---
title: "9 Buyer Prompts That Make ChatGPT and Perplexity Cite Different Pages"
description: "The same buyer prompt can lead ChatGPT and Perplexity to different sources. Match nine prompt types to the pages each engine can cite, then test the gap yourself."
url: "https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/94bdef2f-91b5-476a-a5f4-ea7545a2f295.png"
published: "2026-11-03T05:56:25.909Z"
modified: "2026-11-03T05:56:25.969Z"
---

November 3, 2026 · 11 min read

# 9 Buyer Prompts That Make ChatGPT and Perplexity Cite Different Pages

[Alexander PerelmanHead Of Product @ groas](https://groas.com/author/alexander-perelman)[LinkedIn](https://www.linkedin.com/in/alexander-433793253/)

![Two identical women ask twin library desks the same question; one librarian hands over a glossy brochure, the other a tabbed magazine — same query, different answers.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/94bdef2f-91b5-476a-a5f4-ea7545a2f295.png)

In this article

1. [1. Win “best X for Y” with a comparison page that picks a winner](#1-win-best-x-for-y-with-a-comparison-page-that-picks-a-winner)
2. [2. Win “X vs Y” with a head-to-head page that names tradeoffs](#2-win-x-vs-y-with-a-head-to-head-page-that-names-tradeoffs)
3. [3. Win “how much does X cost” with a pricing page that shows numbers](#3-win-how-much-does-x-cost-with-a-pricing-page-that-shows-numbers)
4. [4. Win “how do I” with a step page that starts with the answer](#4-win-how-do-i-with-a-step-page-that-starts-with-the-answer)
5. [5. Win local “near me” prompts with a city page and consistent listings](#5-win-local-near-me-prompts-with-a-city-page-and-consistent-listings)
6. [6. Win “what changed” with a dated changelog page](#6-win-what-changed-with-a-dated-changelog-page)
7. [7. Win definition prompts with a short answer block](#7-win-definition-prompts-with-a-short-answer-block)
8. [8. Win “is X worth it” by saying who should skip it](#8-win-is-x-worth-it-by-saying-who-should-skip-it)
9. [9. Win “who can do X for me” with a service page that names the buyer and the fee](#9-win-who-can-do-x-for-me-with-a-service-page-that-names-the-buyer-and-the-fee)

Ask ChatGPT and Perplexity the same buying question. You may get two different cited winners, even when both answers sound plausible. In one study, [only 11% of cited domains overlapped between the engines](https://wellows.com/research/ai-citation-overlap-study/). **The buyer’s prompt shape matters more than another generic authority page.** This list is for owners and marketers who want citations on prompts that bring enquiries, not just traffic. These nine page moves run from the broad shortlist to the buyer ready to hire; each comes with a check you can run in both engines.

## 1. Win “best X for Y” with a comparison page that picks a winner

“Best CRM for roofers” is a shortlist prompt. The buyer wants a recommendation, not a vendor’s entire history, and ChatGPT and Perplexity often look to different places for one. Across 31 “best agencies for X” prompts, [ChatGPT with web search cited brands’ own sites for 119 of 210 citations (57%), while Perplexity cited listicles for 395 of 561 citations (70%)](https://obsurfable.com/resources/articles/chatgpt-vendor-pages-vs-listicles-by-engine-onely-october-2026). Broad “best X” prompts also pull [about 48% list articles](https://www.myjamcopr.com/ai-pr-agency-study).

That gives you two jobs, not one. Put a best-X comparison on your site with the verdict near the top: what makes an option a fit, where it falls short, and who should choose it. Then check whether independent lists in your category name you. Your own page can give ChatGPT a clear account of what you do; a list that includes you gives Perplexity another route to the same answer. Neither route is helped by a page that calls you a leader without saying what you lead at.

