PropPilot.ai
The funnel changed and nobody warned you
A buyer in London is considering buying a holiday home in Marbella. In 2022, they typed “best real estate agency Marbella” into Google and clicked the third organic result. In 2026, they ask ChatGPT: “Which real estate agencies in Marbella specialize in international buyers and have good reviews?”
ChatGPT answers in two paragraphs. It mentions four agencies. The buyer clicks on one of them.
The other three agencies — even if they ranked #1, #2, and #3 on Google — were not part of the conversation. From that buyer’s perspective, they don’t exist.
This is the silent shift happening across real estate. SEO didn’t die, but it’s no longer the only game. There’s a new one: getting cited inside the AI’s answer.
That game is called GEO — Generative Engine Optimization.
The numbers behind the shift
The data has moved fast — and unlike most “AI is changing everything” posts, the figures here are public and traceable:
- ChatGPT reached 800 million weekly active users in October 2025 (TechCrunch, Sam Altman) and crossed 900 million by early 2026. Altman has stated that roughly 10% of the world now uses the platform.
- Google’s AI Overviews appeared in 60.3% of US queries by late 2025 (Advanced Web Ranking), and on 39.4% of informational queries specifically.
- When AI Overviews appear, organic CTR drops by 61% on those queries — from 1.76% to 0.61% — according to a Seer Interactive study covering 3,119 queries and 25.1M impressions across 42 organizations.
- But the GEO opportunity sits in the same study: brands cited inside AI Overviews earn 35% more organic clicks and 91% more paid clicks than uncited competitors. Citation isn’t just a vanity metric — it redistributes traffic.
- Perplexity went from 230M to 780M monthly queries in under a year (CEO Aravind Srinivas, May 2025).
- On the buyer side, SEMrush research shows 85% of US consumers use AI tools weekly for shopping research as of December 2025. ChatGPT referrals to US websites grew 367% in 2025 (Euromonitor).
- Critical caveat: only 2% of consumers would buy from an unfamiliar brand based solely on an AI recommendation (Idea Grove survey, April 2026). The other 98% verify first.
For real estate, the implication is direct: a meaningful share of the buyers you most want to reach are filtering options through AI before they touch a portal or Google. Being cited is the entry ticket. Being cited and having a digital footprint that survives verification is what closes deals.
SEO vs GEO — they aren’t the same problem
SEO optimizes for Google’s ranking algorithm. The goal is a click. The metric is rank.
GEO optimizes for an AI model’s citation. The goal is a mention. The metric is citation rate.
| Dimension | SEO | GEO |
|---|---|---|
| Optimization target | Search ranking algorithm | LLM training and retrieval |
| Goal | Click to your site | Citation in the AI’s answer |
| Primary metric | Organic position, CTR | Citation rate, share of voice |
| Crawler | Googlebot, Bingbot | GPTBot, ClaudeBot, PerplexityBot, Google-Extended |
| Format that wins | Long-form articles, lists | Direct answers, structured data, verified facts |
| Brand signals | Backlinks, domain authority | Brand mentions across the web, schema, E-E-A-T |
The signals overlap in places — E-E-A-T (Experience, Expertise, Authoritativeness, Trust), schema markup, content quality — but the optimization is genuinely different. A page that ranks #1 in Google isn’t automatically cited by ChatGPT, and vice versa.
The 7 GEO levers for real estate
1. Make sure AI crawlers can actually reach you
Step zero. Many real estate sites unintentionally block GPTBot, ClaudeBot, and PerplexityBot through inherited robots.txt rules or aggressive WordPress plugins.
Open yoursite.com/robots.txt and verify these user-agents are not blocked:
GPTBot(OpenAI)ClaudeBot(Anthropic)PerplexityBot(Perplexity)Google-Extended(Google AI training)CCBot(Common Crawl, used by many models)
If any are disallowed, your content can’t be ingested. You’re invisible to those models by definition.
2. Publish an llms.txt file
llms.txt is the emerging convention for guiding AI crawlers — a more focused signal than robots.txt. Place it at the root of your domain:
# yoursite.com/llms.txt
# About
PropPilot Realty is a real estate agency specializing in
international buyers in Marbella, Spain.
# Key pages
- /about: company information, founders, team
- /services: services offered
- /properties: current listings
- /blog: market analysis and guides
Adoption isn’t universal yet — OpenAI’s stance is in flux — but Anthropic and Perplexity already respect it, and the cost is negligible.
3. Write content that’s directly citable
LLMs prefer content that gives a clear, attributable answer to a specific question. Three rules:
- Lead with the answer. Open the section with a direct, factual statement. Then expand. Models are more likely to cite the first sentence of a section.
- Use named entities. “Marbella” beats “the area.” “AED 35M” beats “luxury price range.” Specific numbers and names are quotable.
- Avoid floral introductions. “In today’s fast-paced world…” is filler. Models skip it. Buyers skip it. Skip it yourself.
