How to Get Your Business Cited in ChatGPT and Perplexity: A Practical Guide for Small Teams

Type "best SEO tool for a Shopify store" into ChatGPT. You'll often get three or four brand names and a few source links, and one of those links is usually a Reddit thread. Your brand isn't there, even though you rank on page one of Google.

That gap is the reason to learn how to get cited in Perplexity and ChatGPT. When we audited our own site, cocoseo.co, it scored 24 out of 100 for citability. This guide shows the fixes we're making, so you can apply them to your site.

What You'll Build

By the end, your crawlers will be open to AI search bots. Your top pages will have answer blocks, sourced stats, and FAQ schema. You'll also have a mention plan for off-site channels and a spreadsheet that tracks how often ChatGPT and Perplexity cite you across repeated runs.

Before You Start

• CMS admin access. You need WordPress 6.x with Yoast or Rank Math, Shopify with theme editing rights, or Ghost with theme upload access.
• Access to your robots.txt. On Shopify, that's the robots.txt.liquid theme template. On WordPress, use your SEO plugin's file editor or edit the file at your site root.
• Google Search Console, verified for your domain.
• Free accounts on ChatGPT and Perplexity.
• A spreadsheet. Google Sheets is fine.
• Basic comfort editing a text file. No coding beyond copy and paste.
• Time: about 3 to 4 hours for the first pass, then roughly 30 minutes a week.

Step 1: Your robots.txt Decides Whether AI Search Can See You

AI engines can only cite pages their crawlers can reach. Many sites block these bots by accident through old rules, security plugins, or CDN settings. This step opens access to the search crawlers and gives you a separate choice about training crawlers.

OpenAI's documentation (2026) says sites that block OAI-SearchBot won't appear in ChatGPT search answers. That crawler is separate from GPTBot, which collects training data. So you can show up in ChatGPT search without handing over your content for model training.

1. Open your robots.txt file.
2. Add these rules:

User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Perplexity-User
Allow: /

# Optional: opt out of model training only
User-agent: GPTBot
Disallow: /

3. Save and publish.
4. Skip llms.txt for now. SE Ranking (2025) looked at 300,000 domains. Sites with llms.txt were cited no more often than sites without it.

✅ Check: Run this in your terminal:

curl -A "OAI-SearchBot" -I https://yoursite.com/

You should see HTTP/2 200. A 403 means something upstream is blocking the bot.

Step 2: Answer Blocks Are the Core of How to Get Cited in Perplexity and ChatGPT

AI engines pull one passage at a time, not your whole page. If the text under a heading is a slogan, there's nothing to quote. This step puts a short, factual answer directly under each heading so every section can stand on its own.

Our audit flagged this first. The cocoseo.co homepage led with "SEO and social content, on autopilot." That's a tagline, not a fact an engine can lift.

Here's the rewrite:

CocoSEO is an AI content engine that researches keywords, writes cited articles, and publishes them to WordPress, Shopify, or Ghost. You review each draft in about 60 seconds. It also tracks citations in ChatGPT, Perplexity, and Google AI Overviews.

Apply the same pattern to your pages:

1. Open Search Console and list your 10 pages with the most clicks.
2. For each H2 on those pages, write one sentence that answers the heading directly.
3. Put that sentence right under the heading.
4. Name the subject. Write "Acme Coffee ships to all 50 states," not "We ship everywhere."
5. Keep each block between 40 and 60 words.
6. Add one number if you have one: a price, a count, a date.

For more templates, see our guide on writing answer blocks for AI search.

✅ Check: Copy one section into a blank document. If a stranger can understand it without the rest of the page, it passes.

Step 3: Sourced Statistics Give Engines a Reason to Quote You

Generative engines favor passages with specific, attributed evidence. Vague claims like "most customers love us" give a model nothing to verify. This step swaps general statements for named sources, real numbers, and quotes from people.

The most neutral research here comes from Aggarwal et al. (Princeton/Georgia Tech, KDD 2024). They found that adding quotations raised visibility in generative engines by 41%. GEO methods overall raised it by up to 40%.

1. Find every unsupported claim on your top 10 pages.
2. Replace each one with this pattern: "According to [Source] ([Year]), [specific finding]."
3. Link to the primary source, not a blog that summarized it.
4. Add one quote from a named person. Your own customers count.
5. Publish data only you have. Think average delivery time, return rate, or the number of stores you serve.

Original data matters most. A Shopify candle shop that publishes "We shipped 4,200 orders in 2025, and 1.8% were returned" owns a fact nobody else can cite first.

✅ Check: Count your attributed numbers. You want roughly one every 150 to 200 words.

Step 4: FAQ Schema Makes Your Answers Machine-Readable

Schema markup labels your content so machines can tell a question from an answer, a price from a feature. Our audit flagged missing FAQPage and SoftwareApplication schema on cocoseo.co. This step adds FAQPage markup to the pages where people ask the same questions repeatedly.