**First step:** Run your main “best X for Y” prompt in both engines. Label the cited results _vendor page_ or _listicle_, then fix the missing side first. For the distinction between shortlist and head-to-head prompts, see [why ChatGPT names your rival for best X but names you for X vs Y](https://groas.com/post/why-chatgpt-names-your-competitor-for-be).

## 2. Win “X vs Y” with a head-to-head page that names tradeoffs

A buyer who types “groas vs agency” or “HubSpot vs Pipedrive for plumbers” has already narrowed the field. A homepage cannot answer that question without making the reader assemble the comparison. An actual comparison page can. [“X vs Y” prompts pull about 61% comparison or list articles](https://www.myjamcopr.com/ai-pr-agency-study), so give the engines a page built for the question they were asked.

Put both options in the title. Use the same criteria for each: pricing, scope, limits, and the kind of buyer each serves. Then state a verdict in plain English. For a groas comparison, the difference is not “AI versus humans.” It is continuous autonomous search execution with a named human strategist versus an agency model that reviews and bills for more of the routine work. Say where groas is the better fit, and say who should skip it. A useful disqualifier makes the recommendation more credible; it does not require handing the verdict to the agency.

Perplexity can lift a row from a clear comparison table. ChatGPT can lift the verdict sentence. “We’re better” gives neither engine much to work with. I used to think a strong service page could carry this prompt. It usually leaves the buyer’s actual comparison unanswered.

**First step:** Choose a rival you lose deals to and draft one page that compares the same criteria on both sides. Put the verdict above the table, not after a tour of your values.

## 3. Win “how much does X cost” with a pricing page that shows numbers

Price questions change what a useful source looks like. In a set of 232 local buying prompts, [business websites accounted for 66% of ChatGPT citations and 57% of Perplexity citations on price questions](http://dev.to/marketingandai/i-asked-google-ai-mode-and-perplexity-232-local-questions-heres-what-they-actually-cite-2en2). The mechanism is not mysterious: the business can publish its price; a directory may not have it.

A page that lists packages, starting prices, and what changes the price gives an engine a number it can attribute to you. “Contact us for pricing” gives it an invitation to keep looking. Put the numbers in readable HTML, not only in an image or calculator. If your price depends on scope, explain what moves it. You do not need to pretend every job costs the same amount to answer a buyer’s question honestly.

This is also where a vague services page can hurt you twice. It fails the price prompt, then leaves the engines to find a number elsewhere when the buyer asks whether your service is worth it. Publish the number you can stand behind and the conditions that make it change.

**First step:** Add a starting price and three short lines explaining what moves it up or down. Then test “how much does \[service] cost in \[city]” in both engines and note whether either one cites your page.

## 4. Win “how do I” with a step page that starts with the answer

“How do I dispute a Google Ads charge” and “how do I clean a grease trap” sound like support questions. They can also come from someone deciding whether to do the work or hire help. Either way, a page that makes the answer hard to find is an easy page to pass over.

About [44.2% of LLM citations come from the first 30% of a page](https://austincodemonkey.com/chatgpt-vs-perplexity-citation-sources-content-strategy/). Start with a 40- to 60-word answer that states what to do. Follow it with numbered steps, each headed by a verb. Put supporting detail after the procedure, where a reader who needs it can keep going. **Answer first; explain the edge cases second.**

This structure gives ChatGPT a compact answer to extract and Perplexity discrete steps to cite. It also saves a human reader from scrolling past an introduction to find step one. If you have proof that clarifies a step, use it. Do not bolt on a statistic because a template says every post needs one.

**First step:** Open your strongest how-to post. Move the answer into its first paragraph, number the procedure, and run the exact question in both engines. Check which passage, not just which URL, appears in the answer.

