How do you get your firm cited by ChatGPT?

Seven steps, in the order they pay off. Most of this is work an owner or their web person can do without hiring anyone, and the first step is the one that most often turns out to be the whole problem.

By Austin Brewer, founder of AI Visibility Partners · Updated 24 August 2026 · 10 min read

How do you get your firm cited by ChatGPT?

In seven steps, and the order matters more than the effort. Let the AI crawlers in, measure a baseline before changing anything, rewrite headings as the questions buyers type, state plainly what the firm does and where, add structured data, publish the cost and comparison pages competitors have not written, and earn mentions on sites other than your own. Steps one through five are usually a few days of work. Steps six and seven are where the compounding happens, and they never really finish.

None of this makes an engine like your firm. It makes your firm quotable, findable, and hard to contradict, which is the only part of the process anyone outside OpenAI can influence.

Step one: are the AI crawlers even allowed in?

Check this before anything else, because it is the most common single point of failure and it takes minutes to fix. Open yoursite.com/robots.txt and look for the AI crawlers by name: GPTBot and OAI-SearchBot for ChatGPT, ClaudeBot, PerplexityBot, Google-Extended, Applebot-Extended, and CCBot. Some website platforms block them by default under a privacy setting, and many owners have no idea it is switched on. If your content only renders after JavaScript runs, that is the second half of the same problem.

• Check the file. Visit yoursite.com/robots.txt. A blanket Disallow: / under any of those user-agents means you are invisible to that engine by instruction.
• Check the platform setting. Squarespace, Wix, Webflow, and several WordPress plugins expose an "block AI crawlers" or "AI training" toggle. It is sometimes on by default.
• Check without JavaScript. View the page source. If your service copy is not in the HTML, an engine may never see it.
• Publish an llms.txt. A short plain-text summary of what the site contains, at the root. Support is still uneven and it costs almost nothing to add.

Step two: what does the engine say about you right now?

Establish the baseline before you change anything, or you will never be able to prove what worked. Open signed-out sessions in ChatGPT, Perplexity, Gemini, and Google, and run the questions your buyers actually ask, three times each. Record who gets named, which sources are cited, and whether any answer contains something factually wrong about your firm. That record is the thing every later measurement is compared against, and it is worth screenshotting with dates rather than summarising from memory.

Keep the prompt list fixed from this point on. The most common self-inflicted measurement error is improving the prompts over time, which quietly guarantees that later numbers look better than earlier ones. Sample prompt sets are published here, verbatim, if you want a starting list.

Step three: are your headings the questions buyers type?

Most service pages are organised around what a firm wants to say. Engines reward pages organised around what a buyer asks. Convert each heading into the actual question, then put a standalone answer of roughly 50 to 150 words directly beneath it, before any background or context. Standalone means the answer makes sense lifted out of the page entirely, without a preceding sentence to explain what it refers to. If a paragraph opens with the word it, this, or they, it is not standalone.

Step four: does your site say plainly what you do and where?

An engine cannot recommend a firm whose scope it has to infer. Replace slogans with facts: the services offered, the cities and neighbourhoods served, the typical project size, the price band, and who the principal is. This feels blunt to owners who have spent years on brand voice, and it is the single change that most reliably moves an unbranded prompt. Keep the brand voice for the parts of the page a human reads. Put the facts somewhere an engine can lift them in one sentence.

The 5W Public Relations study of local service categories measured an eight-fold entity-strength multiplier on citation share, the largest single signal effect it recorded. Entity strength is largely a matter of a business being described consistently and unambiguously everywhere it appears.

Step five: what structured data should you add?

Three JSON-LD blocks cover most of what a local firm needs, and adding them is a one-off job rather than ongoing work. A ProfessionalService or LocalBusiness block naming the business, its area served, its price range, and its founder. A Service block for each main service line. And an FAQPage block whose answers match the visible text on the page word for word, because a mismatch causes the markup to be ignored. Schema does not persuade an engine of anything. It removes the guesswork about facts the engine would otherwise infer from prose or from a stale directory listing.

Two common errors are worth avoiding. Do not put a price range like "500-900" in a field expecting a single value; use AggregateOffer with lowPrice and highPrice. And do not let the FAQ schema drift from the visible answers as the page gets edited, which happens more often than anyone admits.

Step six: which pages should you publish next?

The ones your competitors have not written, which in this vertical are almost always the same three: a cost guide with real ranges, a comparison page answering the choice buyers agonise over, and a how-to-choose page. Cost and comparison prompts are where homeowners form a shortlist, and they are the prompts directories currently win by default because no local firm has published an answer. A single honest cost page with visible numbers and an updated date frequently outperforms a year of blog posts.

• The cost guide. "How much does a kitchen remodel cost in [city] in 2026" with real ranges, what drives them, and what is excluded. Most firms avoid this. That is exactly why it is available.
• The comparison page. "Design-build firm vs. general contractor", "interior designer vs. architect", "remodeler vs. custom home builder".
• The how-to-choose page. "Questions to ask before hiring a remodeling contractor", "red flags when hiring a kitchen remodeler".

Then maintain them. Seer Interactive found that 75% of the pages AI engines cite had been updated within the previous year, and that refreshed pages beat newly published ones: 72% of cited pages showed a recent update against only 42% newly published.

Step seven: who else on the internet says you exist?

A firm mentioned nowhere but its own website is a firm with one source, and engines weight corroboration heavily. The work here is unglamorous: keep directory listings consistent across Houzz, Angi, Yelp, Thumbtack, HomeAdvisor, Porch, Nextdoor, and Google Business Profile, get named in local press and trade publications, contribute to industry associations such as NARI, NAHB, NKBA, and ASID, appear on supplier and partner sites, and keep the business name, address, and phone identical everywhere they appear. This is the slowest of the seven steps and it is usually the reason a firm that fixed everything else still plateaus.

Consistency matters as much as volume — the discipline SEO practitioners call NAP consistency, for name, address, and phone. Two variants of a business name across a dozen listings can split an engine's understanding of the entity in half, which is a cheaper problem to fix than to earn your way out of. An association membership or a trade-press mention is also the kind of corroborated, attributable signal that search and AI systems weight under what Google's guidelines call E-E-A-T: experience, expertise, authoritativeness, and trust.

Where do these numbers come from?

Every figure on this page comes from a named third-party study, listed below. None of the numbers are ours. Full samples, the scope limits each study states, and the four commonly repeated statistics removed from this site rather than published with a vague attribution are all set out on the sources page.

1. 5W Public Relations, The Local Services AI Visibility Crisis 2026: national chains, platform aggregators, and the structural exclusion of independent operators, April 2026. View the study ↗
2. Seer Interactive, AIO Impact on Google CTR: September 2025 Update, 4 November 2025. View the study ↗
3. Seer Interactive, Study: Content Recency’s Impact on AI Visibility in 2026, 24 July 2026. View the study ↗

Full sources, samples, scope limits, and the four removed figures →

What else do owners ask about getting cited?

Where should you read next?

Once the seven steps are underway, the next questions are usually about measurement, cost, and whether to hire.