1. Pick a page with real customer questions, like pricing or shipping.
2. Add this JSON-LD in the page head, or use Rank Math's FAQ block on WordPress:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [{
    "@type": "Question",
    "name": "How long does shipping take?",
    "acceptedAnswer": {
      "@type": "Answer",
      "text": "Orders ship within 2 business days and arrive in 3 to 5 days in the US."
    }
  }]
}

3. Make sure the same questions and answers appear as visible text on the page.
4. Add Organization schema to your homepage. Software companies can add SoftwareApplication schema with price and features. Store operators can add Product schema.

One honest tradeoff: Google limited FAQ rich results to well-known government and health sites in 2023, so you probably won't see the snippet in Google. The case for FAQ schema now is cleaner parsing. Evidence that it directly raises AI citations is thin. It takes about 20 minutes, so treat it as cheap insurance, not a growth lever.

✅ Check: Paste your URL into validator.schema.org. You should see zero errors.

Step 5: Off-Site Mentions Carry More Weight Than Backlinks

Most AI citations point to pages you don't own. Engines lean heavily on Reddit, Wikipedia, review sites, and LinkedIn. This step builds honest mentions on the platforms engines already trust, which on-site fixes alone can't replace.

The numbers are lopsided. Ahrefs (2025) studied 75,000 brands. Branded web mentions correlated with AI Overview visibility at 0.664, compared with 0.218 for backlinks. Similarweb (2026) found that Wikipedia and Reddit each make up roughly 12–13% of U.S. ChatGPT citations.

Source pools also shift fast. Semrush (2025) saw Reddit's share of ChatGPT citations fall from about 60% to about 10% between August and mid-September 2025. Profound (via Contently, 2026) saw LinkedIn rise from about #11 to #5 among cited domains between November 2025 and February 2026.

1. Claim your profiles. SaaS teams should use G2 and Capterra. Local businesses should use Google Business Profile.
2. Answer three Reddit threads a week where you can genuinely help. Disclose that you work for the company.
3. Turn each article into a LinkedIn post. CocoSEO creates these drafts for you, or you can write them by hand.
4. Ask happy customers for one review on a third-party site.

A note on gaming this: Peec AI (2026) found that about 1 in 10 AI citations came from self-promotional listicles, meaning a vendor's own "best tools" post that ranks itself first. It works today. We wouldn't build a strategy on it, because readers notice and engines will likely adjust.

✅ Check: Ask Perplexity "What is [your brand]?" and look at the sources panel. At least one should be a page you don't own.

Step 6: Repeated Tracking Shows How to Get Cited in Perplexity and ChatGPT Over Time

AI answers change every time you ask. A single check proves nothing. This step sets up a simple tracking sheet that measures how often you appear across many runs, and it treats ChatGPT and Perplexity as different targets.

SparkToro/Gumshoe (2025–26) tested 2,961 prompts. ChatGPT or Google AI returned the same brand list less than 1 time in 100. Rand Fishkin's advice: measure how often you appear across many runs, not your "rank."

The two engines also behave differently:

| Factor | ChatGPT | Perplexity | |---|---|---| | Overlap with Google top 10 (Ahrefs, 2026) | About 12% of cited URLs | About 1 in 3 citations | | Share of AI referral traffic (SE Ranking, 2026) | 74.78% | 7.23% | | What to prioritize | Off-site mentions, answer blocks | Google rankings, fresh content |

1. Write 20 prompts your customers would type, like "best eco candle brand under $30."
2. Run each prompt 5 times in ChatGPT and 5 times in Perplexity, using fresh chats.
3. Record two columns: mentioned (yes/no) and cited URL.
4. Calculate your mention rate: mentions divided by total runs.
5. In GA4, filter referral traffic for chatgpt.com and perplexity.ai.
6. Repeat monthly.

A caveat on our sources: Ahrefs, SE Ranking, Peec AI, and Profound all sell AI visibility tools, and so do we. Treat their numbers as directional. CocoSEO automates this tracking, but a manual spreadsheet gives you the same baseline.

✅ Check: You have one baseline number per engine, such as "mentioned in 14 of 100 ChatGPT runs."

Troubleshooting

OAI-SearchBot gets a 403 even though robots.txt allows it

A 403 error for OAI-SearchBot, even when robots.txt allows the bot, means your CDN or firewall is blocking OpenAI's crawler before OAI-SearchBot ever reads robots.txt. Robots.txt is a set of instructions for crawlers that reach your site. Robots.txt cannot override a security layer, such as Cloudflare, that rejects the request first.

In Cloudflare, check the bot and AI crawler settings under Security. Allow OAI-SearchBot, OpenAI's search crawler, and PerplexityBot, Perplexity's crawler, then rerun the curl test. If the curl test still returns a 403 after you allow both bots, the block is coming from another rule in your CDN or firewall. Review any other security rules in Cloudflare or your hosting firewall that filter automated traffic. Adjust those rules so OAI-SearchBot and PerplexityBot can reach your pages, and rerun the curl test after each change until the response is no longer a 403.