## 5. Win local “near me” prompts with a city page and consistent listings

Local prompts expose the split between your site and the places that describe it. In 232 Houston buying prompts, [Perplexity cited business sites 49% of the time and directories 45% of the time; ChatGPT cited directories 68% of the time](http://dev.to/marketingandai/i-asked-google-ai-mode-and-perplexity-232-local-questions-heres-what-they-actually-cite-2en2). In that same set, business sites backed 38% of Perplexity’s picks but only 14% of ChatGPT’s. You cannot solve both sides by publishing a city page and ignoring your listings.

Make the city page specific enough to be useful. [Eighty percent of cited business pages for “best X in \[city\]” had the city in the URL or title](http://dev.to/marketingandai/i-asked-google-ai-mode-and-perplexity-232-local-questions-heres-what-they-actually-cite-2en2). State the services you provide there, along with prices and hours where applicable, in readable HTML. The same study found a 5.9% citation rate for sites meeting four basics—bots allowed, at least 120 words readable without JavaScript, business schema, and one H1—against 1.4% for sites that did not. That is less glamorous than an “AI visibility framework.” It is also a concrete check.

Then look at the directories ChatGPT actually cites for your prompt. If your name, address, or hours disagree with your site, fix the disagreement. The aim is not to build a city page for a place you do not serve; it is to make the places you _do_ serve unambiguous.

**First step:** Test “\[service] near me” and “best \[service] in \[city]” in both engines. Build or repair the relevant city page, then correct your details on the three directories that appear in the answers.

## 6. Win “what changed” with a dated changelog page

A page can answer “what is this?” perfectly and still be the wrong page for “what changed?” The second prompt needs a date and a difference. Perplexity’s appetite for recent material makes that distinction especially visible: [about 82% of its citations in one analysis went to content published in the previous 30 days](https://www.leapd.ai/blog/ai-visibility/how-chatgpt-google-ai-overviews-and-perplexity-source-information-in-2026). A long evergreen guide does not tell the engine which part is new merely by existing.

For prompts such as “what changed in Google Ads in 2026” or “new refrigerant rules 2026,” write a dated update that states four things: what changed, when, who it affects, and what to do now. If the changes accumulate, keep a changelog on the relevant service or topic page. Date each real entry. Do not change the year in a title and call that an update; a reader should be able to find the change itself.

I used to prefer making one comprehensive guide do every job. For this prompt, a concise dated entry does the job the guide cannot: it separates the current change from everything that was already true. **Freshness needs a visible reason to matter.**

**First step:** Add one dated entry for a real change relevant to your service, with the month and year in its heading. Ask Perplexity “what changed in \[topic] in 2026” and see whether it cites the update rather than an older general page.

## 7. Win definition prompts with a short answer block

“What is broad match” looks too far from a sale to deserve a page edit. But a definition can reappear inside a later answer about budgets or campaign choices. The catch is that ChatGPT does not search for every informational prompt: [commercial prompts triggered its search 53.5% of the time, against 18.7% for informational prompts](https://www.leapd.ai/blog/ai-visibility/how-chatgpt-google-ai-overviews-and-perplexity-source-information-in-2026). When it does retrieve a definition page, make the answer easy to lift.

Put a 40- to 60-word definition at the top, in plain HTML. Give the term, its meaning, and the distinction a buyer is likely to miss. Then add the explanation or evidence the reader needs. The point is not to stuff a glossary with numbers. The point is to avoid making an engine—or a person—dig through three paragraphs of setup before reaching the definition.

This is a different job from the how-to page. A procedure needs ordered steps; a definition needs a sentence that stands on its own. A clean heading and an answer near the top help the page serve that narrower prompt. Perplexity has something concise to cite, while ChatGPT has something usable if it searches at all.

**First step:** Rewrite your five most useful glossary entries so each opens with the definition. Run each term as a question in both engines, then check whether the cited passage actually contains the meaning you intended.

