A few weeks ago, a prospect told me how she found me. She hadn’t searched Google or asked a colleague. She’d described her content problem to Claude, and it recommended me by name.
That conversation changed how I think about visibility. When a buyer asks ChatGPT for a shortlist, the brands it names get the first conversation, and the brands it leaves out may never hear from that buyer at all.
To show up in ChatGPT, you need three things. You need pages that answer your buyers’ exact questions, mentions on the third-party sites ChatGPT reads when it searches, and a site its crawler can access. This playbook covers all three, along with five weeks of my own AI visibility data showing which of my pages get cited and why.
How ChatGPT decides which brands to recommend
ChatGPT recommends brands from two sources. The first is what the model learned during training. The second is live web search, where ChatGPT rewrites your question into search queries, reads the pages that come back, and cites the ones it uses. For questions about current products, prices, and vendors, ChatGPT often turns to search.
ChatGPT search sometimes uses third-party search providers, and it rewrites your question into one or more targeted queries before sending them. After reading the first results, it may send more specific follow-up searches.
Here’s what that means for a SaaS brand. Say a buyer types, “We’re a 40-person startup preparing for our first SOC 2 audit. What tools should we look at?” ChatGPT might search for something closer to “best SOC 2 compliance software for startups 2026,” then read whatever ranks for it. Those results are usually listicles, comparison pages, review sites, and Reddit threads.
So you can show up in two ways. You can be named on the pages ChatGPT reads, or you can own one of those pages yourself. The strongest brands do both.
OpenAI also runs separate crawlers for each job. OAI-SearchBot surfaces websites in ChatGPT’s search features, and GPTBot collects data for training models. Sites that block OAI-SearchBot won’t be shown in ChatGPT search answers, which makes it the one crawler you can’t afford to block.
What five weeks of my own AI visibility data shows
I track my own brand in Peekaboo, an AI visibility tool, across ChatGPT, Gemini, Perplexity, Google AI Overviews, and Google AI Mode. I started with 10 general prompts a buyer might ask when looking for a B2B SaaS writer, which Peekaboo ran 1,329 times between August 26 and September 25, 2026. In late September, I added 23 prompts written the way a content lead actually asks for help. In their first week, those new prompts ran 417 times.
A quick caveat on method before the findings. Peekaboo runs prompts through each model’s API, which isn’t identical to what a logged-in user sees in the ChatGPT app, and a week of data on 23 prompts is a small sample for one small brand. Treat these as patterns worth testing on your own brand, not universal rules.
The prompts I track
Here are some of the prompts in full, with the visibility score each earned. The first three are from my original set of 10, scored over their first month, and the rest are from the 23 I added later, scored over their first week.
| Prompt | Visibility score |
| Our team is on a tight budget. Are there experienced freelance writers who offer package deals for ongoing SaaS content? | 27 |
| How do freelance writers typically charge for long-form content in 2026, especially for the tech industry? | 14 |
| I’m a content manager for a MarTech company. What kind of writing expertise should I look for to improve our inbound strategy? | 0 |
| I lead content at a Series B B2B SaaS company. We have a small in-house team and need a freelance writer who can own SEO articles without a lot of editing. Who should I look at? | 73 |
| Find me a writer who can refresh underperforming SaaS content instead of just producing new articles. | 68 |
| Our SaaS blog gets decent traffic, but leadership wants to see more pipeline from content. I need a freelancer who understands BOFU SEO and can help us create comparison, alternatives, and use-case content. | 53 |
| I manage content for a technical SaaS product, and our internal experts are too busy to write. I need someone who can interview SMEs and turn those conversations into strong articles. | 4 |
| Our CEO keeps asking why competitors appear in ChatGPT and AI Overviews more often than we do. Are there freelance content strategists who specialize in SEO and AI search visibility? | 0 |
Notice how differently the two sets are written. My first prompts read like general research questions. The newer ones read the way a content lead actually asks for help, with their situation, a constraint, and the job they need done. That’s closer to how buyers talk to AI assistants, so it’s the style I’d use for your own list.
I show up when a page answers the exact request
The new hiring prompts split sharply. Eight of the 23 scored 50 or more out of 100, and seven scored zero. Here are the highest and lowest, shortened.
| Prompt (shortened) | Visibility score |
| A content lead at a Series B company needs a writer who can own SEO articles | 73 |
| A writer who can refresh underperforming SaaS content | 68 |
| A freelance writer for bottom-of-funnel SaaS content | 65 |
| A B2B SaaS writer who understands SEO and technical products | 56 |
| A content strategist for SEO and AI search visibility | 0 |
| Someone to turn SME interviews into thought leadership and SEO content | 0 |
| A strategist who can identify content gaps and also write the articles | 0 |
The high scorers line up with pages I’ve built to sell those services. I have a dedicated page for bottom-of-funnel content, one for content refreshes, and one for B2B SaaS content writing. On the Series B prompt, ChatGPT named me in three of four runs and listed me first each time, and Perplexity named me in all four.
