Skip to content
pseo-geo-aeoaeogeochatgptperplexitycitationspillar

How to Get Your Business Cited by ChatGPT and Perplexity

How to get your small business cited by ChatGPT, Perplexity, Claude, and Gemini: answer-shaped content, schema, llms.txt, and consistent facts.

RJ Murray, Founder · June 10, 2026 · 5 min read

We have worked in search through a lot of algorithm shifts, and we have never seen one move as fast as this. Two years ago, getting found meant ranking in a list of ten blue links. Today, a growing share of a business's customers ask ChatGPT or Perplexity a question and get an answer with two or three sources cited inline, and never see a results page at all. If your business is not one of those cited sources, you are invisible to that customer, no matter how well you rank the old way.

The good news: getting cited is more learnable and less contested than classic SEO. Here is exactly how.

tl;dr

To get cited by ChatGPT, Perplexity, Claude, and Gemini: put the direct answer to each page's core question in the first 80 words, add FAQ schema on visible questions and answers, publish an llms.txt file at your domain root, and keep your core facts identical on every page. Answer engines cite pages that state things plainly, consistently, and with structured data behind them. This is answer engine optimization, and it is winnable for a small business faster than classic search rankings are. This post is the field manual.

Why being cited is now its own job

Classic SEO and answer engine optimization are related but not the same, and conflating them is the first mistake. We laid out the full distinction in what is answer engine optimization, but the short version is this: SEO tries to rank your page in a list, AEO tries to make your page the answer the assistant gives and names.

The surfaces are different. A search engine returns links and lets the human choose. An assistant returns a synthesized answer and chooses the sources itself, usually two to four of them, and attributes them. Winning a citation is winning the only slot that gets seen. And because the assistant is choosing, the signals it weighs are slightly different from the ones that move a ranking: clarity, structure, and consistency matter even more than raw authority.

The four things that get you cited

After running this for dozens of small-business sites, the work reduces to four moves. None of them require buying links or pitching anyone.

1. Answer the question in the first 80 words

An assistant reads a page looking for a clean, liftable answer. If your services page opens with "Welcome to our family-owned business, established in 1995, where quality is our passion," there is nothing to quote. If it opens with "We are a licensed plumbing company in Kitchener offering 24-hour emergency service, drain cleaning, and water heater installation across Waterloo Region," the assistant has a sentence it can cite verbatim.

Write the direct, factual answer to the page's core question in plain language, up top, before any throat-clearing. This single change moves more citation outcomes than anything else.

2. Add FAQ schema on visible questions

FAQPage structured data turns a question and its answer into a clean, machine-readable unit an assistant can quote with confidence. The rule, which is also a Google policy rule, is that the questions and answers must actually appear on the page. Marking up content that is not visible is a manual-action risk and we never do it. We go deeper on this in the schema markup guide for local business.

In our crawl of 391 small-business sites for the State of Small-Business Websites report, 36.8 percent had no structured data at all. That is more than a third of businesses handing the citation to a competitor for free.

3. Publish an llms.txt file

llms.txt is a plain-text file at the root of your domain that gives an AI crawler a curated, human-written summary of your site and the URLs that matter most. It is the answer engine equivalent of sitemap.xml, except it optimizes for understanding rather than completeness. It takes about fifteen minutes to write and it improves the odds that an assistant cites you correctly instead of guessing from a partial crawl. The generative engine optimization guide covers where this fits in the broader picture.

4. Keep your facts identical everywhere

This is the one almost everyone gets wrong, and it is the most damaging. If your homepage says one thing about what you do or what it costs, and your services page says something subtly different, the assistant cannot quote you safely, so it quotes someone clearer instead. Consistency is a ranking factor for citations. Pick your core facts, what you do, what it costs, where you serve, and state them the same way on every page. Concrete and consistent beats polished and varied, every time.

The part that proves this works: this site

The reason we can write this with confidence is that the site you are reading runs the playbook on itself. Every key page on atlasforge.one opens with an answer-shaped lead. FAQ schema sits on the visible questions. There is an llms.txt at the root. And there is a single source for the facts, one file, so the price and the build time and the service description are identical on every page and never drift.

That is the AtlasForge model in one sentence: every site we build ships this way, because we test it on ourselves first. If you want to see what that looks like for your business, the guides hub walks through each capability, and you can enter your site to watch a finished, AEO-ready version get built.

Where to start this week

Pick your three most important pages, the homepage and your two top services or products. Rewrite each opening so the first two sentences answer the page's core question plainly. Add FAQ schema to the questions you already answer on those pages. Write a fifteen-minute llms.txt. Then audit your core facts and make them identical across all three. That is the minimum viable citation setup, and for most small businesses it is the difference between being the source an assistant names and being the business it never mentions.

Frequently asked

How do I get my business cited by ChatGPT?
Structure your pages so an assistant can lift a clean, true answer: put the direct factual answer in the first 80 words of each key page, add FAQ schema on visible questions and answers, publish an llms.txt file at your domain root, and keep your core facts (what you do, what it costs, where you serve) identical everywhere. ChatGPT cites pages that state things plainly and consistently, not pages full of marketing language.
What is the difference between SEO and getting cited by AI?
SEO aims to rank a page in a list of search results. Getting cited by an AI assistant, called answer engine optimization, aims to make your page the source the assistant quotes when it answers a question directly. The techniques overlap heavily, both reward clear content and structured data, but AEO adds answer-shaped writing and an llms.txt manifest aimed at the assistant rather than the search results page.
Do I need an llms.txt file to be cited by AI?
It is not strictly required, but it helps and it is cheap to ship. llms.txt is a plain-text file at your domain root that gives an AI crawler a curated summary of your site and the URLs that matter most. It takes about fifteen minutes to write and improves your odds of being cited correctly rather than the assistant guessing from a partial crawl.
Why does an AI assistant cite one business and not another?
An assistant cites the source that gives it a clean, quotable, trustworthy answer with the least ambiguity. A page that states exactly what a business does, what it costs, and where it operates, backed by consistent schema, is far more citable than a page of vague claims. Inconsistent facts across pages are the most common reason a business is passed over: if one page says one price and another says something else, the model cannot quote you safely.
How long does it take to start getting cited by AI assistants?
Faster than classic SEO, often weeks rather than months, because answer engines re-crawl frequently and the bar is structural rather than authority-based. Once your key pages are answer-shaped, schema is in place, and llms.txt is published, assistants can begin citing you on the next crawl. The slow part is consistency: the facts have to stay identical across every page over time.

About the author

RJ Murray

Founder

RJ Murray founded AtlasForge and builds the product himself. AtlasForge is the agency that builds a business its entire website for free, shows it on a private link within 48 hours, and charges a monthly subscription only to publish it. RJ writes the engineering behind that: the Next.js builds, the programmatic SEO engine, the answer-engine work that gets a site cited by ChatGPT and Perplexity, and the autonomous content pipeline that drafts these posts for him to review and publish. He works from Waterloo, Ontario, and most of the live sites in the AtlasForge portfolio are businesses across the Waterloo-Kitchener region and Ontario. He writes from what he ships: every claim on this site is something the site itself does, and these posts describe the builds behind the receipts, not theory.

Want your site to read like this does?

We use analytics to understand which pages help, with PII redacted and session inputs masked. Your form submissions always reach us regardless of this choice.