It’s 11 o’clock at night. Someone in your area can’t sleep because they need to sell their house to buy the next one, and they’re worried about the timing. They don’t open the portals. They open ChatGPT.

That moment – long before anyone fills in an appraisal form – is where listings might be being won and lost. AI search has changed the game for real estate agents, and most don’t know it yet.

“Your mission, if you choose to accept it, is to be the exact local answer your clients ask AI for, so that you win the lead before a single click happens,” Samantha McLean told agents in Week 3 of the AI Sprint.

It’s a big mission. Here’s how she broke it down.

Three ways to get a lead – and the one that’s getting harder

There are really only three ways to generate a lead, Samantha said. Outbound, where you interrupt people with calls, flyers and social posts. Inbound, where they come to you. And bought leads, through the likes of Local Agent Finder, Open Agent, or Zillow if you’re overseas.

No judgement on any of them. But her preference is clear.

“Buying leads is okay, but it’s a bit like being handed a fish,” she said.

“Inbound and SEO is like building a boat and learning to fish forever. You get it right, and people start coming to you.”

The problem is that inbound itself has changed.

Roughly one in three people on Earth now touch an AI every month, by Samantha’s count, and they’re using it to research you well before they ever think about calling.

“They’re doing their own research, and they’re doing it in a different way than they ever have before,” Samantha said.

The magic mirror problem

Her first challenge to the room was one she’s set in several emails: go and ask ChatGPT who the best agent in your area is.

Just don’t do it yourself.

“It’s a bit like the queen in Snow White,” she explained.

“Magic mirror on the wall, who’s the fairest of them all? And the mirror always says, ‘You are.’ Because the LLM wants to please you.”

An AI you’re already logged into knows your location and your context, and it’s built to be your assistant – so it flatters you.

To get an honest read, swap with a trusted colleague. Get them to run the prompt on you while you run it on them.

The results can be sobering. On a call with an Insider the day before, Samantha found Google’s deep research had reported the agent’s days on market as 55. It wasn’t true.

“The LLMs will find out whatever they can about you, whether it’s true or not, because they’re looking to surface an answer,” she said.

The prompt she suggests handing to a colleague: “I’m meeting three agents this week to appraise my home. For each one, search the web for their recent sales, reviews and social media. Build a dossier on who’s genuinely active in my street and who’s just shopping for listings.” For extra detail, turn deep research on.

What your clients are actually asking

The instinct is to optimise for “best real estate agent in Broadbeach Waters”. But the average prompt people type into ChatGPT runs to about 60 words – and almost nobody searches that way.

“They’re searching for much longer, more detailed, probably more fear-based, painful questions than they ever have,” Samantha said.

The 2am questions sound more like this: “I need to sell my current house to buy a new one, but I’m worried about the timing. What’s the best way to handle the financing and logistics so I don’t get stuck paying two mortgages?” Or the inherited, outdated property the siblings want to sell as-is, in probate.

These are the long-tail, emotional queries agents can actually answer with targeted content. And being the answer is the goal.

How AI search rewrote SEO

The reason this matters now, Samantha said, is that Google has rolled out its biggest interface change in 25 years, with AI mode as the default experience.

“This whole SEO thing we used to optimise for has changed,” Samantha said.

The ten blue links aren’t gone – they can still be found off to the side – but they now sit beneath an AI overview that often answers the question outright. That’s the zero-click search: the user can get what they came for and never visit anyone’s website.

It’s not just text, either. Google has flagged an Ask YouTube feature that shifts video search from keywords to natural-language prompts and, by Samantha’s reading, will deep-link users to the exact timecode where their question gets answered.

“If you aren’t doing video market updates or talking-head market updates, you’re missing out on the ultimate zero-click use case,” she said. Google can already watch, transcribe and understand any video you put on YouTube.

The goal is no longer just website clicks. It’s building a hyperlocal authority ecosystem so that when the answer engine responds, you’re the answer it surfaces.

EEAT: why Google is hardest on agents

To be that answer, you need to understand what the engines are looking for. Samantha introduced two concepts.

The first is EEAT – experience, expertise, authoritativeness and trustworthiness.

Real estate gets some of the toughest treatment here, because Google files it under “your money or your life”.

“If someone writes a blog post on changing a zipper and it’s wrong, the world’s not going to end,” she said.

“But incorrect advice on finance or real estate or health falls into that ‘your money or your life’ category. So Google wants to make sure, before it surfaces an answer, that you have the expertise and experience to be trusted.”

Authoritativeness is why being mentioned on a high-authority site matters. Sites that have published good content for years – Elite Agent among them – get weight. If you’ve been written about or appeared on a podcast, that coverage tends to surface.

This piece connects directly to last week’s session on AI-powered DISC profiling and the earlier work on building AI buyer personas – the same theme of doing the homework before you ever sit down with a vendor.

Information gain, and the slop bucket

Trustworthiness is the part of the algorithm Samantha says has changed most. She pointed to a Google update in March 2026 aimed at AI-generated content created to game the system.

The phrase to remember is information gain.

“What the AIs want is information they don’t already know – something new,” Samantha said. “By all means use AI to generate content, but make sure you add something the AI wouldn’t know.”

What do kids actually love about that local park? What’s the dish people queue for at the cafe down the road? That human, on-the-ground detail is the signal.

Miss it, and you know exactly where you land.

