How is getting cited in AI engines different from ranking on Google?

Ranking gets your link shown in the results. Citation gets your domain named inside the answer itself, with or without a click. The success metric shifts from clicks to mentions and cited URLs. Technical SEO and content quality still matter for both, but extractable structure and sourced claims matter much more for citation.

Which AI engine should I optimize for first?

Whichever your audience uses most. Google AI Overviews reach the most people incidentally because most searches still start at Google, while ChatGPT and Perplexity are where more deliberate research happens. The universal tactics (sourced statistics, definition-first answers, question-shaped structure) work on all of them, so the first choice mostly affects what you measure.

Why do AI engines cite Reddit so much?

Reddit threads answer specific questions in first person, which is the shape of content engines lift for experiential questions. Google also has a direct pipe: in February 2024 it announced an expanded partnership that gave it access to Reddit's Data API. In Profound's analysis, Reddit was the most cited source for both Perplexity and Google AI Overviews.

Should I create Reddit content as a marketer?

Participate, do not promote. Communities detect and punish promotional accounts. The teams who earn Reddit mentions have people who genuinely take part in relevant subreddits over months, answering questions as practitioners.

Is llms.txt worth implementing?

It is cheap, and useful for assistants and agents that read it, particularly for sites with documentation. It is not a ranking lever for Google: Google says you do not need AI text files or special markup to appear in AI Overviews or AI Mode.

How quickly do changes show up?

Perplexity and ChatGPT search retrieve from the web when they answer, so a change to a page their crawlers can reach can appear quickly. Google AI Overviews draw on Google's index, so they follow its crawling and indexing timeline. Because answers vary from run to run, judge changes over weeks of tracking rather than a single check.

Can I track citations programmatically?

Yes. Otterly.AI lists API and MCP access on its Standard and Premium plans, and other trackers offer exports. For a do-it-yourself start, a fixed set of 30 to 50 prompts run on a schedule works; the hard part is choosing prompts your buyers actually ask.

How to Get Cited in ChatGPT and Perplexity

Why citation matters more than rank

A user asks ChatGPT, Perplexity or Google a question. They get an answer paragraph with a handful of sources beneath it. Most read the answer and stop. Some click a source.

If you are not among those sources, you are not in the consideration set, even when you rank first in the results underneath. Ahrefs analysed 300,000 keywords using December 2025 data and found that when an AI Overview appears, the top organic page's clickthrough rate is 58% lower on average. Authoritas, studying 3,500 UK news searches in April 2025, found clickthrough fell by close to half when an AI Overview appeared: 47.5% on desktop and 37.7% on mobile. Google called that study "inaccurate and based on flawed assumptions", which is worth knowing; every independent study so far has pointed the same direction.

Ranking is no longer enough on its own. The currency is being cited inside the answer the user actually reads. This guide is the playbook for earning that.

What works across all three engines

A small set of interventions help regardless of engine. Do these first.

Add quotations, statistics and source citations

This is the best-documented intervention in the GEO research. The Princeton GEO paper (Aggarwal et al., 2023) ran controlled experiments on how content changes affect visibility in generative answers:

The three strongest methods were adding quotations from credible sources, adding statistics and citing sources.
The best methods improved visibility over unoptimised content by up to 41% on the paper's position-adjusted word count and 28% on its subjective impression measure.
Citing sources raised the visibility of a site ranked fifth in search results by 115%, while the top-ranked site's visibility fell on average. Sourcing is one of the few levers that helps a smaller site against a bigger one.
Keyword stuffing performed worse than doing nothing.

The working rule follows directly: every quantitative claim links to its primary source on first mention. That includes your blog posts, and it is the rule this post follows.

Use definition-first paragraphs

Answer engines lift passages, not pages. The most extractable passage opens with the answer, then supports it.

Narrative: "When we first started exploring this question, we noticed something interesting about how marketing teams approach content scoring..."

Definition-first: "Real-time content scoring is the practice of grading content as it is drafted, not after publication. Three things drive an effective score: coverage of the topic, structural completeness, and sourced claims."

The second can be lifted whole into an answer. The first cannot.

Structure around the questions people ask

People ask answer engines questions, so pages built from clearly headed questions with direct answers match the shape of the request. A substantive FAQ on important pages adds more citable passages, and FAQPage markup makes each pair unambiguous to any system that reads it. Six to eight real questions, answered in the first sentence, is a good format.

Earn presence on the sites each engine trusts

The engines do not lean on the same sources. In Profound's analysis of citation patterns:

ChatGPT: Wikipedia was 47.9% of citations among its top-10 sources.
Perplexity: Reddit was 46.7% of its top-10 citations.
Google AI Overviews: Reddit led at 21.0%, followed by YouTube at 18.8%.

For a B2B brand, your own pages will struggle to beat those sources on general questions. GEO therefore has an off-site component:

An accurate Wikipedia presence where you are genuinely notable.
Real participation on Reddit, as practitioners rather than marketers. Google announced an expanded partnership with Reddit in February 2024 that gave it access to Reddit's Data API.
YouTube content with clear titles and descriptions for how-to and comparison questions.
Mentions in the third-party lists and reviews your category is researched through.

Google AI Overviews specifics

AI Overviews and AI Mode are built on Google Search, and Google's own guidance is unusually direct. According to Google's documentation on AI features, "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." A page needs to be indexed and eligible to be shown with a snippet. Google also says you "don't need to create new machine readable files, AI text files, or markup" and that there is "no special schema.org structured data that you need to add."