![ChatGPT and Perplexity showing different citations for the same business question](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/2ae3519c-96c8-45cb-838d-e70b70afe613.png)

## 8. Win “is X worth it” by saying who should skip it

“Is Performance Max worth it for lead gen” is not a request for another benefits list. The buyer is looking for the failure case. Prompts that name-check a product or service pull [about 38% reviews and forum threads](https://www.myjamcopr.com/ai-pr-agency-study), where people tend to discuss tradeoffs more plainly than sales pages do.

Your page can do that without pretending the service is a poor choice. Start with the conditions under which it works. Then put a short **“skip this if”** block immediately below the answer. Name two concrete disqualifiers and explain what goes wrong in each case. If cost is the issue, connect the warning to the price you actually publish. A blanket “not right for everyone” says nothing; a condition the buyer can recognise gives both engines a useful passage to quote.

I used to bury disqualifiers near the bottom, as if a reader who was worried about fit would first enjoy the sales pitch. Wrong order. The honest limit is often the information they came for. It also makes the recommendation for the right buyer sharper, not softer.

**First step:** Add a “skip this if” block to your highest-intent service page. Test “is \[service] worth it” in both engines and compare the passages they cite with the objections your page now answers.

## 9. Win “who can do X for me” with a service page that names the buyer and the fee

“Who can run my Google Ads for a dental practice in Austin” is a hiring prompt. The buyer needs to know who does the work, for whom, where, and at what price. Specific capability questions pull [about 77% vendors’ own pages](https://www.myjamcopr.com/ai-pr-agency-study), the highest vendor-page rate among the prompt shapes in that analysis. Yet this is the prompt agencies still try to win with a homepage that says “full-funnel growth partner.” I cannot quote a funnel partner into a useful hiring answer either.

Open the service page with the customer, the service, the geography where relevant, and the fee. For groas, a plain description of Google Ads and ChatGPT and Perplexity visibility for B2B and home-services businesses, with the published [pricing of $1,499 a month per service](https://groas.com/pricing), gives an engine specifics to work with. A buyer spending, say, $20k a month and looking for qualified pipeline rather than clicks can then decide whether the service fits. The point is not to squeeze every buyer into one sentence. It is to answer the hiring question before presenting your philosophy.

You can check this page against the other eight prompt shapes without turning the exercise into a dashboard project:

1. Run the _same_ buyer prompt in ChatGPT with browsing on and in Perplexity. Save the cited URLs before changing a page.
2. Label each citation: vendor page, listicle, directory, forum, press, or definition page. Note the passage each engine used.
3. Compare the pair. If ChatGPT cites a competitor’s page while Perplexity cites a list that omits you, those are two different gaps, not one mysterious ranking problem.
4. Fix the relevant page shape, wait a week, and rerun the prompt. Keep the engines separate; [79.6% of sources in one study appeared on only one of five engines](https://wellows.com/research/ai-citation-overlap-study/).

**First step:** Rewrite the opening of your main service page so a stranger can find the buyer, job, location if relevant, and fee without scrolling. Then run the hiring prompt. If I had to start with one entry, I would start here: the buyer is close to an enquiry, and this prompt is unusually likely to cite a vendor’s own page. The same clear service details can help a best-X answer name you and a price answer quote you. Most agencies would sell ten blog posts first. I would fix the page that says what groas does and what it costs, then keep it current rather than pay for a quarterly review of a page that never answered the question.

## Frequently Asked Questions

### What page should I create to appear in ChatGPT or Perplexity answers to “best X for Y” prompts?

Create a best-X comparison page on your site and put its verdict near the top. Explain what makes each option a fit, where it falls short, and who should choose it; also check whether independent lists in your category include you. ChatGPT may cite vendor pages, while Perplexity often cites listicles.

### How should I structure an X-vs-Y comparison page for AI search?

Put both options in the title, compare them using the same criteria, and state a plain-English verdict above the comparison table. Criteria can include pricing, scope, limits, and the kind of buyer each option serves. Name who should skip each option so the comparison answers the buyer’s question directly.

