AI Search Is Changing What It Means to Be an Authority in Real Estate

Yesterday was the first day of the LuxuryRealEstate.com conference. Between sessions and at the reception that evening, I kept hearing a question successful luxury agents ask me more and more often:

Why do agents who sell far less real estate sometimes show up more prominently in ChatGPT, Google's AI answers and Perplexity?

It's a fair question.

This year I'm here on a panel with Rob Thomson and Jack Thomson about Waterfront Properties being named LuxuryRealEstate.com's Affiliate of the Year. But I've presented on AI at these conferences many times, and some version of this question comes up almost every time.

The answers agents get don't help much. Speakers with similar credentials give conflicting advice, and we haven't even agreed on what to call it: SEO, AEO, GEO, AI search optimization or search everywhere optimization.

So rather than add one more opinion, I want to start with a more basic question:

When someone asks AI a question about real estate, where does it get the information to answer?

New Labels for a Familiar Challenge

The terminology matters less than the debate suggests.

Google's own guidance for generative AI search continues to emphasize established SEO practices. Google also says on its AI features page that there are no additional technical requirements to appear in AI Overviews or AI Mode, and no special AI files or markup are required simply to participate in those experiences.

That doesn't mean nothing has changed.

A great deal has.

Instead of scanning ten blue links and choosing which sites to visit, consumers can increasingly ask a complicated question and receive a synthesized answer.

That makes the source of the answer increasingly important.

And I believe it creates a real opportunity for people who actually know their markets.

You Know Your Market. Does AI Know You Do?

This is where the frustration makes sense.

A top producer may have decades of experience, hundreds of transactions and deep knowledge of a market.

Much of that expertise may never have been published.

Referral businesses are especially exposed. According to the National Association of REALTORS®' 2025 Profile of Home Buyers and Sellers, as summarized by Virginia REALTORS®, 66% of sellers found their agent through a referral or had worked with them before, while 43% of buyers used a referred agent.

That's an outstanding way to build a business.

But phone calls, private conversations, listing presentations and dinners don't necessarily create a public record AI can find.

Real estate has a second problem.

Some of the most authoritative evidence of an agent's production sits inside MLS and brokerage systems and may not be readily accessible or independently verifiable by the AI system answering a consumer's question.

I believe production matters.

If the question is who has sold the most in a market or price range, verified transaction history should carry enormous weight.

The agent with the most sales isn't automatically the right choice for every client, but real transaction experience is an important measure of expertise.

The problem is that AI may not be able to weigh evidence it can't reliably access or verify.

And there's a bigger point.

Most Questions Aren't "Who Sells the Most?"

Think about what buyers and sellers actually want to know:

• What should I know before buying in this market?
• What makes one development different from another?
• How should I think about pricing my home?
• Why can two similar-looking homes sell for very different prices?
• How do flood zones and insurance affect waterfront property?
• How do different types of country club memberships work?
• Are buyers negotiating?
• Are they paying cash or financing?
• What should I renovate before selling?
• What mistakes should I avoid?

People ask AI questions like these every day.

Selling the most homes doesn't put those answers on the public web.

But experience often means that agent has excellent answers.

The opportunity is to make more of those answers available.

AI Has Created an Unlimited Supply of Average Content

This is where I think many people will take the wrong lesson from AI search.

Anyone can now produce a well-written article on almost anything in seconds.

Ask AI for "Seven Things to Know Before Buying a Luxury Waterfront Home," and you'll probably get a perfectly competent article.

So will thousands of other agents.

More of that isn't much of a competitive advantage.

Writer Nicolas Cole, in a recent conversation with Greg Isenberg, offers a useful way to think about it. He separates content into three kinds, and the framework fits real estate remarkably well.

Google obviously isn't endorsing Cole's framework, but there's an interesting similarity in direction.

Google's current guidance for generative AI search emphasizes useful, original, non-commodity content and first-hand experience rather than simply producing more generic material.

That distinction matters.

Think of Your Expertise as a Data Set

Cole calls our experiences, lessons and points of view an "unmade data set."

It's a useful idea for real estate, because a successful luxury agent may carry a remarkable amount of knowledge that exists almost nowhere publicly.

It lives in:

• Thousands of conversations with clients
• Years of negotiations and listing presentations
• Knowing why one property sold quickly while another sat
• Understanding distinctions between streets, buildings, clubs and waterfront locations that no database adequately captures
• Recognizing when the numbers don't tell the whole story
• Lessons learned from transactions that didn't go as expected
• Questions buyers and sellers have asked repeatedly over many years

AI doesn't automatically know any of this.

If you've never articulated it, much of that knowledge remains an unmade data set.

So the opportunity isn't simply to create more content.

It's to make what you genuinely know available, understandable, verifiable and attributable.

What We Know, and What We're Still Learning

AI search is moving fast, and an entire industry already claims to know exactly how to optimize for it.