My first 10 prompts showed the same thing on a smaller scale. The two where I scored highest were about pricing and packages, at 27 and 14 out of 100, and I have pages that answer both, including a guide to content writer rates. Most of the others scored zero.
Articles about a service don’t get you hired for it
The zeros taught me more than the high scores. I scored zero on all three prompts that asked for a freelance content strategist, even though I’ve published an explainer on what a freelance content strategist does, and it ends with a section on working with me. I also scored zero and four on the two prompts about turning SME interviews into content, even though I’ve written a detailed guide to conducting expert interviews. None of my service pages mentions SME interviews, and none of them offers content strategy as a service in its own right.
An article about how to do something shows expertise, but it doesn’t tell an AI tool that you’re available to hire for it. If you want to be recommended for a service, give that service its own page that says clearly that you offer it, who it’s for, and what it includes.
The pages that get cited answer one question directly
Peekaboo recorded 499 citations of my site in the last week, spread across 63 URLs. These pages drew the most, not counting my homepage.
| Page | Citations | Cited by |
| Best SEO content writers (roundup) | 47 | Google AI Mode, AI Overviews, Perplexity |
| Best content writers for hire (roundup) | 40 | Google AI Mode, AI Overviews, Perplexity |
| B2B SaaS content writer (service page) | 40 | ChatGPT, Google AI Mode, AI Overviews, Perplexity |
| Best B2B SaaS content writers (roundup) | 37 | AI Overviews, Perplexity |
| Best SEO content writing services (roundup) | 27 | Google AI Mode, AI Overviews, Perplexity |
Four of the five are roundups of writers or writing services that I publish, and I include myself in each alongside writers and services I’d recommend. When a buyer asks an AI tool who to hire, a well-sourced list is exactly what it wants to read. If you publish lists like these, be fair to the other entries and clear about your own.
Each of these pages answers one question, and the answer sits near the top. In my first month of tracking, the most cited page was my rates guide, which opens with a “Quick answer” sentence giving the typical price range. When Gemini cited that page, its link jumped straight to that sentence, which shows exactly which passage the model pulled.
My site now leads, but LinkedIn and marketplaces are close behind
In my first month, with 10 general prompts, LinkedIn (630 citations) and Reddit (423) were both cited more often than my own site (351). Once I added hiring prompts, my site moved to the top, with 499 citations in a week. LinkedIn was nearly level at 492, and freelance marketplaces like Upwork (393), Freelancer, and Toptal were close behind.
The sources shift with the question. General questions pull in LinkedIn and Reddit discussions, while hiring questions pull in marketplaces and other writers’ portfolio sites. Your website is one source among many, so it’s worth knowing which sources AI tools read for your buyers’ questions.
When ChatGPT mentions me, it mentions me early
In the last week, ChatGPT gave me a visibility score of 20 out of 100, and when it mentioned me, my average position in its answer was 1.4, so I was usually the first name listed. Perplexity scored highest, at 24, with an average position of 1.9, and Google AI Overviews placed me first on average, at 1.1, with a score of 21. Gemini was the outlier, with a score of 11 and an average position of 5.4, so it tended to list me further down.
My overall score rose to 17 in the last week, but that mostly reflects the new prompts, which match my services more closely than the first 10 did. The useful comparison is prompt by prompt, not the headline number.
The takeaway for your own brand is that appearing less often doesn’t mean appearing weakly. Once an answer engine decides you fit a question, it tends to lead with you, which is another reason to build pages around specific requests instead of broad topics.
Step 1: Map the prompts your buyers ask
Start with the questions, because every other step depends on them. Aim for 20 to 30 prompts, written in your buyers’ words rather than your positioning deck’s. The prompts I track, shown earlier, are a good model for the wording.
You’ll find the raw language in sales call recordings, G2 and Capterra reviews, support tickets, and Google Search Console. Search Console is underrated here. On my own content marketing agencies page, it showed about 330 queries over 28 days, and many read like prompts, with full sentences, context about the company, and a specific constraint.
