type Ranking = {
  keyword: string
  position: number
  engine: "google" | "chatgpt"
}|

// deploy to production
await vercel.deploy({
  project: "6aiatx",
  env: "production"
})

const agent = new AIAgent({
  tools: [webSearch, codeExec],
  memory: vectorStore
})

const leads = await db.query(
  "SELECT * FROM inbound WHERE new"
)

// automate the repetitive part
for (const msg of inbox) {
  await agent.reply(msg)
}

export const optimize = () =>
  crawl().then(fix).then(rank)

model.predict(input).then((res) => {
  return deploy(res)
})

const site = await audit.run({
  url: "client.com",
  checks: ["speed", "schema", "meta"]
})

← Back to Blog

When an Owner Asks ChatGPT for a Property Manager

GEO is the other half of search now. Here is what it is, and what we actually do about it.

Roman Lubetzky

Co-Founder & CEO, 6AI · August 2026

model.predict(input).then((res) => {
  return deploy(res)
})|

const site = await audit.run({
  url: "client.com",
  checks: ["speed", "schema", "meta"]
})

async function trackRankings(domain: string) {
  const terms = await keywords.get(domain)
  return agent.optimize(terms)
}

import { claude } from "@anthropic/sdk"
const response = await claude.complete({
  model: "claude-sonnet-4-6",
  prompt: buildPrompt(context)
})

type Ranking = {
  keyword: string
  position: number
  engine: "google" | "chatgpt"
}

// deploy to production
await vercel.deploy({
  project: "6aiatx",
  env: "production"
})

const agent = new AIAgent({
  tools: [webSearch, codeExec],
  memory: vectorStore
})

const leads = await db.query(
  "SELECT * FROM inbound WHERE new"
)

An owner with a rental property to hand off used to open Google, search “property management” and their city, and work down the results. A growing share of them now open ChatGPT or Perplexity and ask who they should use.

Those are completely different games. In the first one you are competing for a position on a page. In the second there is no page — there is one answer, naming two or three companies, and you are either in it or you do not exist. Nothing in your rankings report will tell you which.

What GEO actually is

GEO stands for generative engine optimization. SEO optimizes for ranking in a list of results. GEO optimizes for being the source an AI quotes when it answers a question directly.

Some people call it answer engine optimization, or AEO. The label matters less than the shift: for the first time in twenty years, the thing being optimized is not a position but a citation.

It is not a replacement for SEO. The two overlap heavily — clean structure, clear content, and technical health help both. But there are things that only matter for GEO, and if you are only doing traditional SEO you are not doing them.

Why property management is especially exposed

Three reasons this hits our industry harder than most.

It is local, so the answer set is small. “Who manages rental property in Austin” resolves to a handful of names, not a page of ten blue links with more below. Being fourth is fine in search results. Being fourth in an AI answer often means being left out of it.

It is referral-driven, which means owners are used to asking for a recommendation. Asking an assistant is a very natural substitute for asking a friend, and it feels more like advice than like searching.

And it is high-consideration. Nobody hands over a property on impulse. Owners research, and research is exactly the behavior these tools have absorbed.

What we actually do about it

The work is less mysterious than it sounds. An AI system has to be able to read your site, understand what you are, trust it, and lift a specific claim out of it. Most property management sites fail at least two of those.

So: structured data, so a machine can tell what each page is rather than guessing from the layout. Clear entity signals — who you are, which markets you serve, what you actually do — stated consistently in the same words everywhere, because inconsistent descriptions make an entity harder to resolve. Content organized so a single sentence can be quoted without the surrounding paragraph. A machine-readable summary at llms.txt for the crawlers that look for one. And crawler access for the AI bots that respect robots rules, which is a one-line decision most sites have never consciously made.

None of that is a trick. It is the same instinct as good SEO — be legible — pointed at a reader that is a model instead of a person.

How to tell if it is working

This is the part I like. You do not have to wait on a rank tracker. Open an AI tool with browsing and ask it the question an owner would ask: who manages rental property in your city, or what a specific company does. Read the answer.

If it describes you wrongly, or confuses you with a similarly named company, or does not mention you at all, that is your baseline and it took ninety seconds to get. Do the work, wait for a crawl, ask again.

One honest caveat: prompted answers update faster than unprompted ones. An engine that will describe you correctly when handed your URL may still not name you unprompted for weeks, because that depends on a broader index refresh. Do not read early silence as failure.

Where this is going

I do not think AI answers replace search. I think they become the first step for a meaningful slice of owners, and that slice grows. The companies that get legible to these systems early will be the defaults in their markets by the time everyone else notices there was something to optimize.

That is why we treat SEO and GEO as one job rather than two services. Owners use both. Here is how we do it, or tell us where you want to show up.