How to Get Your Business Cited by ChatGPT and Perplexity

Your Next Customer Might Never Open Google

A growing share of B2B and local search is shifting to AI-powered tools. When a founder asks Perplexity "who does AI consulting for small businesses?" or a procurement manager asks ChatGPT "best logistics software for mid-size fleets," the answer is a synthesized paragraph with inline citations -- not ten blue links. If your business is not one of those citations, you are invisible to that buyer.

This is not a theoretical problem. We started tracking how inbound leads found OneWave AI in late 2025, and by January 2026, a meaningful share of qualified leads were coming from AI search -- people who typed a question into Perplexity, ChatGPT, or Gemini and our site was cited in the response. No traditional Google ranking involved. No ad spend. Just content that the model chose to reference.

This post is the tactical playbook we used. No theory, no buzzwords -- just the specific things we did to go from zero AI citations to consistent inbound from ChatGPT, Perplexity, and Claude-powered search.

SEO is about getting ranked. Getting cited by AI is about being the source the model trusts enough to quote. The tactics overlap but the mindset is completely different.

How AI Search Engines Pick Sources

Before the tactics, you need to understand the mechanism. When Perplexity answers a question, it does not crawl the web in real time the way Google does. It uses a retrieval-augmented generation (RAG) pipeline: it searches an index, pulls the top-matching pages, reads them, then synthesizes an answer with citations back to those pages. ChatGPT with browsing works similarly -- it issues search queries, reads the results, and constructs a response.

This means the game is not about keyword density or backlink volume. It is about whether your page contains a clear, direct, well-structured answer to the question someone is asking. AI models are selecting for information density, specificity, and credibility signals -- not PageRank.

What Gets Cited (and What Gets Ignored)

After analyzing which of our pages got cited and which did not, and studying the research from Georgia Tech and Princeton on generative engine optimization, we identified clear patterns:

• Pages with specific numbers get cited. "We saved one client $240K per year" beats "we help clients save money" every time. AI models gravitate toward concrete claims with attached metrics because they make the synthesized answer more useful.
• Pages that directly answer a question get cited. If someone asks "what is an AI agent?" and your page has a section titled "What Is an AI Agent" with a clean two-sentence definition in the first paragraph, you are far more likely to be selected than a page that buries the answer in the fourth section.
• Pages with structured data get cited more. JSON-LD schema -- particularly FAQPage, Article, and HowTo -- gives the retrieval system metadata that helps it understand what your page is about before it reads the full content.
• Expert attribution matters. Research showed that adding expert quotations to content boosted AI citation rates by over 40 percent. A quote from your CEO with their title carries weight that anonymous copy does not.
• Freshness wins tiebreakers. When two pages answer the same question equally well, AI models tend to cite the more recently published one. Keep publish dates current and update older content.

The 7-Step Playbook We Used

Here is exactly what we did at OneWave AI to start getting cited. Every step is something a small business can execute this week.

1. Audit Your Content for "Citable Answers"

Go through your top pages and ask: if an AI model read this page, could it pull out a clean two-to-three sentence answer to a specific question? If the answer is buried in marketing fluff, the model will skip you and cite whoever gave a straighter answer.

We rewrote every service page to lead with a direct answer to the most common question about that service. Our Claude Agents page does not start with "we offer world-class solutions." It starts with what a Claude agent is and what it does for your business.

2. Add FAQ Sections With Real Questions

FAQ sections are citation magnets. When someone asks Perplexity a question and your page has that exact question as an H3 with a direct answer below it, the retrieval system treats that as a near-perfect match.

But the questions have to be real. Do not write "Why is [Your Company] the best?" Write the questions your customers actually ask you. We added FAQ sections to every case study with questions like "Can you build a real website with Claude Code?" and "How much does a custom website cost compared to hiring an agency?" -- questions people are actually typing into AI search.

3. Implement FAQPage Schema (JSON-LD)

This is the technical step most businesses skip. FAQPage structured data tells search engines (both traditional and AI-powered) that your page contains question-and-answer pairs. The implementation is straightforward:

{
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Can a non-developer build a website with AI?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. OneWave AI built their production site
                 using Claude Code with zero prior dev experience..."
      }
    }
  ]
}

Every page on our site that has FAQ content also has the corresponding FAQPage JSON-LD. It takes fifteen minutes to implement and makes your answers machine-readable.

4. Publish Case Studies With Hard Numbers

Case studies are the single best content type for AI citations. They combine specific metrics, real company names, concrete outcomes, and narrative detail -- everything AI models look for when constructing a credible answer.

Our case study on Nationwide Haul gets cited when people ask about AI for logistics or trucking companies because it includes specific numbers: how much time was saved, what tools were used, what the before and after looked like. Generic "we helped a client succeed" pages do not get cited.