What that means in practice:

Classic SEO is the entry ticket. If a page cannot rank and show a snippet, it cannot be a supporting link.
Cover the related questions. Both features may use what Google calls query fan-out, "issuing multiple related searches across subtopics and data sources." Semrush found People Also Ask alongside 90.03% of AI Overviews. Pages that answer the neighbouring questions give Google more to draw on.
Answer early. A page that answers the literal question in its opening paragraph gives any extractive system a clean passage to use.
Treat schema as hygiene. Structured data still helps Google understand a page, but Google is explicit that no special markup is needed for these features.
Know the controls. nosnippet, data-nosnippet, max-snippet and noindex limit how content appears in AI features.

The same Semrush study found AI Overviews spreading beyond informational queries during 2025, with commercial and transactional queries making up a growing share. They are no longer only a top-of-funnel concern.

Perplexity specifics

Perplexity answers from a live web search and shows numbered sources. Two crawlers matter, and Perplexity documents both:

PerplexityBot is used "to surface and link websites in search results on Perplexity." If you block it, you are asking not to be surfaced.
Perplexity-User visits a page when a user's question needs it, and Perplexity says it "generally ignores robots.txt rules," because the fetch is user-initiated.

Beyond access, the Profound data above is the practical guide: Reddit dominates Perplexity's citations, so for questions about your category the answer often includes community discussion and press alongside your own pages. Owned content that is comprehensive, specific and sourced gives Perplexity a reason to cite you next to them. Keep important pages current; an engine that searches live is always comparing your page against whatever was published last.

ChatGPT search specifics

OpenAI launched ChatGPT search on October 31, 2024, and by February 2026 OpenAI said ChatGPT had 900 million weekly active users.

Access comes first. OpenAI's crawler documentation is explicit that "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers." GPTBot, the training crawler, is controlled separately: "Each setting is independent of the others," so a site can allow OAI-SearchBot to appear in search while disallowing GPTBot. Many sites that blocked AI bots in 2024 blocked both without meaning to. Our AI crawler checker shows which of these your robots.txt allows.

On sources, ChatGPT leans heavily on reference material: Wikipedia was nearly half of its top-10 citations in Profound's analysis. For B2B brands the implication is to publish unambiguous canonical answers to the questions prospects ask (definitions, comparisons with real tables, step-by-step guides), because those are the pages that read like references.

Where llms.txt fits

llms.txt was proposed by Jeremy Howard in September 2024 as a way "to provide information to help agents use a website": a markdown file at the root of your domain with a title, a short summary and sections of annotated links.

It is cheap to add and useful for assistants and agents that read it, especially on sites with documentation. It is not an official standard, and it is not a Google ranking lever: Google says AI text files are not needed to appear in AI Overviews or AI Mode. Treat it as a service to the tools that use it, not as a way to move rankings. Our llms.txt generator builds one from your pages.

Tracking your citations

Some of the tools in this space, with details from their own pages as of September 2026:

Otterly.AI. Tracks ChatGPT, Google AI Overviews, Perplexity and Microsoft Copilot on every plan, with Claude, Google AI Mode and Gemini as add-ons. Plans start at $29 a month for 15 prompts, refreshed daily. It was named a Cool Vendor in the 2025 Gartner Cool Vendors for AI in Marketing.
Profound. A dedicated enterprise platform that raised a $180 million Series D at a $1.8 billion valuation in September 2026. It offers a free trial of 7 days on three engines; ongoing tracking is Enterprise.
Veritas (our product). Tracks ChatGPT, Perplexity, Gemini and Google AI Overviews on every plan, from 5 prompts on Starter, re-checked every two weeks, with the editor that drafts the fix in the same workspace.
Manual sampling. A fixed set of 30 to 50 prompts run on a schedule is a real starting point. It is noisy and free.

For the full field, see the best AI visibility tools, or the head-to-heads with Otterly.AI and Profound.

Whatever you use, measure the right thing. SparkToro's research across 2,961 prompts found AI tools returned the same list of brands less than once in a hundred runs. A single "position" means little. Track brand mention rate across many prompts, cited URL rate, and share of voice against named competitors, for each engine separately.

A practical execution sequence

If you can only do five things this quarter, do these in order:

Check crawler access. Confirm OAI-SearchBot, PerplexityBot and Googlebot can reach your important pages, and decide separately what to do about training bots.
Add sourced statistics, quotations and inline citations to your top ten pages. This is the intervention with the strongest published evidence.
Add a real FAQ to pillar and product pages, answer-first, covering the questions around the main one.
Audit your presence on Wikipedia, Reddit and YouTube, and plan three months of genuine participation where you have gaps.
Set up tracking across at least Google AI Overviews, Perplexity and ChatGPT, and establish a baseline before you change anything else.

Expect movement to show up unevenly across engines. Track each one separately and judge the trend over weeks.

Closing

The hard shift in GEO is accepting that the first reader is often a machine deciding whether you are worth quoting. Search engines rewarded content a person would click. Answer engines reward content a system can lift cleanly and trust. The teams that write for that, with every claim sourced and every answer up front, are the ones being cited.

Veritas drafts marketing content from your own documents with a citation on every claim, the format answer engines can trust. Try Veritas free or see how it tracks AI visibility.

How to Get Cited in ChatGPT, Perplexity, and Google AI Overviews