Search stopped sending people to your site

For twenty years, search meant a list of links and the visitor picked one. That is ending. Google now writes an answer at the top of the page, and most people read it instead of clicking through.

13.7%
of Google searches now return an AI Overview

Source: Xu et al., WUSTL, 2026

64.7%
of question-form searches trigger one, against 9.5% of keyword searches

Source: Xu et al., WUSTL, 2026

29.8%
of the sources cited never appear on the first page of results

Source: Xu et al., WUSTL, 2026

For property search this matters more, not less. "Is the East Cape a good place to buy," "what are closing costs in Los Cabos," "is Cabo safe to live in" are full questions, and questions are exactly what the answer engine takes over. Google reports that more than two billion people now encounter these answers every month.

What does that cost a real estate site?

When the answer sits on the results page, the click that used to reach you often never happens. Your best page can hold the top organic spot and still lose the visit.

This is not only a publisher problem. A separate study measured a drop of roughly 15% in traffic to Wikipedia once AI Overviews appeared on related searches (Khosravi and Yoganarasimhan, 2026), and Pew Research found in 2025 that people are markedly less likely to click any link when an AI summary is present. The pattern is the same for a brokerage site as it is for an encyclopedia. The answer satisfies the question, so the visit does not occur.

Why does AI cite one source and not another?

Getting cited is not the same as ranking. The two run on different machinery. In the largest study to date, almost a third of the domains cited inside AI Overviews did not appear anywhere on the first page of organic results for the same query, and each answer drew on a median of eight sources pulled from a pool the ranking algorithm does not fully surface.

Traditional SEO

  • Earn a ranked link on the page
  • The visitor chooses what to click
  • Ten organic slots to compete for

AEO

  • Earn a citation inside the written answer
  • The model chooses what to surface
  • A median of eight cited sources, often pulled from off-page

Source: Xu et al., WUSTL, 2026

There is a second pattern worth your attention, and it changes how you spend effort.

For a brokerage that means a listing page or a company blog post, on its own, is the weakest kind of source in the eyes of an answer engine. Being referenced by a market report, a local publication, or an industry site counts for more than another post on your own domain.

How do you get cited?

The work is concrete, and most of it is structural rather than clever.

Answer the question in the first two sentences. Lead each section with the direct answer, then expand. The engine lifts the clear, early statement, not the paragraph that warms up for four lines first.

Use the buyer's real questions as your headings. Question-form queries are the ones that trigger an answer, so a page built around them gives the model the shape it is looking for.

Give it something to extract. Definitions, clear comparisons, ordered steps, and numbers with their sources are the material answers are built from. Vague prose is not.

Earn references off your own site. Because earned media outweighs brand-owned content, a mention in a market report or a local outlet does more for your AI visibility than another post you publish yourself.

Be accurate, and be sourced. The same large study found that 11% of the claims inside AI Overviews were not supported by the pages they cited. The sources that get trusted over time are the ones whose facts hold up, with the reference visible. Sloppy numbers are a liability twice over.

Add structured data. Schema markup tells the machine what each part of your page is, which makes you easier to parse and to cite.

How do you know if it is working?

You cannot see AI citations in Google Analytics, which is why most businesses have no idea whether they are being quoted or ignored. Blueprint tracks client visibility across ChatGPT, Gemini, Claude, and Perplexity with Visby, and runs AIQ as a first diagnostic to show where you stand before any work starts. The metric that matters is your share of the answer across these assistants, not your position on a page of links almost no one clicks through anymore.

The work that wins AI is the work that wins buyers

None of this is a trick. Clear answers, honest numbers, a credible reputation off your own site, and a page structured so it can be read at a glance. That is what an answer engine rewards, and it is the same thing that earns a buyer's trust. Build for one and you build for both.