Luxury Real Estate Marketing in 2026: How Buyers in Miami and Mexico Find Agents Through AI

Who Is Buying Luxury Real Estate Between Miami and Mexico?

Foreign buyers purchased $56 billion of existing U.S. homes between April 2024 and March 2025, up 33.2% year over year, according to the National Association of Realtors’ 2025 International Transactions in U.S. Residential Real Estate report, released 14 July 2025. They bought 78,100 properties, a 44% jump, at a record median price of $494,400. Florida took 21% of that volume — still the top state — and Mexico returned to NAR’s top five countries of origin, ranking third at $4.4 billion (8% of foreign-buyer dollar volume) across roughly 6,200 homes.

Miami carries an outsized share of that Florida total. MIAMI REALTORS’ own international report, released 27 January 2026, put South Florida’s foreign-buyer share at 15% of residential dollar volume — seven times the national figure (2%) and more than three times the statewide figure (5%) — and said international purchasers accounted for 49% of new South Florida construction, pre-construction and condo-conversion sales over the 18 months ending July 2025. Florida has now held the #1 spot for 17 straight years. In that same MIAMI REALTORS’ data, Colombia led foreign buyers of South Florida real estate at 15% of foreign-buyer share, Argentina followed at 12%, and Mexico ranked third at 7%.

Put the two reports together and the shape is clear: Mexico is a top-five source of foreign capital into U.S. housing nationally, and a top-three source of foreign buyers into the market — Miami — that absorbs a fifth of all foreign purchases in the country. A buyer moving between Ciudad de México, the Riviera Maya and Miami is not a niche case; it is a meaningful, measured slice of two active markets.

How Do These Buyers Actually Search for an Agent?

Neither report measures search behavior, so we will not pretend it does. What is true, and does not need a survey to state, is that a buyer moving between two countries and two languages is used to typing (or speaking) a direct question and expecting a direct answer — in Google’s AI Overviews, in ChatGPT, in Perplexity — rather than reading ten blue links and cross-referencing a second site in the other language. That habit is now the default across most consumer search, luxury real estate included.

Our own scan of 63 Miami real-estate websites, run 25 September 2026, found every one of them technically reachable by ChatGPT’s search crawler and Googlebot — so the crawlers are getting in. The gap, on that scan and on the one below, is not access. It is whether the page is built so an AI engine can actually lift a clean, attributed answer out of it once it arrives.

What Does an AI Answer Engine Need to See on the Page?

Google’s own structured-data documentation is narrower than most agencies claim. For a page to be eligible for a LocalBusiness rich result, Google requires exactly two properties: name and address. RealEstateAgent is a real schema.org type under LocalBusiness and Organization, and Google’s general guidance is to use the most specific business sub-type available rather than a generic block. Google does not require structured data for its generative AI features specifically — but it also cannot cite what it cannot parse, and a page with no business-identity markup leaves an AI engine to guess who is behind it from prose alone.

Three structural things follow from that: one clean Organization or RealEstateAgent block naming who you are, where you work and how to reach you; headings written as the questions buyers actually ask (not labels); and a single, named, licensed person on the byline. That last point is not a schema requirement — it is an E-E-A-T signal AI engines and Google both weigh, and it is the one thing a franchise page or a portal listing usually cannot put a real, verifiable name behind.

Are Mexican Real Estate Sites Built This Way Yet?

We wanted to know whether the structural gaps we found in Miami show up on the other side of this same buyer flow. On 25 September 2026 we ran a small, informal check — the same kind of pass-fail checks behind our free AI Visibility Checker — against 39 Mexican real-estate and hospitality sites; 33 returned a real homepage (three answered with an empty 170-byte shell and are excluded). We are reporting totals only, not naming any site.

Of those 33: 13 (39%) had no business-identity structured data at all — no Organization, no RealEstateAgent, nothing naming who runs the site. 12 (36%) shipped no JSON-LD of any kind. 28 (85%) had fewer than two headings written as a real question. Smaller gaps showed up too: 16 (48%) had no visible email link on the homepage, and 11 (33%) were missing an Open Graph title and image for social previews. Across the sample, sites passed an average of roughly 77% of the checks we ran.

That is a small, self-selected sample checked on one day, not a market survey — treat it as directional. But it lines up with what we found scanning Miami: the crawlers are rarely the problem, and the missing business identity and the absence of question-shaped structure are.

What Does a Bilingual Listing Page Actually Need?

A translate widget bolted onto an English page is not a bilingual page. It gives a Spanish-reading buyer machine-translated prose sitting on an English URL, with no schema, no headings and no FAQ actually written for that audience — and it gives an AI engine two mismatched signals instead of one clean one. A real bilingual pair is two URLs, each with its own meta title and description, its own question-shaped headings answering the questions a buyer in that language actually asks, and matching (not identical, but equivalent) structured data on both sides.

Our own Spanish marketing-inmobiliario page is built that way on purpose: not a translation of the English page, a page built for the Spanish-language query on its own terms, with its own headings and its own FAQ. That is the model for a listing or agent page serving buyers who move between Ciudad de México or the Riviera Maya and Miami — one identity, two languages, both structured the same way.

What Are the Five Steps to Make a Listing AI-Ready?

In the order we would tackle them on an existing site:

None of these require a rebuild. Most Miami and Mexico agent sites we have looked at already have a homepage, a contact form and a working CMS — the fixes above are structural additions to what is already live.

What Should a Cross-Border Agent Fix First?

Start with identity, then structure, then language. A missing business-identity block is the single most common gap on both sides of this market — 38% of the Miami sites we scanned and 39% of the Mexican sites lacked one. Add that first. Then rewrite two or three of your most-visited page’s headings as the actual questions buyers type. Only after that is it worth building the Spanish twin, because a well-structured English page is the template the Spanish page should match, not diverge from.

For the fuller Miami picture and the method behind these checks, see our Miami real estate AI-readiness study. For how we build cross-border and bilingual marketing programs day to day, see Miami real estate marketing and real estate marketing. You can run the same checks on your own site in about a minute with the free AI Visibility Checker.