You can't buy a place in an AI answer, and there's no special tag that earns one. What gets a business cited is being crawlable by the right bots, having a page that answers the question in a sentence a machine can lift, and saying something the other pages don't. The first part is a checklist. The second is the work.
Many guides on this topic come from vendors that sell AI-visibility tracking, so they list tactics with confident percentages attached. The official documentation from Google, OpenAI and Perplexity is much shorter, and it says less than you'd expect. This article starts from that documentation and adds only what can be checked, as a practical look at organic search and AI visibility.
What does a site need to be cited in AI answers?
It needs to be reachable and eligible, and then it needs to deserve the citation. Google says there are no additional technical requirements for AI Overviews and AI Mode beyond being indexed and eligible to show with a snippet. Everything after that is content quality, and nothing guarantees inclusion.
Google's own page on AI features is blunt about this. It says there are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary," and that you don't need new machine-readable files or special markup. It also warns that "indexing and serving isn't guaranteed."
So the technical layer is short. Each assistant has its own crawler, and each can be switched off with a robots.txt line:
- Google: normal Googlebot crawling and indexing. A page blocked from Search can't appear in its AI features.
- ChatGPT: OpenAI describes OAI-SearchBot as the crawler used "to surface websites in search results in ChatGPT's search features." Sites that opt out won't appear in ChatGPT search answers, though they may still show as plain navigation links. A robots.txt change can take about 24 hours to be picked up.
- Perplexity: PerplexityBot is "designed to surface and link websites in search results on Perplexity," and Perplexity says it isn't used to crawl content for AI foundation models.
If you blocked AI crawlers in 2024 out of caution, check that you didn't block the search ones too. The separate question of training bots versus search bots is covered in this guide to robots.txt and AI crawlers. This is the cheapest fix on the list and the one most often missed.
Does schema markup help you get cited?
Probably not by itself. Google says no special schema is needed for AI features, and a large public test from Ahrefs found no clear citation gain after pages added JSON-LD. Use schema because it's good practice and keeps your data consistent, not as a citation lever.
That test is worth a closer look because so many guides claim the opposite. Ahrefs identified 1,885 pages that added JSON-LD between August 2025 and March 2026 and compared them with 4,000 control pages that never added it, measuring citations for 30 days before and after. The results: Google AI Overviews moved −4.6% against controls, AI Mode +2.4% and ChatGPT +2.2%, the last two statistically indistinguishable from zero. Ahrefs notes that cited pages were almost three times more likely to have JSON-LD than uncited ones, and attributes that to other quality signals rather than the markup.
That last point is the trap. Schema and citations correlate because sites that care about schema also tend to care about everything else. If someone shows you that correlation as proof, ask what they controlled for.
How do you write a page that gets cited in AI answers?
Answer the question in the first two sentences under its heading, in plain words that still make sense if copied out alone. Then add what only your business can add: a number you measured, a decision you made, a mistake and its cost. Pages that only restate what already ranks give an AI nothing new to cite.
The first half is mechanical. A passage that begins "Yes, but only if…" or "It costs between X and Y because…" can be quoted. A passage that begins with three sentences of background can't, because whoever extracts it has to find the answer first. Write each section as if it will be read with the heading cut off.
The second half is where most small-business sites fail, and Google's guidance on helpful content explains why. Among the questions it asks about your page is whether it provides "original information, reporting, research, or analysis." It adds that "attribution or giving credit to other sources doesn't replace the need for original effort." A page that cites five sources and adds nothing is a summary of the top results, and an AI answer is already a summary of the top results. It has no reason to point at yours.
What counts as original for a business site is usually modest:
- Your real prices, ranges or timelines, stated with what drives them.
- How a process works in your company, step by step, including the awkward steps.
- A comparison where you say which option you'd pick and why.
- Numbers from your own work, with the date and the sample behind them.
One academic study of generative engine optimization, from 2023, reported that some content changes could raise visibility in generated answers by up to 40%, and noted that the effect varies across domains. That is a lab result on a benchmark, not a promise for your site, but it fits the pattern: specific, supported content beats general content.
Why does a page get cited for a question it never mentions?
Because AI search doesn't run your customer's question as typed. Google says AI Overviews and AI Mode "may use a 'query fan-out' technique," issuing multiple related searches across subtopics. Your page can be cited for one of those hidden sub-questions, so cover the adjacent questions a buyer would ask next.
In practice, a page titled "website redesign cost" can be pulled in for a sub-question about what makes redesigns run over, if it has a clearly headed section on exactly that. This is the argument for a small cluster of linked pages on one service, each answering one narrow question well, rather than one long page that tries to cover everything. Narrow pages are the ones that get matched.
How do you know whether it's working?
Watch Search Console for Google, and test the assistants by hand. Google counts AI Overview and AI Mode traffic in the Performance report under the "Web" search type, so it doesn't get its own line. For ChatGPT and Perplexity, ask your ten most important customer questions every month and note whether you're cited, and who is instead.
Keep expectations honest. Citation tracking is noisy, answers differ from one run to the next, and most published percentages come from vendors with something to sell. Treat a manual log of cited or not cited, month by month, as more trustworthy than a dashboard score.
If you'd rather have onlyspell audit your crawler settings and rebuild your key pages around the questions your buyers ask, talk to us about your project.
Frequently asked questions
Do I need to pay or register to appear in ChatGPT or Google AI answers?
No. Google says there are no additional requirements beyond being indexed and eligible for a snippet, and OpenAI and Perplexity both recommend allowing their search crawlers in robots.txt. Appearing isn't guaranteed, though, and nobody can sell you a guaranteed place.
Should I block AI crawlers to protect my content?
You can block the bots that collect training data without blocking the ones that power search answers. Blocking OAI-SearchBot means your site won't appear in ChatGPT search answers, and PerplexityBot is how Perplexity surfaces and links sites. Decide bot by bot, not with one blanket rule.
How long until my site starts being cited?
There's no reliable published timeline. OpenAI says robots.txt changes can take about 24 hours to take effect, but earning citations depends on indexing, ranking and the quality of your pages. Be wary of any source that promises a number of days without showing how it measured.
Sources
If you still need additional help in developing and designing your website, our magic is here to help. Contact us today, and our wizards will help you immediately with your digital marketing needs!