5. Write for Perplexity Specifically

Perplexity is the AI search engine most likely to cite you right now because it always shows sources. Here is what we learned about how it selects them:

• It uses Bing's index as its primary retrieval layer. This is the single most overlooked fact about AI search. Perplexity and ChatGPT browsing both pull from Bing, not just Google. If you have not submitted your sitemap to Bing Webmaster Tools, do it today -- it takes five minutes and directly affects whether AI search can find you.
• It prefers pages with clear section structure. H2s and H3s that match common questions act as retrieval anchors.
• It cites pages that include data and sources. If your page references a study, includes a stat with attribution, or quotes an expert, Perplexity is more likely to include it.
• It rotates sources. Ask the same question three times and you might get three different source sets. This means your page does not need to be the single best result -- it needs to be in the top handful of relevant pages.

6. Target Long-Tail Questions, Not Keywords

Traditional SEO targets keywords: "AI consulting," "business automation." AI search optimization targets questions: "how do I automate my customer support with AI," "what does an AI consulting engagement actually look like."

The shift matters because AI models are answering full questions, not matching keywords. Your content needs to anticipate and directly answer the specific questions your ideal customer is asking. We use our actual sales calls to find these -- every question a prospect asks is a content opportunity.

7. Keep Content Fresh and Update Old Posts

AI models have a recency bias. Between two pages with similar quality answers, the one published or updated more recently tends to win. We update our highest-performing blog posts monthly -- adding new data, refreshing examples, and updating publish dates.

This is also why a blog matters more for AI citations than static service pages. Service pages are set-and-forget. Blog posts can be continuously updated to stay current, which keeps them in the AI citation pool.

Perplexity vs. ChatGPT vs. Claude vs. Gemini: How Each Cites Differently

Not all AI search engines work the same way. Here is what we have observed from tracking our own citations across all four major platforms.

The takeaway: Perplexity is the easiest to crack for small businesses because it actively crawls and cites live pages. ChatGPT browsing is similar but less consistent. Claude is harder because it draws more from training data, so being cited by other well-known sources matters more. Gemini is the one to watch -- Google's AI Overviews are replacing traditional search results for a growing number of queries, and they pull directly from Google's index, so strong traditional SEO is your best lever there. The good news is that every tactic in this playbook improves your chances across all four.

Add an llms.txt File to Your Site

This is the tactic almost nobody is talking about yet. The llms.txt standard is a proposed convention -- similar to robots.txt -- that tells AI models what your site is about and which pages to prioritize when crawling for answers. You place it at yoursite.com/llms.txt and it contains a structured summary of your site, your key pages, and what topics you are authoritative on.

Think of it as a cover letter for AI crawlers. Instead of hoping the model figures out that your case study page is more important than your privacy policy, you tell it directly. The format is simple -- a markdown-like file listing your site name, a one-line description, and links to your most important content with brief summaries.

It is early days for the standard, but the downside of adding it is zero and the upside is that AI models increasingly use it to prioritize which pages to read. If you are already doing the other tactics in this post, adding llms.txt takes ten minutes and gives AI crawlers a map to your best content.

Reviews and Third-Party Mentions Matter More Than You Think

AI models do not just read your website. They read everything -- review sites, directories, forums, social media, and other sites that mention you. When Perplexity answers "who is the best AI consultant in Florida," it is not just pulling from your homepage. It is triangulating across Google Business reviews, Clutch profiles, LinkedIn posts, and anywhere else your name appears.

This means your off-site presence is part of your AI citation strategy. Ask happy clients for Google reviews. Make sure your business profiles on directories and review platforms are complete and current. If a client mentions you on LinkedIn or in a blog post, that mention becomes training data and retrieval fodder for AI models.

Industry Review Platforms Are AI Citation Gold

Platforms like G2, Trustpilot, Capterra, and Clutch carry outsized weight in AI search results. When someone asks ChatGPT "best AI consulting firms" or Perplexity "top workflow automation tools," these platforms are almost always in the citation set. The reason is simple: AI models treat aggregated third-party reviews as high-credibility signals. A claim on your own website is marketing. The same claim validated by 50 reviews on G2 is evidence.

Here is the practical playbook for review platforms:

• Claim and complete your profiles. G2, Trustpilot, Capterra, Clutch, and Google Business Profile are the big five. An incomplete profile is worse than no profile -- it signals abandonment. Fill out every field, add screenshots, and write a description that matches the keywords you want to rank for.
• Systematize review collection. After every successful engagement, send a direct link to your G2 or Trustpilot page. Make it part of your offboarding process, not an afterthought. Five detailed reviews from real clients beat a hundred generic ones.
• Respond to every review. AI models can read review responses. A thoughtful reply to a review adds another indexed mention of your company with relevant keywords. It also signals active engagement, which platforms reward with higher visibility.
• Use review badges on your site. G2 badges, Trustpilot widgets, and Clutch verified badges create a bidirectional link between your site and the review platform. This reinforces the connection for both traditional search crawlers and AI retrieval systems.

