How to get cited by AI: what actually works

Getting cited by AI is not a trick. It is three plain things done in order, and doing them honestly beats chasing whatever is trending this week. This page leads with the levers that have real evidence behind them, flags the ones built on vendor or simulated tests, and names the fake levers you can skip. If you want the simplest version, the numbered steps below mirror the answer at the top of this page.

Why bother. Because more of your customers now ask an AI first. BrightLocal's consumer research found the share of people using AI to find local businesses jumped from about 6 percent to about 45 percent in a single year. When AI answers the question, the business it names wins the customer. Everyone else is invisible.

How does AI decide what to cite?

AI does not recall your business from memory. AI retrieves it. When someone asks a question, the AI runs a live search, pulls a small set of pages, writes its answer from those pages, and names its sources. That process is called retrieval, and the whole game is getting retrieved. If AI never pulls your page, it can never cite you. To get cited by AI, you have to be one of the pages it pulls.

One question is really many questions. Behind a single prompt, the AI quietly fans out into several smaller sub-searches, then stitches the results into one answer. So you are not trying to rank once for one phrase. You are trying to be present across all the little questions your buyer's prompt breaks into. Being the single top result for one keyword is not the goal. Showing up again and again across the sub-searches is.

That is why one honest pattern wins, and it is the pattern this whole page is built on. Make a claim on your own site. Frame why the claim is true. Then prove it somewhere else. The proof is the part that decides it, and the strongest proof lives off your own site, on the third-party pages AI already trusts. You can say you are the best plumber in town all day on your own homepage. It counts for little until other trusted places say it too.

How do you get cited by AI?

You clear three gates in order, then you measure. There is no shortcut past this, and doing them in order saves you money. Here is the whole plan as a numbered how-to.

1. Let AI crawlers read your site. This is the gate before every other step. Make sure your robots.txt and your CDN allow the AI bots to fetch your pages. The most common cause of zero AI visibility is a security setting quietly blocking them. See our full guide to AI crawler access to check yours.
2. Publish clear, quotable content with facts and citations. Write short, direct pages that answer the real questions buyers ask, and back your claims with specific facts, statistics, and named sources. This is the exact kind of text AI lifts into an answer.
3. Earn brand mentions on the sources AI trusts. Get your business named on the third-party sites AI reads, like directories, review sites, and industry pages. This is the single strongest known signal, and it is the slowest to build.
4. Check monthly whether AI is naming you. Ask your buyer's real questions in ChatGPT, Perplexity, and Google, and write down whether you are named and linked. That is how you know the work is landing.

Notice the order. Access is the floor. Content is what gets lifted. Mentions are what tip the pick your way. Skipping to step two or three while a crawler is blocked is spending money on a locked door.

What actually makes AI cite you? (the evidence)

The honest answer is that a few levers have real evidence, several have vendor or simulated evidence, and some have none. So here is every lever, whether it works, and the exact study behind it. We flag the vendors and carry every caveat, because that honesty is the whole point.

What this means for you. Put your effort where the evidence is strongest. Fill your pages with real facts and cited sources, and work to get your brand named on other trusted sites. Those two levers have the best proof we have, even with their caveats. This is the core of real generative engine optimization, and it is why a good business can still stay invisible if it skips them. For the deeper why, read why isn't my business showing up in ChatGPT.

Does it work the same on ChatGPT, Perplexity, and Google?

No. The three levers apply across the board, but each engine weights sources differently, so the fine detail changes by engine. Treating them as identical is a common mistake.

Two things change by engine. The first is how the engine finds pages. ChatGPT's search and Microsoft Copilot lean on Bing's index, which is why claiming Bing Webmaster Tools matters more than most people think. Perplexity leans harder on fresh, recent content. Google AI Overviews run on Googlebot's own crawl and blend in classic ranking signals.

The second is which third-party sources each engine tends to trust. As a rough guide, Perplexity often leans on Reddit and YouTube. ChatGPT often leans on Wikipedia and Reddit. Google AI Overviews often favor YouTube and Quora. These are tendencies, not rules, and they shift over time as the engines change, so do not treat them as a fixed checklist. The point stands. Getting named on the sources one engine trusts is not the same as getting named on another's, which is why a cross-engine plan beats chasing one.

The playbook is the same three gates, but the tuning differs. When you are ready to go engine by engine, we have separate guides for how to rank in ChatGPT, how to rank in Perplexity, and how to rank in Google AI Overviews. This page is the cross-engine foundation they all build on. Rather than cram every engine's detail here, follow the link to the engine you care about.

What does NOT work?

Four things get sold as citation levers that the evidence does not support. Skip them as standalone plays and put the money where the proof is.

Schema markup by itself. Ahrefs ran a controlled test on 1,885 pages and adding schema moved AI citations within noise, roughly minus 5 to plus 2 percent across AI Overviews, AI Mode, and ChatGPT. Google says no special schema is needed for AI Overviews. Schema is still worth having for rich results and entity clarity, so it is not useless, it just is not a citation lever on its own. We wrote the honest version at schema markup for AI search.

The llms.txt file. Google's John Mueller called it "purely speculative," and Gary Illyes said in mid-2025 that Google does not support it. No AI system is on record reading it. Adding one does no harm, but do not expect a citation from it. Our full take is at does llms.txt work.

Keyword stuffing. In the Princeton study, stuffing keywords scored below the plain baseline. It reads badly to humans and it does not help with AI either.

Buying domain authority. A high domain rating is a weak signal here. Ahrefs found brand mentions correlated about 3 times stronger than backlinks, and low-authority pages still get cited often. Chasing a big authority score is the wrong place to spend.

How do I know if AI is citing me?

Run a manual prompt test, because there is no clean dashboard for this yet. The test is simple and you can do it in twenty minutes a month.

1. Write down the real questions a buyer would type, like "best plumber in Austin" or "who does HVAC repair near me."
2. Ask each question in ChatGPT, Perplexity, and Google's AI Overviews, one at a time.
3. For each answer, record whether your business is named, whether it is linked, and who got named instead.
4. Do it from a signed-out or private browser window, so your past chats do not skew the result.
5. Repeat the same list every month and watch the trend, not a single day.

That log is your scoreboard. If you would rather not track it by hand, our free AI-Visibility Checklist walks you through the same test and scores your site. One honest caution while you read the results. Independent research from Pew found people click a link inside an AI summary only about 1 percent of the time, so being cited is about being the trusted answer, not just chasing clicks.

How long does it take?

Months, not days. Anyone promising to get you cited in days is guessing, or selling. Here is the honest timeline.

Fixing crawler access can pay off fast, sometimes within a single crawl cycle, because you are removing a block rather than building something new. Improving your content sits in the middle. But the biggest lever, earning brand mentions on the sites AI trusts, is also the slowest. That work typically shows up over about three to six months or more, because you are waiting on other people to publish, on AI systems to re-crawl, and on the signal to build. It compounds, which is the good news. The businesses that started earning mentions early are the ones getting named now.

Common questions about getting cited by AI