4. Schema markup, properly applied
Real estate-relevant schemas include RealEstateAgent, Residence, Apartment, LocalBusiness, Review, FAQPage, and Article. Every property listing and every blog post should carry the appropriate JSON-LD.
This is foundational SEO too, but it’s becoming more important for GEO. AI engines use structured data to confirm facts and resolve ambiguity. A listing without schema is a listing the model has to guess about.
5. Brand mentions across the web (E-E-A-T at scale)
The strongest GEO signal isn’t your own site — it’s how often and how positively your brand is mentioned elsewhere. Models trained on the open web learn that “PropPilot Realty” is a credible Marbella agency by seeing the name appear in trustworthy contexts: industry publications, Reddit threads, Quora answers, podcasts, news mentions, partner directories.
Practical implication: PR, partnerships, podcast appearances, and well-curated profiles on relevant directories are now GEO investments, not just brand-building.
6. Expose data via MCP for direct AI integration
This is where the real differentiation lives. Most agencies are stuck competing on text. The leaders are exposing their inventory, availability, and FAQs via the Model Context Protocol (MCP) — the open standard that lets AI assistants query your data directly.
When a buyer asks Claude “show me 3-bed apartments under €600K in Málaga,” and the model can call your MCP server in real time, your agency becomes the source of truth — not a paragraph on a blog.
We covered the architecture in Connected AI Agents — The Future of Real Estate. The relevant point for GEO: a published MCP server moves you from “content the AI might cite” to “data the AI must use.”
7. Track your AI Citation Rate
You can’t optimize what you don’t measure. The metric to track is AI Citation Rate (ACR): the percentage of times your brand appears in answers to a defined set of queries across the major AI engines.
A simple monthly process:
- Define 20-30 queries a real buyer in your market would actually ask. Mix branded (“best real estate agency in Marbella”) and unbranded (“how do I buy property in Spain as a foreigner”).
- Run them across ChatGPT, Claude, Perplexity, and Google AI Overviews.
- Record whether your brand is mentioned and in what context.
- Track the trend over time.
This is the GEO equivalent of weekly rank tracking. It’s manual today, but it works.
Quick audit — are you currently being cited?
Pause for five minutes. Open ChatGPT, Claude, and Perplexity. Try these queries:
- “Best real estate agency in [your city]”
- “How do I find a trustworthy property agent in [your country]”
- “What’s the most reliable proptech tool for real estate brokerages”
- “[Your agency name] reviews”
If your brand doesn’t appear at all in the first three, you have a GEO gap. If it doesn’t appear cleanly in the fourth, you have a brand-data gap.
This isn’t theoretical. Run the test now. The result usually is sobering — and it’s the most direct argument for taking GEO seriously.
Frequently asked questions
What is GEO (Generative Engine Optimization)?
GEO is the practice of optimizing your content and digital presence so that AI search engines — ChatGPT, Claude, Perplexity, Google AI Overviews — cite your brand when answering user queries. It’s the natural successor to SEO for a market where buyers increasingly start inside an AI assistant.
How is GEO different from SEO?
SEO optimizes for Google’s algorithm and aims for clicks. GEO optimizes for AI model citation and aims for mentions in the answer the AI gives — even if the user never clicks through. The signals overlap (E-E-A-T, schema, quality content), but the format, channels, and measurement are different.
Are AI crawlers allowed on my site by default?
By default, yes, unless your robots.txt explicitly blocks them. Many sites accidentally block GPTBot, ClaudeBot, and PerplexityBot through inherited rules from old SEO plugins or templates. The first GEO audit step is verifying these crawlers are allowed.
What is llms.txt and do I need one?
llms.txt is an emerging convention for telling AI crawlers what content is most useful and how to interpret it. Adoption is uneven — Anthropic and Perplexity respect it, OpenAI’s policy is in flux — but the implementation cost is low and the upside is meaningful. For real estate, worth doing.
Should I worry about the AI Overviews killing my organic traffic?
If your traffic depends on informational queries that AI Overviews can answer in the SERP, yes — you’ll see clicks decline regardless of what you do. The defense is two-fold: GEO (be cited inside the Overview) and direct channels (email lists, brand searches, MCP-driven discovery) that don’t depend on Google’s funnel.
Conclusion
The buyers you want most — qualified, international, time-pressed — are spending more of their decision time inside AI assistants than ever before. If your brand doesn’t show up in those conversations, your funnel is leaking buyers you’ll never even know existed.
GEO isn’t a replacement for SEO. It’s the next layer. The work is concrete: unblock the crawlers, add schema, publish llms.txt, expose your data through MCP, build brand mentions, track your citation rate.
Citation in an AI Overview correlates with 35% more organic clicks and 91% more paid clicks (Seer Interactive). The cited slots are still uncrowded in real estate, especially outside English-speaking markets. The window for getting in early is open — but it isn’t going to stay that way.