Your page ranks in Google, but ChatGPT never cites it

A Google ranking and a ChatGPT citation are separate outcomes, so a page that ranks in Google but never appears in ChatGPT answers is normal. Ahrefs (2026) found that only 12% of URLs cited by ChatGPT, Gemini, and Copilot also ranked in Google's top 10. A top-10 Google position does not guarantee an AI citation.

Because Google rankings and AI citations overlap so little, focus on off-site mentions and answer blocks instead of chasing more Google rankings.

You appeared once, then vanished

AI answers are close to random from run to run, per SparkToro (2025–26). Don't react to one result. Compare mention rates month over month.

The schema validator shows duplicate FAQPage errors

Duplicate FAQPage errors in a schema validator mean two sources on your site are outputting the same FAQ structured data at the same time, usually your theme and your SEO plugin. The fix is to turn off FAQ schema in one of the two sources so the page outputs a single FAQPage block.

To decide which FAQ schema to keep, compare each version against the questions and answers readers can actually see on the page. Keep the FAQPage schema that matches your visible on-page text, and disable the FAQ schema setting in the other source, whether that source is the theme or the SEO plugin. FAQ schema should describe the same questions and answers that appear on the page, so the version that mirrors your visible text is the one to keep. After you turn off the duplicate source, run the page through the schema validator again. The duplicate FAQPage error should no longer appear. If the error remains, check whether another source is still adding FAQ schema to the page.

Perplexity cites an outdated version of your page

An outdated Perplexity citation is a freshness problem. Perplexity favors fresh content, so the fix is to update the page, change the visible "last updated" date, and request reindexing in Google Search Console. After those three steps, check Perplexity again in one to two weeks to confirm Perplexity cites the current version.

What to Do Next

Once your top 10 pages pass, expand to pages that answer comparison questions: pricing, alternatives, and integrations. Our audit flagged missing integration pages for Shopify, WordPress, Ghost, and Webflow as the next gap on cocoseo.co, and that's likely true for many software sites. Build one focused page per integration, each with its own answer block and FAQ schema. Then review your tracking sheet monthly and double down on the channels producing mentions. Our guide to AI crawler settings in robots.txt covers the advanced bot rules.

Further reading

• Aggarwal et al., "GEO: Generative Engine Optimization" (KDD 2024): the most neutral research on which content changes raise AI visibility.
• OpenAI's crawler documentation: the official list of OpenAI bot names and what each one does.
• Perplexity's PerplexityBot documentation: the user agents and IP ranges to allowlist in your firewall.
• SparkToro's AI recommendation consistency research: Rand Fishkin's case for measuring mention rates instead of rankings.
• Schema.org FAQPage reference: the exact properties your JSON-LD should include.

Sources

1. Overview of OpenAI Crawlers — OpenAI, 2026. Supports: Sites that block OAI-SearchBot won't appear in ChatGPT search answers, and OAI-SearchBot is separate from GPTBot, which collects training data.
2. GEO: Generative Engine Optimization — Aggarwal et al., Princeton/Georgia Tech/IIT Delhi (KDD 2024), 2024. Supports: Adding quotations raised visibility in generative engines by about 41%, and GEO methods overall raised it by up to 40%.
3. Only 12% of AI Cited URLs Rank in Google's Top 10 for the Original Prompt — Ahrefs, 2026. Supports: Only 12% of URLs cited by ChatGPT, Gemini and Copilot rank in Google's top 10, compared with nearly 1 in 3 Perplexity citations.
4. An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied) — Ahrefs, 2025. Supports: Branded web mentions correlated with AI Overview visibility at 0.664, compared with 0.218 for backlinks.
5. LLMs.txt: Why Brands Rely On It and Why It Doesn't Work — SE Ranking, 2025. Supports: Across about 300,000 domains, sites with llms.txt were cited no more often than sites without it.
6. Analysis of Top AI Search Engines: Who Is Catching Up to ChatGPT? — SE Ranking, 2026. Supports: ChatGPT sends 74.78% of AI referral traffic and Perplexity sends 7.23%.
7. The Most Cited Domains by LLMs — Similarweb, 2026. Supports: Wikipedia and Reddit each make up roughly 12–13% of U.S. ChatGPT citations.
8. The Most-Cited Domains in AI: A 3-Month Study — Semrush, 2025. Supports: Reddit's share of ChatGPT citations fell from about 60% to about 10% between early August and mid-September 2025.
9. NEW Research: AIs are highly inconsistent when recommending brands or products — SparkToro, 2025–26. Supports: Based on 2,961 prompt runs, AI brand recommendations change almost every time, so brands should track how often they're mentioned instead of their "rank."
10. Self-promotional listicles analysis: Data from 232,000 citations — Peec AI, 2026. Supports: About 1 in 10 AI citations (roughly 11%) came from self-promotional listicles.
11. The rise and fall of FAQ schema — Search Engine Land, 2025. Supports: In August 2023, Google limited FAQ rich results to well-known government and health sites.