Vendor consent for success stories. If you used predictive tools to identify a vendor, don’t share their story in marketing without explicit consent.

The Competitive Pressure

Here’s the uncomfortable reality: agents who use these tools aggressively may win more listings in the short term. Those who hold back on ethical grounds may lose to competitors without such concerns.

Market forces don’t automatically reward ethical behaviour.

But there’s a counterargument: as consumers become aware of these practices, the agencies known for them may face reputational consequences. Trust is a long-term asset in real estate.

What Vendors Should Know

If you’re selling property, you should understand:

Your behaviour is being analysed. Searching on Domain or realestate.com.au, getting rates quotes, enquiring about renovations—these activities can trigger prospecting.

You can ask how agents found you. A direct question—“How did you know to contact me?”—may reveal more than you expect.

Off-market interest often has reasons. When an agent approaches with a buyer who “might be interested” before you’ve listed, consider why that buyer happens to be ready for your property specifically.

You have some control. Asking to be removed from prospecting lists won’t stop all contact, but legitimate agencies will comply.

The Regulatory Horizon

Privacy reform is coming, eventually. The Attorney-General’s review of the Privacy Act has recommended extending coverage to small businesses and strengthening individual rights.

If those changes become law, many current prospecting practices may need revision. Agencies building ethical frameworks now will adapt more easily than those forced to change later.

The Bottom Line

Technology that identifies likely sellers before they list isn’t inherently wrong. But how we use it matters.

The real estate industry has a choice: develop norms that balance effective prospecting with privacy respect, or wait for regulations to impose standards externally.

Self-regulation doesn’t have a great track record in real estate. But the alternative—government-mandated restrictions that may be blunter than necessary—isn’t obviously better.

We should be having this conversation more openly. The vendors whose life events we’re analysing deserve to know how the game works.