Getting "cited" by an AI search engine means your business is named, quoted, or linked as a source when someone asks ChatGPT, Perplexity, Google AI Overviews, or Gemini a question in your industry. It is quickly becoming the most valuable real estate in organic marketing — and unlike a Google ranking, there is no page two. An AI answer names a few sources or none at all.

I'm David Campbell, founder of Nerd Stack. We build websites for businesses in Denver and across the U.S., and over the past year "how do we show up in ChatGPT?" has gone from a curiosity to the first question new prospects ask. This guide is the tactical companion to our complete Answer Engine Optimization guide: where that piece covers the strategy, this one covers execution — exactly what to change so an AI engine picks your page as a source.

When someone asks an AI engine a question, it doesn't return ten links. It writes a single synthesized answer and attributes a small number of sources — usually two to seven. Getting cited means being one of them.

There are two forms of citation, and both matter:

  • Linked citations — a clickable source link beside or beneath the answer. These drive referral traffic you can see in analytics.
  • Unlinked mentions — the answer names your business in its text ("Denver studios like Nerd Stack...") without a link. No click, but the AI has positioned you as an authority in front of a buyer at the exact moment they're deciding.

The mistake most businesses make is treating this like SEO — chasing a ranking position. AI citation is less like ranking and more like being the expert a journalist calls for a quote. You earn it by being genuinely quotable.

Why Most Businesses Never Get Cited

AI engines don't cite websites. They cite chunks — a passage, usually the paragraph or two underneath a single heading. The model retrieves the chunk that best answers the question and folds it into the response.

That one fact explains why most business websites are invisible to AI search. They're written as brochures — atmospheric copy about "passion" and "partnership" — not as a set of self-contained, liftable answers. The four most common reasons a page never gets cited:

  • No question-shaped sections. A heading that says "Our Services" can't be matched to a question. "How much does a website cost?" can.
  • The answer is buried. The useful sentence is in paragraph four, after the throat-clearing. The model extracts the top of the chunk.
  • The content is generic. Vague claims anyone could write give the model no reason to pick you over a thousand identical pages.
  • The page is invisible to AI crawlers. If your content only renders after JavaScript runs, most AI crawlers never see it. We cover the fix in our Core Web Vitals 2026 guide.

First, Can the Crawler Even Reach You?

Everything below this section assumes something nobody checks: that an AI engine can actually fetch your pages. Plenty of sites fail at that first step, and the reason is usually one line in a file nobody has opened in years.

Not all AI crawlers do the same job, and the distinction matters more than the names suggest:

  • The ones that answer questions — OAI-SearchBot and ChatGPT-User, PerplexityBot and Perplexity-User, ClaudeBot, and Google-Extended. These fetch a page because someone is asking something right now. Block these and you cannot be cited at all, no matter how good the page is.
  • The ones that collect training data — GPTBot, CCBot, Applebot-Extended. Blocking these is a legitimate editorial choice about whether your archive trains a model. It is a different decision, with different consequences, and the two get confused constantly.

The confusion is not hypothetical. At the time of writing, The Verge's robots.txt allows GPTBot — the training crawler — while disallowing ChatGPT-User, ClaudeBot, PerplexityBot and Google-Extended. Whatever you make of that trade, it is clearly deliberate. Somebody sat down and decided it.

The problem for everyone else is that those same lines travel. They get copied out of a template, shipped as a plugin default, or pasted from a blog post about "blocking AI scrapers," and they land in small business robots.txt files where nobody reads them. You can be invisible to AI search because of a file you have never opened.

The second thing that hides you: JavaScript

AI crawlers overwhelmingly do not execute JavaScript. If your copy only exists after the page hydrates, a browser shows a full page and a crawler sees close to an empty one. This is the same problem that hurts your Core Web Vitals, which means fixing it pays twice — but it is worth testing separately, because a page can feel fast and still be empty to a crawler.

What about llms.txt?

An emerging convention rather than a standard: a markdown file at /llms.txt pointing engines at the pages you consider canonical. Adoption is uneven and the honest answer is that nobody can yet show it moves citations on its own. It costs about twenty minutes, a handful of engines read it, and we cover the detail in what llms.txt actually does. Treat it as cheap insurance, not a lever.

All three of these are checkable in about ten seconds with our free AI Search Readiness check — no email required. Start there, and only then work through the rest of this guide.

The Uncomfortable Truth: Most of the Power Is Off Your Own Website

Here's the finding that surprises business owners most. According to Search Engine Land's 2026 analysis of the content traits LLMs quote most, the strongest single predictor of being cited by ChatGPT isn't your domain authority, your backlink count, or how much content you publish. It's branded web mentions — how often other people write about your business by name.

Put bluntly: ChatGPT cites you because other people already cited you. The model has learned, across its training data and live retrieval, that your name co-occurs with expertise in your field. Your own website confirms the story; it rarely originates it.

That reframes the whole job. Getting cited is partly a content task and partly a reputation task:

  • Earn mentions in places AI trusts. Industry trade publications, local press, podcasts, and well-moderated communities. One genuine feature in a publication your industry respects outperforms fifty directory listings.
  • Be described consistently. AI engines build an "entity" for your business. If you're "Nerd Stack" on your site, "Nerd Stack LLC" on Google, and "NerdStack Web" on a directory, you're diluting that entity. Pick one name, one description, one category — everywhere.
  • Mind the high-trust platforms. Industry analyses in early 2026 found Wikipedia accounts for a meaningful share of ChatGPT citations, and YouTube overtook Reddit as the most-cited social platform. You won't get a Wikipedia page as a small business — but the principle holds: presence on platforms AI already trusts compounds.

How to Make a Page Genuinely Citation-Worthy

Off-site reputation gets you considered. On-page craft gets you extracted. This is the part you fully control, and it's mechanical once you see it. Every important page should be built from self-contained, answerable chunks.