“Otherwise you end up in the slop bucket,” she said, “because you’ll end up with words like ‘delve’ and ‘nestled in a cute corner of’, and it’ll just sound like AI stuff.”

Brand defence: own the ground they land on

The second concept is brand defence – an SEO strategy for keeping control of your own narrative.

When someone asks whether you’re any good at your job, you want something you control to surface. A neglected LinkedIn profile or a listing on a portal where buyers can also see five other agents is weak brand defence.

“It’s somewhere you own, no competitor ads, controlled by you,” Samantha said. Your own website is the best home for it – somewhere a prospect lands inside your ecosystem, not someone else’s.

She also made a case for transparency over polish. The flawless just-sold flyer tells an AI – and a human – almost nothing about whether you can handle a deal when it goes sideways.

“If I’m on a plane with a pilot in a storm, I want somebody flying the plane who’s been through a few storms,” she said. “I don’t want someone whose first storm is this one.”

The agents winning at this aren’t just posting the result. They’re explaining the process: what the hurdles were, how many buyers missed out, how they got it over the line.

The hardest signal to manufacture is earned media – being cited or mentioned by someone else.

Historically that meant selling a trophy home or a derelict place with a great story, the kind of thing mainstream media will chase for clicks. Without one of those, your options were to ring a journalist and hope, or pay for it.

A human-written profile on Elite Agent takes a day and a half to two days of work across several people, plus an editorial sign-off, and runs to about $1,000. As Samantha put it, that’s actually cheap compared with paid media elsewhere – Forbes charges several thousand for an article.

But the full human treatment doesn’t scale to every agent. Which is exactly the gap that started a different idea.

How Ailsa was born

“What if we could use a conversational AI as a journalist?” Samantha said. “This is where Ailsa was actually born.”

Ailsa is built to deliver information gain. She doesn’t stop at the facts of the sale – she probes the layers underneath: what your strategy was, what you found hard, the local knowledge only you would have.

In a live demonstration, Ailsa walked Samantha through a just-sold story: the price, the method, days on market, then the harder questions. An 11-day sale through a short, two-week auction campaign in an area where auctions aren’t the norm, because the vendor had already committed to their next purchase.

“Walk me through how you actually made that work,” Ailsa pressed. “What did you do differently to generate that urgency?”

“You can see how she’s leading me down a path of creating the content the AIs are looking for,” Samantha said – the process, who missed out, the length of the campaign. The lived experience that proves expertise.

Every story comes formatted with the SEO Google wants – the suburb and the agent’s name in the URL, the right headings and an agent bio at the end – and a direct-answer summary is now being added, a clean 40-to-60-word block an AI Overview can lift and cite. Crucially, it carries the authority from eliteagent.com that a do-it-yourself version can’t replicate.

The maths: a human-written Elite Agent story is around $1,000. An Ailsa story is $97 each, with five- and ten-packs available. The cheapest route by far is Elite Agent Insiders at $297 a year, which includes four Ailsa story credits, the DISC profiling and “Before We Meet” tools from earlier sessions, and a half-hour onboarding with Samantha on the annual plan.

The DIY version – with one big caveat

For anyone who wants to try it themselves, Samantha shared the prompt:

“Act as an expert real estate copywriter specialising in answer engine optimisation. Write a just-sold case study that satisfies Google’s EEAT requirements and ranks in AI overviews.” Then add your suburb and property type, the friction you overcame, your local secret, a 40-to-60-word direct-answer summary up top, and run it against your voice DNA profile.

Property managers shift the focus from a sold transaction to a problem solved or a yield maximised, and turn the raw data of a recent management win into a 400-word case study.

The caveat is real.

“The DIY route will only work if you put the sweat equity in to build your voice DNA and brand,” she said. “Otherwise you still end up in the slop bucket.”

And there’s the quieter question worth sitting with: even when done is better than perfect, does done actually get done – or are you already off to the next open home?

Your 2026 inbound checklist

Samantha closed with what to actually do next:

  • Don’t stop anything you’re already doing. Keep your Google, Rate My Agent and Facebook reviews coming, plus written and video proof. AI search has expanded the search world, not necessarily completely replaced anything.
  • Write case studies, not just sold flyers – show the process, the hurdles and how many buyers missed out.
  • Start creating video content, and add timecodes to your YouTube uploads. Ask Google for a description with timecodes and it’ll generate them for you.
  • Build a website hub that answers the long, fear-based questions your clients are really asking, so they land somewhere you own.
  • Comment as well as post – commenting is content, and Google now sees all of it.
  • Pursue earned media and exact-URL citations from trusted, high-authority sites.

The bigger point is that the people on the other side of the table have AI too. Buyers and sellers are running these prompts before they meet you, so run them yourself and be ready.

But there’s a line AI can’t cross. “People will have access to the same information you do, but the AI can’t provide the insight you can,” Samantha said – the kind of local knowledge that explains why the units on one side of a building are worth more than the other.

“Don’t automate your relationships – protect them. Don’t give away your voice to the AI, and don’t give away that last layer of customer care.”

This is Week 3 of The AI Sprint. Watch Week 1 and Week 2 on youtube now!

Did we mention you can join plenty of other AI-curious agents and Sam- live in our next sprint, which is June 2nd (that’s next Tuesday!)

It’s free, and you can register here.

Want to use AI like a professional, not a tourist? Join Elite Agent Insiders and get the tools to become so good they can’t ignore you.