A property manager I know told me she processed 147 rental applications in a single week last month. One hundred and forty-seven. For context, that’s across a portfolio of about forty properties in Sydney’s Inner West. The rental market is that competitive, and the administrative burden on property managers has become genuinely unsustainable.

So it’s no surprise that AI-powered tenant screening tools have arrived. The question is whether they’re actually an improvement — and whether they’re legal.

What These Tools Actually Do

Let me describe what’s on the market, because the terminology is vague and the marketing is predictably overblown.

The core function is automated application assessment. A tenant submits their rental application — identity documents, employment details, rental history, references — and the AI system processes all of it. It verifies documents, cross-references databases, checks for red flags, and produces a risk score or recommendation.

Some platforms go further. They analyse income stability by looking at payslip patterns rather than just a snapshot number. They check social media for potential concerns (which immediately raises my eyebrows). They cross-reference tenancy tribunal records across states. And a few claim to predict the likelihood of a tenancy going wrong based on pattern matching against historical data.

Companies like Snug, 2Apply, and a handful of newer entrants are offering various levels of AI-assisted screening. The major property management platforms — PropertyMe, Console, Rex — are also integrating screening features, though with varying degrees of sophistication.

The Speed Advantage Is Real

I’ll give credit where it’s due. These tools genuinely speed up the process. What used to take a property manager several hours — calling employers, chasing references, verifying TICA records, reviewing bank statements — can now happen in minutes for the automated portions.

For landlords, faster screening means shorter vacancy periods. For good tenants, it means they’re not waiting weeks to hear back while juggling multiple applications across different agencies.

One property management firm I spoke with said they’ve cut their average time-to-decision from five business days to less than 24 hours since implementing AI screening. That’s significant. In a market where tenants are sometimes applying for twenty properties simultaneously, speed matters.

The teams at Team400’s AI consultancy page have been working with property management companies on exactly this kind of workflow — figuring out where AI genuinely saves time versus where human judgment is still essential. It’s not a binary choice.

The Fairness Problem

Here’s where my enthusiasm cools considerably. Any system that uses pattern matching to predict human behaviour carries bias risk. And in rental screening, the consequences of biased decisions are severe — we’re talking about whether someone has a roof over their head.

Consider: if the AI learns from historical data showing that tenants from certain postcodes, age groups, or employment types are more likely to default, it will reproduce those patterns. It doesn’t matter whether the correlation is causal or coincidental. The algorithm doesn’t know the difference.

Australia’s Residential Tenancies Act (varying by state) prohibits discrimination based on protected attributes. The Australian Human Rights Commission has flagged AI decision-making in housing as an area of active concern. And yet, many of these screening tools are essentially black boxes — neither the landlord nor the tenant can fully explain why a particular score was assigned.

I’ve personally seen situations that worry me. A self-employed freelancer with excellent rental history getting flagged as “high risk” because the AI couldn’t pattern-match their income against standard employment models. A young couple being scored lower because they’d only had one previous tenancy. These aren’t hypothetical edge cases.

What Landlords and Property Managers Need to Know

Never let the AI make the final decision. Use it as a filter, not a judge. It shouldn’t be the sole reason you reject someone.

Understand what data the tool uses. If it’s scraping social media or using demographic proxies, you’re exposing yourself to discrimination claims. Ask the vendor directly.

Keep records. If a rejected applicant challenges your decision at the tribunal, “the computer said no” won’t fly. You need to show that a human reviewed the application.

Check state-specific regulations. Victoria, Queensland, and NSW all have different rules around tenant selection. The AI tool might not understand state nuances.

My Take After 25 Years

I’ve been in this industry long enough to remember when tenant screening meant calling someone’s boss and asking “do they seem reliable?” That wasn’t better. It was just differently biased — biased toward people who interviewed well, who had employers willing to take calls, who fit a subjective notion of “good tenant.”

AI screening has the potential to be more consistent and more objective than human screening. But only if it’s designed carefully, audited regularly, and used as a tool rather than an oracle.

The technology is here. The regulation hasn’t caught up. And the gap between those two things is where real harm can happen. Property managers who care about doing this right need to be asking harder questions of their software vendors than most are currently asking.

We owe that to our landlords, and we owe it to the tenants.