What happens when AI does the agent’s job.
You paid an agent for three things: a price, an audience of buyers, and someone to run the deal. Software now does all three. Watch what that does to a fee built on scarcity.
The argument
The real estate commission was never a fee for opening doors. It bought you access: recent sale prices, a list of buyers, and a process you only do a few times in your life. For decades, the agent held all three. Today, software holds them better.
Strip the commission back to what it actually bought, and there are only three jobs underneath it. Price the home. Put it in front of buyers. Run the negotiation to a result. Every one of those jobs has been automated, and the percentage on top hasn’t moved.
Robots aren’t replacing people. A fee priced for information scarcity is just outliving the scarcity, into a world where the information is free.
The short version
Four things worth knowing.
3 jobs
Pricing, buyer reach and negotiation are what a selling fee actually pays for. All three are now done by software, in the open, at near-zero marginal cost.
2–3%
The commission rate hasn’t changed since before the internet, even as the tools behind it were rebuilt from scratch.
Public data
Comparable sold prices, suburb trends and median values are now public. The pricing advantage an agent once sold is available to anyone.
The portals
Buyers begin on realestate.com.au and domain.com.au, not in an agent’s contact book. Buyers moved online and stayed there.
Then and now
The three jobs you were paying for.
Ask a seller what the commission is for and the answer is usually vague: “the agent sells the house.” But selling a house breaks into a small number of concrete tasks. Each one below: what it used to require, and how it’s done in 2026.
| The job | What it used to require | How it is done now |
|---|---|---|
| Pricing the home | An agent’s “feel for the area” and private access to recent sale prices | Automated comparable-sales analysis from public sold data |
| Reaching buyers | The agent’s database, window display and newspaper column | National property portals every active buyer already uses |
| Matching buyers | Phone calls to a hand-kept contact list | Search alerts and buyer-matching software, instant and at scale |
| Running the campaign | An agent coordinating photography, copy and listings by hand | Guided onboarding that assembles the campaign in a day |
| Negotiation | A closed back-and-forth only the agent could see | Transparent bidding where every buyer competes in the open |
| Availability | One person, juggling a dozen other listings | Software that does not take a weekend off or chase the next listing |
Read down that right-hand column. None of it depends on a percentage of your sale price. The cost of valuing a $700,000 home and a $1.7M home with the same software is identical.
A fee for scarcity
Where the fee came from.
Percentage commission made sense in a world where the agent was the only way in. They held the comparable sales, the buyers, and the shopfront on the high street. What you bought was access nobody else had, and a percentage was a reasonable way to price it.
That world ended when buyers moved online. Sold-price data went public. Valuation models trained on millions of transactions started outperforming gut feel. Once the scarcity was gone, the fee should have followed. It didn’t, because it is deducted at settlement, once, by people who rarely sell twice.
“Once buyers could see everything an agent could see, the percentage was no longer paying for skill. It was paying for habit.“
Ben Williams · Founder, Unreserved
What the AI actually does
Not “AI-powered.” Named, and specific.
Naming “AI” without saying what it does is marketing. This is the work that replaces the three jobs, for a flat fee instead of a percentage.
AI Valuation Tool
Runs comparable-sales analysis on real market data to price a home from evidence, not a listing pitch.
AI Pre-Sale Advisor
Handles onboarding without an agent visiting your kitchen to win the listing.
Bid-Smart
A reverse-auction mechanic that flips the bidding so buyers compete transparently, not blind against a single gatekeeper.
What this means for sellers
The question to ask before you sign.
The question is no longer “are they good?” It is “what, specifically, am I paying a percentage for that software does not already do?” If the honest answer is pricing, reach and negotiation, then you’re paying today’s percentage for tools that existed before the internet. The tools are available directly. The only thing standing between most sellers and tens of thousands of dollars is the assumption that an agent is still the only door in.
Notes & sources
This briefing describes the shift in who performs the core tasks of a home sale, not any individual agent’s competence. The “then and now” comparison reflects the agent-mediated model that established prevailing commission rates against tools available directly to sellers in 2026.
Typical residential commission of 2–3% is drawn from 2025–26 industry commission surveys. “Public data” refers to the now-routine availability of sold-price and median-value information through property data services and portals. Unreserved is licensed to sell real estate in five states (VIC, NSW, QLD, SA, WA). Background data and interviews available to media on request.
For journalists
Ben Williams is available for interview on AI in real estate, the economics of agent commission, and what automation does to a percentage-based fee. We can walk through the valuation model, the Bid-Smart mechanic, and the data behind this briefing.
Media enquiries: ben@unreservedrealestate.com.au · Press centre: Press & Media
See what the AI values your home at.
Get a free AI valuation, priced from real comparable sales, not a listing pitch.
Get your free valuation