Group the prompts by buying stage, because each stage calls for a different page. The examples below are for a hypothetical SOC 2 compliance tool.
| Buying stage | Example prompt | Page that answers it |
| Problem | How long does a first SOC 2 audit take for a startup? | Guide or explainer |
| Shortlist | Best SOC 2 compliance software for a 40-person startup | “Best X for Y” page, third-party listicles |
| Comparison | [Your product] vs [competitor] for startups | Comparison page |
| Alternatives | Cheaper alternatives to [competitor] | Alternatives page |
| Pricing | How much does SOC 2 compliance software cost? | Pricing page or pricing guide |
Once you have the list, run each prompt in ChatGPT and note who gets named and which pages get cited. That gives you a baseline, and it shows you which third-party pages you need to get onto.
Step 2: Publish pages that answer those prompts directly
My data points to one rule. If you want to be recommended for a question, publish a page that answers it. Bottom-of-funnel pages are the place to start, because they match the shortlist, comparison, and pricing prompts that come right before a purchase.
These are the page types I’d build first.
- A dedicated page for each use case you want to be recommended for. In my data, articles about a service didn’t get me recommended for it, while pages that offered the service did. For a SaaS product, that means a page for each use case or product line, not just blog posts about the topic.
- Pricing pages with real numbers. “Contact sales” gives ChatGPT nothing to quote. Even a starting price or a typical range gives the model an answer.
- Comparison pages. “[Your product] vs [competitor]” pages give ChatGPT a side-by-side answer it can cite when buyers compare you directly.
- Alternatives pages. Buyers who are unhappy with a competitor ask for alternatives, and a fair, well-sourced page puts you in that answer.
- “Best X for Y” and use-case pages. Prompts usually include a constraint, like team size, industry, or budget. A page built around that constraint matches the rewritten search query.
Structure each page so the answer is easy to lift. Put a 40- to 60-word answer directly under each heading, name products and features explicitly instead of saying “our solution,” use tables for comparisons, and date anything that changes, like pricing. The process for finding bottom-of-funnel keywords will surface most of these topics, and the guide to optimizing content for AI search goes deeper on page-level structure.
Step 3: Get onto the lists ChatGPT reads
When a buyer asks for a shortlist, ChatGPT often builds its answer from existing shortlists. If your brand appears on the listicles and comparison pages that rank for your category, you have a much better chance of appearing in the answer.
I’ve seen how much a single list placement can be worth. One agency I included in my roundup of SaaS SEO agencies told me a buyer found them on that list and reached out. That deal was worth $36K in annual recurring revenue, and it closed within two weeks.
Here’s how to find and approach those lists.
- Take your shortlist and comparison prompts from Step 1 and turn them into the search queries ChatGPT would likely run, such as “best [category] for [use case] 2026.”
- Search those queries in Google and Bing, and list every third-party page in the top 10.
- Check which of those pages ChatGPT cites when you run the original prompts.
- Contact the authors with specific, useful facts, like your current pricing, the use case you’re strongest for, and a customer result they can verify.
- Keep your review profiles on G2, Capterra, or the review sites in your category current, and ask customers to mention their use case in reviews.
Publishers get plenty of generic pitches, so the specifics do the persuading. A short note with a current price, a clear “best for,” and one verifiable result is easy for an author to add, and it gives both the author and ChatGPT something concrete to say about you.
Step 4: Show up on LinkedIn, Reddit, and YouTube
In my data, LinkedIn was cited almost as often as my own site, and Reddit and freelance marketplaces weren’t far behind. If those platforms show up in answers for your prompts too, your brand needs a presence there that says something useful.
On LinkedIn, that usually means founders and subject-matter experts posting about the problems your buyers ask AI tools about, with specific numbers and opinions. On Reddit, it means answering questions in the communities your buyers use, as yourself, with your affiliation disclosed. Reddit users are quick to spot a sales pitch, and a helpful answer that mentions your product in context does more than a promotional post.
YouTube works well for product demos, walkthroughs, and comparisons. Give each video a descriptive title and an accurate transcript, since the text is what AI tools can read.
Step 5: Describe your product the same way everywhere
ChatGPT has to decide what your product is before it can recommend it. If your homepage calls you a “compliance automation platform,” G2 lists you as “GRC software,” and your LinkedIn page says “security for startups,” you’re making that job harder.
Write one clear description and use it everywhere buyers and AI tools look. It should name your category, your ideal customer, and two or three differentiators that are true and specific. Then use the same wording on your homepage, pricing page, review profiles, LinkedIn, and partner listings.
Consistency helps the model connect your brand to a category and a use case. When a buyer’s prompt includes that use case, you’re easier to match.
Step 6: Make sure ChatGPT can read your site
The technical checks are quick, and a failure here can cancel out everything above.
