How a Small Business Gets Cited by ChatGPT, Perplexity and Google AI Overviews

To get cited by ChatGPT, Perplexity and Google's AI Overviews, you need three things: their crawlers must be able to read your pages, each important question needs a page that answers it plainly in the first few sentences, and your site needs the same basic search visibility it always needed. There is no special file or markup that skips those steps.

I spend a lot of time on this because I build ChatBento, and it's the question small business owners ask me most: "How do I show up when someone asks an AI about my kind of business?" Here's what I tell them, and what I do on our own site.

How do AI engines decide what to cite?

Each engine searches the web, reads a handful of pages, and writes an answer with links to the pages it used. So the question is really two questions: can the engine find and read your page, and is your page the clearest answer to the question it was asked?

Google is unusually direct about the first part. Its documentation on AI features says there are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." A page has to be indexed and eligible to show in Search with a snippet. That's it.

ChatGPT and Perplexity run their own search crawlers:

• ChatGPT uses OAI-SearchBot. OpenAI's bot documentation says sites that opt out of it "will not be shown in ChatGPT search answers."
• Perplexity uses PerplexityBot, which its documentation describes as the bot that surfaces and links websites in Perplexity's results.

Step 1: Let the search bots in

Open your robots.txt (yourdomain.com/robots.txt) and check what it blocks. The common mistake is blocking every AI bot in one go because someone read that AI companies train on your content.

Training and search are separate bots:

You can block GPTBot and still allow OAI-SearchBot. If you block the search bots, you're asking not to be cited.

Also check your host or CDN. Some offer a one-click "block AI bots" setting. Make sure it isn't blocking the search crawlers you want.

Step 2: Give every real question its own clear answer

AI engines quote passages, not whole websites. The pages that get cited usually answer one question directly, near the top, in words the engine can lift without rewriting.

For each question that matters to your business:

1. Put the answer in the first two or three sentences. "Yes, we deliver to Northern Ireland. Orders arrive in 3 to 5 working days and shipping costs £6." Then explain.
2. Use the question as a heading where it reads naturally: "Do you deliver to Northern Ireland?"
3. Be specific. Prices, times, places, names, limits. "Affordable" gives an engine nothing to quote. "$8 a month" does.
4. Keep one version of the truth. If an old blog post says one price and your pricing page says another, an engine may quote either. Update or remove the old page.

This is the same advice I give for training a support chatbot, and it's not a coincidence. An AI support bot and an AI search engine both read your pages and answer from them. Content that's clear enough for one is usually clear enough for the other.

Step 3: Find out what people actually ask

The hardest part isn't writing the answers. It's knowing the questions, worded the way customers word them. Good sources:

• Your inbox and DMs. Any question you've answered three times deserves a page or a FAQ entry.
• Your chatbot transcripts. If you run a support bot, read what visitors type. You get the exact phrasing, typos included, and you see which questions your site can't answer yet. ChatBento keeps every conversation in your dashboard for exactly this.
• Search Console. The Queries report shows the longer questions people type into Google before they land on your site.
• Ask the engines. Type your customers' questions into ChatGPT, Perplexity and Google. Note who gets cited. Those pages show you what "a clear answer" looks like in your niche.

If you want a head start, our free FAQ generator reads a page on your site and drafts ten FAQs from it. Treat the output as a first draft: check every answer against what you actually do.

Step 4: Keep doing the boring SEO

Google says a page must be indexed and eligible for a snippet to appear in AI Overviews. In practice that means the usual work still counts:

• Pages that load, render without errors, and have a sensible title and description.
• Internal links between related pages, so crawlers find them.
• Mentions and links from other sites: directories, partners, reviews, local press.
• A sitemap submitted in Google Search Console and Bing Webmaster Tools.

Structured data is still useful for regular search results, but it's not a ticket into AI answers. Google's same page says there's "no special schema.org structured data that you need to add."

What about llms.txt?

llms.txt is a proposed plain-text file at the root of your site that summarizes it for AI tools. It's popular advice right now, so here is what the 2026 data says:

• Adoption is growing. Originality.ai counted 36,120 llms.txt files in May 2026, up from 4,088 in June 2025, according to PPC Land's write-up of its study.
• Almost nobody reads them. Ahrefs analyzed server logs from 137,210 domains in May 2026. Of roughly 38,000 domains with a valid llms.txt, 97% got no requests for it that month.
• No measurable effect on citations. SE Ranking studied about 300,000 domains and found no link between having the file and how often a domain was cited by AI.
• Google doesn't use it. Google's AI features documentation says you don't need to create "new machine readable files, AI text files, or markup" to appear in AI Overviews or AI Mode.

We do have one at chatbento.com/llms.txt. It's generated from the same data as our pages, so it never goes out of date. That's the honest case for it: cheap, harmless, possibly useful later. It is not a way to get cited, and I wouldn't spend money on it or put it ahead of steps 1 to 4.

How do you know if it's working?

There's no perfect report yet, but you can get a decent picture:

• Referrals in your analytics. Look for visits from chatgpt.com, perplexity.ai and similar domains. The numbers are usually small, but they tend to be people with a specific question.
• A monthly spot check. Keep a list of ten questions your customers ask. Once a month, ask them in ChatGPT, Perplexity and Google, and note whether you're cited and which page was used.
• Which pages get cited. If one page keeps getting picked, look at why. It's usually the most direct answer. Do more of that.

The short checklist

• OAI-SearchBot, PerplexityBot and Googlebot allowed in robots.txt
• Host or CDN "block AI bots" setting checked
• Your top 10 to 20 customer questions each answered in the first sentences of a page
• Old pages with outdated prices or policies updated or removed
• Sitemap submitted to Google and Bing
• Ten test questions checked in each engine once a month
• llms.txt only if you have an hour to spare, and with no expectations

Answering your customers' real questions clearly pays off twice: AI engines have something to cite, and visitors who land on your site get their answer straight away. Start with the FAQ generator, or see all our free tools.