The way Australians choose a real estate agent used to be reassuringly analogue: the board on the neighbour's lawn, the brother-in-law's recommendation, the agent who door-knocked last spring. Some of that still happens.
But a growing share of the decision now takes place somewhere no principal can see: inside a chat window, where a vendor types "who's the best agent to sell a family home in my suburb?" and an AI answers with names.
For an industry built on local reputation, this is a structural shift, not a marketing fad. The shortlist (the two or three agents who get the appraisal call) is increasingly drafted by ChatGPT, Gemini and Google's AI Overviews before any human conversation begins.
Agents who are named get the call. Agents who aren't never learn the call existed.
The research shift nobody's front desk can see
Vendors and buyers have always researched before getting in touch; what's changed is where. Instead of a dozen Google searches and a portal browse, they're asking one assistant compound, specific questions (which agents dominate this suburb, what's a fair commission, who handles downsizers well, is auction or private treaty right for a property like mine) and getting synthesised answers with recommendations attached.
The 2026 State of AI SEO in Real Estate industry benchmarks, measured across 150+ agents and 8.2 million tracked queries, put a number on how far this has already gone.
67%
of buyers now use AI as their primary research method before contacting an agent, per the 2026 State of AI SEO in Real Estate industry benchmarks
Read that carefully: primary research method, before contact. By the time a buyer or vendor rings an office, much of the evaluation is done.
The phone call is now the outcome of a process that ran inside an AI answer, not the start of the funnel.
The invisibility problem
The natural response is "fine, so am I showing up?" For most of the industry, the benchmark answer is blunt.
91%
of agents are effectively invisible in AI answers in their own market, per the 2026 State of AI SEO in Real Estate industry benchmarks
Effectively invisible, in their own market: the suburbs where they've sold for years. The usual causes are mundane: websites that never state the facts engines need (suburbs serviced, property specialties, recent results), review profiles that are thin or scattered across inconsistent listings, and a public record that doesn't corroborate the agent's actual track record.
None of that is malice or mystery. It's simply that nobody built the profile the machines read, which also means the gap is fixable, and right now, few competitors are fixing it.
What vendors and buyers are actually asking
It helps to be concrete about the queries at stake. On the vendor side: best agent to sell in a named suburb, agents with strong auction clearance for a property type, what commission is reasonable in this market, questions to ask at an appraisal.
On the buyer side: suburbs to consider on a budget, buyer's agents worth engaging, how to read a contract, whether now is a sensible time. Mortgage broking gets the same treatment: best broker for first-home buyers, broker versus bank, how much can I borrow.
Every one of those answers is a chance to be named, or a chance for the engine to hand your market to whoever made themselves citable.
What the benchmarks say happens when agents invest
The same 2026 industry benchmarks track what changes when agents do the work: a 250% average increase in AI traffic for agents who invest in GEO, 90 days or less to a #1 AI recommendation in their market, a 4.2× improvement in lead quality measured by 90-day close rate, and +74% more clicks per impression when AI cites you.
Those are industry benchmark figures, not guarantees: your suburb, competition and starting point all matter, which is why any serious programme measures visibility through to appraisals and listings rather than stopping at a citation count. But the direction is consistent: early movers in a market where 91% are absent capture outsized share, and lead quality improves because an AI-referred vendor arrives pre-sold on the recommendation.
A practical starting plan for agents and brokers
You can begin this week without a consultant. First, run the audit yourself: ask ChatGPT and Google who the best agents in your core suburbs are, and record who gets named.
Second, fix your entity: one consistent name, office address, phone and service-area list across your website, Google Business Profile and every portal and directory profile you hold. Third, make your site citable: suburb pages that state real facts (years active, property types, recent sales context, commission approach) plus a Q&A section answering the questions vendors actually ask.
Fourth, build the third-party record: systematic review collection after every settlement, and local media or community mentions where you can earn them.
Then make it compound: fresh market commentary each quarter, tracked AI answers each month, and measurement tied to appraisals booked, not impressions. This is the deepest playbook we run (real estate and mortgage broking are our flagship specialisation), and the full programme is laid out on our real estate industry page.
If you'd rather start with evidence than promises, check your AI visibility and see exactly what the engines say about your market today.
Yes. Portals appear for listing-browsing questions, but "who should sell my house" questions return named agencies and agents, drawn from reviews, local mentions and agency websites. Those recommendation queries are exactly where appraisals come from.
It helps (retrieval still runs on search foundations), but it isn't automatic. AI answers weigh explicit facts, entity consistency and third-party proof differently from rankings, and the benchmarks show plenty of page-one agents missing from AI answers in their own market.
Foundation fixes (profile consistency, citable suburb pages, review momentum) can move answers within weeks. The 2026 industry benchmarks report 90 days or less to a #1 AI recommendation in a market for agents who invest; treat that as a benchmark to test against, not a promise, and measure it in your own suburbs.
Directly. Broker-choice and property-management questions get the same AI treatment as agent-choice ones, and the playbook (explicit facts, consistent identity, reviews, tracked answers) transfers almost unchanged. Broking is part of our flagship real estate specialisation for exactly that reason.