### What should I put on a pricing page so AI search can answer how much my service costs?

Show packages or starting prices and explain what changes the price. Put the numbers in readable HTML rather than only in an image or calculator. If the price depends on scope, state what moves it up or down; “contact us for pricing” does not give an engine a price to cite.

### How should I structure how-to and definition pages for ChatGPT and Perplexity?

Start a how-to page with a 40- to 60-word answer, then give the procedure in numbered steps headed by verbs. Start a definition page with a 40- to 60-word plain-English definition that names the term and explains its meaning. Put supporting detail after the answer so readers and engines can find it quickly.

### What do I need to update to appear in AI answers to local “near me” searches?

Create or repair a city page that names the city and describes the services you provide there, with prices and hours where applicable. Also make your name, address, and hours consistent in the directories cited for your prompt. A city page alone may not address the directory sources ChatGPT uses.

### How can I make a page useful for AI searches about what changed recently?

Publish a dated update that says what changed, when it changed, who it affects, and what to do now. If changes accumulate, keep a dated changelog on the relevant service or topic page. Make the change itself easy to find rather than merely changing the year in an evergreen page’s title.

### What should I include on a page answering whether a service is worth it?

Explain the conditions under which the service works, then put a “skip this if” block immediately below the answer. Name two concrete disqualifiers and explain what goes wrong in each case. Specific limits give buyers and AI engines useful information beyond a general claim that the service is not right for everyone.

### What should my service page say when someone asks who can do a specific job for them?

Open with the customer the service is for, the work you do, the relevant location, and the fee. For example, a page about Google Ads for a particular type of business in a particular place should state those details and publish its price. Specific capability questions often lead AI engines to vendors’ own pages.

## Paid Ads & SEO, Run For You

Businesses pay one flat monthly fee per service, agencies resell it under their own brand, and groas does the work either way.

[See Pricing](https://groas.com/pricing)