Allow OAI-SearchBot in robots.txt. Sites that opt out of OAI-SearchBot won’t be shown in ChatGPT search answers. Blocking GPTBot is a separate decision, and you can allow OAI-SearchBot while disallowing GPTBot if you don’t want your content used for training.
Check that your content is in the HTML. If your pricing or product details only appear after JavaScript runs, crawlers may miss them. A quick test is to view the page source, or fetch the page with a simple command-line request, and confirm the text is there. When I ran that test on my rates guide, the raw HTML returned the full article, about 7,700 words, before any script ran.
Keep key facts out of images. Prices, feature lists, and comparison tables should be text. A pricing table saved as an image gives a crawler nothing to quote.
How to measure whether you show up in ChatGPT
You’ll need three kinds of measurement, because no single tool shows the whole picture.
Prompt tracking. AI visibility tools run your prompts on a schedule and record when you’re mentioned, where you appear, and which pages get cited. Peekaboo is what I use. Semrush includes prompt tracking across ChatGPT, Perplexity, and Gemini from its $199-a-month Starter plan, which covers 50 prompts a day as of September 2026. The roundup of AI visibility tools compares more options.
Referral traffic. ChatGPT adds a tracking parameter, utm_source=chatgpt.com, to many of the links it cites. In my Peekaboo data, most of ChatGPT’s links to my service pages and homepage carried that parameter. In GA4, filter sessions by source to see visits from chatgpt.com, and compare which landing pages they reach.
Self-reported attribution. Referral data misses buyers who read an answer and then type your URL directly, which is how the Claude prospect reached me. Add a free-text “How did you hear about us?” field to your demo form, and tag answers that mention ChatGPT, Claude, Perplexity, or “AI” in your CRM. Those tags let you connect AI visibility to pipeline, which is the number your leadership will ask about.
A 90-day plan to show up in ChatGPT
Three months is enough to publish the core pages, start outreach, and see early movement in your tracking.
| Timing | What to do |
| Weeks 1–2 | Map 20–30 buyer prompts, record a baseline in ChatGPT and your tracking tool, and fix robots.txt and rendering issues |
| Weeks 3–8 | Publish or refresh 6–10 pages that answer your shortlist, comparison, alternatives, and pricing prompts |
| Weeks 4–12 | Pitch the third-party lists ChatGPT cites, update review profiles, and start posting on LinkedIn and Reddit |
| Weeks 6–12 | Add the self-reported attribution field and tag AI-sourced leads in your CRM |
| Week 12 | Re-run your prompts, compare against the baseline, and choose the next set of pages |
Pages can start drawing citations soon after search engines index them. Brand mentions on third-party sites take longer, so expect your tracking to move unevenly, with individual prompts improving before your overall score does.
This is the same structure I use in the Pipeline Sprint, a 90-day engagement where I map buyer-intent searches and prompts, then write the 6–10 pages built to answer them on Google and in AI answers.
Frequently asked questions about showing up in ChatGPT
How do I get my business to show up in ChatGPT?
Publish pages that directly answer the questions your buyers ask, get your brand onto the third-party listicles and review sites ChatGPT reads during search, and make sure OpenAI’s OAI-SearchBot can crawl your site. Then track your prompts to see which pages ChatGPT cites and where you still need coverage.
How long does it take to show up in ChatGPT?
It depends on which source ChatGPT uses. When it searches the web, a new or updated page can be cited once search providers have indexed it. Recommendations that come from the model’s training data only change when a new model is trained, which can take months. Third-party mentions usually take the longest to build.
Does ranking on Google help you show up in ChatGPT?
It helps indirectly. ChatGPT search sends rewritten queries to third-party search providers and reads the pages that come back, so pages that rank well for those queries are more likely to be read and cited. Strong rankings on Google and Bing make that more likely, but they don’t guarantee a mention.
What is ChatGPT SEO?
ChatGPT SEO is the practice of improving how often ChatGPT mentions, recommends, or cites your brand. It builds on traditional SEO, since ChatGPT relies on web search for many answers, but it also covers third-party mentions, consistent brand descriptions, answer-first page structure, and prompt tracking.
Should I block GPTBot?
Blocking GPTBot stops OpenAI from using your content to train its models, but it doesn’t remove you from ChatGPT search. Blocking OAI-SearchBot does. If you want to appear in ChatGPT search answers, allow OAI-SearchBot either way.
Why does ChatGPT recommend my competitors instead of me?
It’s usually because your competitors appear on the pages ChatGPT reads for that prompt and you don’t, or because they have a page that answers the question directly. Run the prompt, check which pages ChatGPT cites, and look for the gap. It’s often a missing comparison page, a missing listicle placement, or pricing ChatGPT can’t find.