The recipe for one chunk:

  1. Write the heading as a real question. The exact words a customer would type or speak. "How long does a website redesign take?" not "Timelines."
  2. Answer it in the first one or two sentences. Directly and completely, before any context. The opening sentence is what gets quoted.
  3. Make the chunk stand alone. Read the passage with the rest of the page hidden. If it still makes sense and still answers the question, it's liftable. If it leans on the paragraph above, rewrite it.
  4. Be specific and named. "Most redesigns take six to ten weeks" beats "redesigns can take a while." Numbers, timeframes, named tools, and concrete examples give the model something to anchor to.
  5. Then expand. After the direct answer, add the nuance, caveats, and detail that make it trustworthy.

A quick before-and-after. Before: a heading "Our Approach" followed by "We believe every project deserves a thoughtful, collaborative process tailored to your unique goals." That chunk answers no question and will never be cited. After: a heading "How does the website design process work?" followed by "A custom website typically takes six to ten weeks across five stages: discovery, design, development, review, and launch." That second chunk can be lifted into an answer verbatim.

Bulleted lists, comparison tables, clear definitions, and a genuine FAQ section all make extraction easier. It's also why schema markup for AI search matters — it labels these chunks so the model doesn't have to guess what they are.

Freshness: The 30-Day Citation Window

AI engines strongly favor recent content. One 2026 analysis of ChatGPT citation patterns found that the large majority of cited pages had been updated within the previous 30 days. Models retrain and refresh their indexes far more often than Google ships a core update, and they treat a recent date as a proxy for "still accurate."

Practically:

  • Pick your 15–20 most important pages — money pages and pillar posts.
  • Put them on a 90-day review cadence. Update real things: a statistic, an example, a price, a new section.
  • Show a visible "Last updated" date, and make sure the date in your sitemap and structured data matches it.

One caveat: freshness means meaningful updates. Changing a date without changing the content is exactly the kind of trick models and search engines are built to ignore.

Publish Things Worth Quoting

AI engines preferentially cite the source that contains a fact — not the tenth site to repeat it. That's a real opening for small businesses, because you sit on data nobody else has.

You don't need a research department. You need to count things you already do:

  • "Across the last 40 sites we launched, the average load time at handoff was 1.4 seconds."
  • "We surveyed 30 of our clients; 73% said their old website generated fewer than two leads a month."
  • "Of our redesign projects, 8 in 10 kept or improved their Google rankings within 60 days."

One original, honest, specific number — published in a clearly dated post — becomes a magnet. Every time an AI answers a question that statistic addresses, you are the citable source. We call this citation gravity, and it compounds over time.

A 30-Day Plan to Earn Your First AI Citations

  1. Week 1 — Audit. List the 20 questions your customers actually ask. Check whether any page on your site answers each one in a clean, liftable chunk. Most won't.
  2. Week 2 — Rewrite. Convert your top five pages to question-shaped headings with direct-answer openings. Add a genuine FAQ section to each.
  3. Week 3 — Publish original data. Write one post built around a number only your business can report. Date it clearly.
  4. Week 4 — Earn one mention. Pitch one local publication or industry newsletter, or contribute one genuinely useful answer in a community your customers read. One real mention is the goal.

How Do You Tell If It's Working?

AI-citation measurement is still young. Three checks that work today:

  • Prompt testing. Keep a list of 20–30 real customer questions. Once a month, run them through ChatGPT, Perplexity, Gemini, and Claude, and note where you appear.
  • Referral traffic. Watch analytics for referrers like chatgpt.com and perplexity.ai. It's your simplest leading indicator.
  • Crawler activity. Check server logs for GPTBot, ClaudeBot, and PerplexityBot. Pages they crawl often tend to be pages they cite.

Our AEO guide goes deeper on measurement tooling.

Frequently Asked Questions

How long does it take to get cited by ChatGPT?

For on-page changes — question-shaped headings, direct answers, schema — measurable improvement usually appears within 30–60 days as AI crawlers re-index. For citations that depend on earned mentions, expect 60–90 days of consistent effort before they start to compound.

Can a small local business get cited, or is this only for big brands?

Local businesses often have an easier time. There's far less competition for the citation slots on local-intent questions ("best web designer in Denver") than on national ones, and a strong Google Business Profile plus a little local press goes a long way. See our Local SEO 2026 guide.

Yes. An unlinked mention still puts your name in front of a buyer as the AI's chosen authority, at the moment they're researching a decision. It shapes the shortlist even when it doesn't drive a measurable click.

Is getting cited different from ranking on Google?

It's related but not the same. AI citation is built on SEO fundamentals, then adds direct answerability, citable formatting, structured data, and earned mentions. The right framing for 2026 is "SEO plus AEO," not one instead of the other.

Do I need to pay for an AI-tracking tool?

Not to start. A spreadsheet of customer questions you test monthly is enough for most small businesses. Paid tools help once you're managing citations across many pages or brands.

Bottom Line

Getting cited by AI search is not a trick or a hack. It's the sum of three honest things: content written as clear, self-contained answers; a reputation other people are willing to mention; and original information genuinely worth quoting. Businesses that build those three things will be the ones AI engines name for the next decade. Businesses waiting for a shortcut will keep waiting.

If you'd like this implemented properly — a site rebuilt around answerable, schema-labeled, citation-ready content — that's the work we do at Nerd Stack. Get in touch and we'll walk through what it looks like for your business.

Sources: Search Engine Land — How to Get Cited by ChatGPT: The Content Traits LLMs Quote Most; Contently — How to Get Your Brand Cited in ChatGPT (2026); Search Engine Journal — Key Enterprise SEO and AI Trends for 2026.