I think it's important to separate what is documented from what research suggests and what remains largely speculation.

The distinction matters.

The GEO research doesn't show that adding a citation to your next article makes ChatGPT 40% more likely to recommend you. It found improvements in measures of source visibility under particular experimental conditions.

Similarly, one study finding a strong preference for earned media shouldn't be turned into a universal rule that AI always trusts someone else's website more than yours.

Research gives us evidence.

It doesn't give us the complete algorithm.

We also don't know how these systems, each built independently by its own AI provider, will change six months from now.

I'd be skeptical of anyone, including me, who claims to have found a permanent formula.

That's why I focus on what makes sense regardless of what we eventually call this.

Your Website Matters, but It Shouldn't Stand Alone

Your website documents what you say about yourself.

Independent sources can provide something different:

corroboration.

That can include:

• Client reviews
• Independent media coverage
• Professional organizations and board service
• Speaking appearances, interviews and podcasts
• Brokerage and MLS profiles
• Awards and professional recognition
• Other authoritative sites referencing your research or analysis

For established top producers, this can be an enormous asset.

Many already have years of reputation, relationships, transactions and recognition.

The goal isn't to manufacture authority for AI.

It's to make existing authority more visible and verifiable.

I increasingly think of this as a combination of three things:

First-party expertise + authoritative evidence + independent corroboration.

What I'm Doing Now

This question is personal for me.

After 18 years managing one of the country's most innovative and recognized luxury brokerages, I've moved into luxury real estate sales as an agent with Waterfront Properties, while continuing my work as an AI evangelist and educator.

So I'm facing the same question:

Knowing what I know about AI, search, data, marketing and real estate, how should I build my business for the next decade?

I'm approaching it with a particular set of strengths.

• Market intelligence. I've spent years analyzing MLS statistics and studying what the data can tell us about markets, communities, buyers, sellers and individual properties.
• Marketing. Waterfront Properties has always been a leader in marketing innovation for real estate, and that work has been recognized repeatedly in our industry. Years inside that environment shaped how I think about creating information people want, presenting it effectively and building visibility over time.
• Technology. Technology has been part of my background for years, and that experience helped me adopt and use AI earlier and more deeply than I otherwise might have.

Not every agent has that combination, and that's the point.

Every experienced agent has their own data set.

One understands new construction extraordinarily well.

Another has spent 25 years selling waterfront property.

Someone else knows condominiums, country clubs, historic homes, relocation, negotiation or a particular development better than almost anyone.

The question is:

How much of that knowledge exists somewhere AI can find it?

For my own business, I'm starting with the questions:

1. What are buyers and sellers actually asking?
2. Which of those questions can I answer especially well because of my experience?
3. Where can authoritative data improve the answer, and where can public information add context?
4. Where does my experience tell me something the numbers alone don't?
5. Where can marketing and presentation make complicated information easier to understand?
6. Where can technology help me analyze, organize and distribute that knowledge?
7. What do I know, or what can I analyze, that isn't already readily available online?

Sometimes the answer rests mainly on data, sometimes on experience.

Often the strongest answer combines both.

AI can then help me research, analyze, organize, reformat and distribute that work more efficiently.

I don't want AI manufacturing expertise I don't have.

I want it helping me make better use of the expertise I do have.

Rather than just telling other agents what to do, I'm building my own business around this idea.

I've Seen This Before

About 15 years ago, I was deep into SEO, and our listings, websites and community pages consistently ranked near the top of Google's organic results.

During successful listing presentations, sellers who owned businesses, such as doctors, dentists, plumbers and other professionals, would see those results.

The conversation often turned to their own companies because they wanted the same visibility.

The technology is very different today, but I suspect one of the most important lessons still holds:

Useful bodies of work compound.

Strong websites built visibility by consistently answering the questions people searched for and by earning links and recognition from other authoritative sites.

That didn't happen in 30 days, and I don't believe the underlying principle disappears because the interface has changed from a search box to a conversation.

Don't Chase Every Acronym

There's plenty to learn about how AI systems discover, retrieve and cite information.

Technical accessibility, site structure, structured data, reviews and third-party references all matter, and we should keep testing and learning as the technology develops.

But I wouldn't build a long-term strategy around today's acronym or someone's latest optimization trick.

I'd build something much harder to replace:

a public record of expertise.

• Answer real questions.
• Use authoritative evidence whenever possible, and show your sources.
• Contribute information, analysis or perspective that adds something useful to what already exists.
• Explain what you've learned from experience.
• Develop clear points of view.
• Earn legitimate third-party recognition.
• Keep your information current.
• Keep publishing, not for three weeks after someone's GEO presentation, but for years.

If you've spent 10, 20 or 30 years becoming genuinely knowledgeable about a market, you already possess the hardest part:

the expertise.

The emerging challenge is making that expertise accessible to the systems increasingly answering your prospective clients' questions.

Don't create content just so AI will find you.

Create information worth finding.