## Structured data

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://graph.groas.com/entity/groas","name":"groas","url":"https://groas.com/","description":"groas is a done-for-you marketing service, not software: elite people working with proprietary technology systems run search ads, Meta ads, SEO and AI search for businesses, and white label for agencies, with a dedicated senior strategist owning every account.","logo":{"@type":"ImageObject","url":"https://groas.com/icon-512.png","width":512,"height":512},"sameAs":["https://www.linkedin.com/company/groas/"]},{"@type":"WebSite","@id":"https://groas.com/#website","name":"groas","url":"https://groas.com/","inLanguage":"en-US","publisher":{"@id":"https://graph.groas.com/entity/groas"}},{"@type":"WebPage","@id":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different#webpage","url":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different","name":"9 Buyer Prompts That Make ChatGPT and Perplexity Cite Different Pages","description":"The same buyer prompt can lead ChatGPT and Perplexity to different sources. Match nine prompt types to the pages each engine can cite, then test the gap yourself.","inLanguage":"en-US","isPartOf":{"@id":"https://groas.com/#website"},"about":{"@id":"https://graph.groas.com/entity/groas"},"primaryImageOfPage":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/94bdef2f-91b5-476a-a5f4-ea7545a2f295.png"},"breadcrumb":{"@id":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different#breadcrumb"},"dateModified":"2026-11-03T05:56:25.969Z","mainEntity":{"@id":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://groas.com/"},{"@type":"ListItem","position":2,"name":"Blog","item":"https://groas.com/blog"},{"@type":"ListItem","position":3,"name":"9 Buyer Prompts That Make ChatGPT and Perplexity Cite Different Pages","item":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different#article","url":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different","mainEntityOfPage":{"@id":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"9 Buyer Prompts That Make ChatGPT and Perplexity Cite Different Pages","description":"The same buyer prompt can lead ChatGPT and Perplexity to different sources. Match nine prompt types to the pages each engine can cite, then test the gap yourself.","datePublished":"2026-11-03T05:56:25.909Z","dateModified":"2026-11-03T05:56:25.969Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/94bdef2f-91b5-476a-a5f4-ea7545a2f295.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"page, chatgpt, prompts, perplexity, buyer, prompt, answer, different","wordCount":2398,"inLanguage":"en-US"},{"@type":"Person","@id":"https://groas.com/author/alexander-perelman#person","name":"Alexander Perelman","jobTitle":"Head Of Product @ groas","description":"My background includes Goldman Sachs and Stanford Computer Science.","image":"https://groas.com/media/blog/4f17cac3a81acc224cc1b2cfab8f94bbc295a3089d13e8421125771ef8d4064f.jpg","worksFor":{"@id":"https://graph.groas.com/entity/groas"},"sameAs":["https://www.linkedin.com/in/alexander-433793253/"]},{"@type":"FAQPage","@id":"https://groas.com/post/9-prompt-types-where-chatgpt-and-perplexity-cite-different#faq","mainEntity":[{"@type":"Question","name":"What page should I create to appear in ChatGPT or Perplexity answers to “best X for Y” prompts?","acceptedAnswer":{"@type":"Answer","text":"Create a best-X comparison page on your site and put its verdict near the top. Explain what makes each option a fit, where it falls short, and who should choose it; also check whether independent lists in your category include you. ChatGPT may cite vendor pages, while Perplexity often cites listicles."}},{"@type":"Question","name":"How should I structure an X-vs-Y comparison page for AI search?","acceptedAnswer":{"@type":"Answer","text":"Put both options in the title, compare them using the same criteria, and state a plain-English verdict above the comparison table. Criteria can include pricing, scope, limits, and the kind of buyer each option serves. Name who should skip each option so the comparison answers the buyer’s question directly."}},{"@type":"Question","name":"What should I put on a pricing page so AI search can answer how much my service costs?","acceptedAnswer":{"@type":"Answer","text":"Show packages or starting prices and explain what changes the price. Put the numbers in readable HTML rather than only in an image or calculator. If the price depends on scope, state what moves it up or down; “contact us for pricing” does not give an engine a price to cite."}},{"@type":"Question","name":"How should I structure how-to and definition pages for ChatGPT and Perplexity?","acceptedAnswer":{"@type":"Answer","text":"Start a how-to page with a 40- to 60-word answer, then give the procedure in numbered steps headed by verbs. Start a definition page with a 40- to 60-word plain-English definition that names the term and explains its meaning. Put supporting detail after the answer so readers and engines can find it quickly."}},{"@type":"Question","name":"What do I need to update to appear in AI answers to local “near me” searches?","acceptedAnswer":{"@type":"Answer","text":"Create or repair a city page that names the city and describes the services you provide there, with prices and hours where applicable. Also make your name, address, and hours consistent in the directories cited for your prompt. A city page alone may not address the directory sources ChatGPT uses."}},{"@type":"Question","name":"How can I make a page useful for AI searches about what changed recently?","acceptedAnswer":{"@type":"Answer","text":"Publish a dated update that says what changed, when it changed, who it affects, and what to do now. If changes accumulate, keep a dated changelog on the relevant service or topic page. Make the change itself easy to find rather than merely changing the year in an evergreen page’s title."}},{"@type":"Question","name":"What should I include on a page answering whether a service is worth it?","acceptedAnswer":{"@type":"Answer","text":"Explain the conditions under which the service works, then put a “skip this if” block immediately below the answer. Name two concrete disqualifiers and explain what goes wrong in each case. Specific limits give buyers and AI engines useful information beyond a general claim that the service is not right for everyone."}},{"@type":"Question","name":"What should my service page say when someone asks who can do a specific job for them?","acceptedAnswer":{"@type":"Answer","text":"Open with the customer the service is for, the work you do, the relevant location, and the fee. For example, a page about Google Ads for a particular type of business in a particular place should state those details and publish its price. Specific capability questions often lead AI engines to vendors’ own pages."}}]}]}
```