Get Cited by Your Partners

If you have technology partners, vendor relationships, or channel partnerships, those partner pages are some of the most powerful citation sources available. When Anthropic lists you in their partner directory, or Intercom features you as a certified implementation partner, or a SaaS vendor includes you in their integrations page -- those pages carry domain authority that flows directly into AI citation rankings.

Most businesses treat partner listings as a nice-to-have. For AI citations, they are essential. A mention on a high-authority partner site tells AI models that you are not just claiming expertise -- a trusted company is vouching for you. Ask your partners to link to your site from their partner pages. Co-author blog posts or case studies with them. If you helped a partner's customer succeed, ask the partner to feature that story.

The compounding effect across all of these channels is real: a client who leaves a G2 review, is featured in a case study on your site, posts about you on LinkedIn, gets mentioned on a partner's page, and writes about you on their own company blog creates five separate signals that AI models use to validate your credibility. One mention is noise. Five mentions from different sources is a pattern that models learn to trust.

Stop Writing Marketing Copy. Start Writing Useful Content.

This is the single most important mindset shift for AI citations, and it is the one most businesses resist. AI models do not cite sales pages. They cite pages that genuinely help the person asking the question. If your content reads like a brochure -- "leverage our synergistic solutions to maximize your operational efficiency" -- no AI model will ever quote it because it does not actually say anything.

Compare that to a page that says: "We helped a five-truck logistics company automate their dispatch scheduling using Claude, cutting 12 hours per week of manual work and eliminating double-booking errors." One is marketing. The other is information. AI models cite information.

The test is simple: would your content be useful to someone who will never buy from you? If a competitor reads your blog post and learns something, that post is good enough to get cited. If it only makes sense as a pitch to prospects, it will not get cited by anything.

Every post on our blog follows this rule. Our MCP servers guide teaches people how MCP works whether they hire us or not. Our agent-building guide gives a real technical blueprint. The posts that teach are the posts that get cited -- and the people they teach often become the people who reach out.

What We Got Wrong (So You Do Not Have To)

Not everything we tried worked. Here are the mistakes we made early on:

• Writing for Google instead of for answers. Our early blog posts were keyword-stuffed and optimized for traditional ranking factors. They ranked decently on Google but never got cited by AI. The posts that get cited are the ones that give clean, direct, useful answers.
• Ignoring structured data for months. We did not add JSON-LD schema until late 2025. Once we added FAQPage and Article schema across the site, we saw a noticeable increase in AI citations within weeks. Do this first -- it is the highest-leverage technical change.
• Not tracking AI referrals. We had no idea AI search was driving traffic until we set up proper analytics. If you are using something like Clarity or GA4, look for referral traffic from perplexity.ai, chat.openai.com, and similar domains. You might already be getting cited without knowing it.

The Minimum Viable AI Citation Strategy

If you do nothing else, do these three things this week:

1. Add FAQPage JSON-LD to your top three pages. Pick the pages most likely to answer questions your customers ask. Add 3-5 real questions with direct answers. Implement the FAQPage schema.
2. Rewrite one page to lead with a direct answer. Pick your most important service page. Make the first paragraph a clean, specific answer to "what is [your service] and why does it matter?" No fluff, no brand positioning -- just the answer.
3. Publish one case study with real numbers. Name the client (with permission), state the problem, quantify the outcome. "We helped [Company] reduce [X] by [Y]% in [Z] months" is exactly the format AI models love to cite.
4. Submit your sitemap to Bing Webmaster Tools. Perplexity, ChatGPT, and Gemini all rely on Bing's index for retrieval. If Bing cannot find you, AI search cannot cite you. This takes five minutes.

That is it. Four changes, one week, and you are ahead of 95 percent of small businesses who have not started thinking about AI search at all.

This Is the Early Mover Window

Right now, most businesses are still optimizing exclusively for Google. That is like optimizing for desktop in 2010 when mobile was about to eat everything. The shift to AI search is not coming -- it is here, and the businesses that position themselves now will own those citation slots for years.

We wrote about the broader strategic shift in our post on GEO and AEO: the new rules of AI search. This post is the tactical companion -- the specific steps to execute on that strategy. If you want help implementing this for your business, we do exactly this kind of work. Or check out our success stories to see what the results look like in practice.