---
title: "To the Marketing Manager Whose CEO Sent a Competitor’s ChatGPT Screenshot"
description: "A competitor appeared in your CEO’s ChatGPT screenshot. Before buying a monitoring dashboard, spend seven days finding which buyer prompts miss you and why."
url: "https://groas.com/post/an-open-letter-to-the-marketing-manager"
image: "https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/8c81f354-e1f8-45f7-ad04-ef58a9f453bf.png"
published: "2026-10-18T05:25:52.882Z"
modified: "2026-10-18T05:25:52.946Z"
---

October 18, 2026 · 9 min read

# To the Marketing Manager Whose CEO Sent a Competitor’s ChatGPT Screenshot

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

![A tiny marketer calmly measures a smartphone that crashed into a model town while a vendor's stall of shiny gauges goes ignored.](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/8c81f354-e1f8-45f7-ad04-ef58a9f453bf.png)

In this article

1. [First, find out whether buyers are asking that question](#first-find-out-whether-buyers-are-asking-that-question)
2. [Do not pay for the meter before you find the miss](#do-not-pay-for-the-meter-before-you-find-the-miss)
3. [Your seven days: build a prompt log, then diagnose the misses](#your-seven-days-build-a-prompt-log-then-diagnose-the-misses)
4. [Tell your CEO what you found, not what a rank promises](#tell-your-ceo-what-you-found-not-what-a-rank-promises)
5. [Keep the sheet if you can ship the fixes; get help if you cannot](#keep-the-sheet-if-you-can-ship-the-fixes-get-help-if-you-cannot)

If you are the marketing manager whose CEO just sent a ChatGPT screenshot with a competitor in it, do not buy a visibility dashboard this week. That screenshot shows one buyer question, one answer, and one uncomfortable omission. You need a week of finding which buyer prompts leave you out and why before you pay to monitor the omission.

I know the message under the screenshot, too: _why not us_. I used to manage Google Ads by hand for brands where one screenshot could ruin a week. That taught me to take the question seriously without treating the first thing someone offers to sell you as the answer. **Start with crawl access and content gaps.** You can check both before you buy a tool.

## First, find out whether buyers are asking that question

That screenshot feels personal. It is not a verdict on your brand. The answer depends on the question asked and the material the system can find and use. I used to tell clients to search their own brand name to check visibility. I was wrong. Nobody buys that way. Test with [buyer prompts, not your brand name](https://mediaofficers.com/insights/how-to-get-cited-by-chatgpt-perplexity-ai-overviews). Your CEO’s prompt belongs in the test, but it cannot be the whole test.

Most prompts do not produce a neat vendor shortlist. One analysis of thousands of ChatGPT prompts found [2.4 brand mentions on average, with 44% of prompts producing zero mentions](https://www.brightedge.com/resources/weekly-ai-search-insights/chatgpt-brand-mentions-vs-citations-what-triggers-visibility). Prompts containing _best_ averaged 4.8 mentions, though a quarter still named no one. Commercial wording such as _where_, _deals_, and _affordable_ drew 4 to 8 times more mentions than informational wording such as _how_, _what_, or _cost_.

That distinction matters when you answer the CEO. If an informational prompt names nobody, you have a different problem from a buying prompt that names three competitors and skips you. If your competitor appears in only a few prompts, you need to know which few. Tonight, reply that **one screenshot is a sample of one** and that you will check the buyer questions before recommending spend.

### Your homepage may not answer what the buyer asked

Your homepage says what you sell. A buyer asks which option to pick, what it costs, or where it falls short. Those are different questions. For your own page to earn a citation, [crawlers need to reach it, the page needs a quotable answer, and your brand needs a presence on trusted third-party sites](https://mediaofficers.com/insights/how-to-get-cited-by-chatgpt-perplexity-ai-overviews). A features page cannot do much for a question that calls for a comparison or a limitation stated plainly.

This is why I would not start by rewriting the homepage. First find the missed prompts; then find the pages that ought to answer them. If you cannot point to a page that addresses the question, the gap is not a mysterious visibility problem. It is an unanswered question. **Fix the page that matches the question**, not the page everyone at the company already knows how to find.

### Look at the pages doing the quoting

Open the citations under the answer. Some will point to sites you do not control. The pattern varies by platform: [ChatGPT leans on Wikipedia, Reddit, and Forbes; Perplexity on Reddit and YouTube; and AI Overviews on Reddit, YouTube, and Quora](https://connectivewebdesign.com/blog/ai-seo). Your site is one possible source among many. In an answer built around a Reddit thread that names three vendors, your homepage may not be the most useful page in the pile.

That stings. It also gives you a place to look. Note which third-party pages name your competitor, what question those pages answer, and whether your brand belongs in that conversation. Do not put every omission in the same bucket. You can fix a missing answer on your site; you cannot edit someone else’s thread. **A missing mention and a missing page are not always the same problem.** Keep them separate in your notes.

## Do not pay for the meter before you find the miss

I understand the appeal of a dashboard. It looks like something you can forward to your CEO by Friday. But [monitoring tools track mentions and prompt triggers; they do not ship the content and crawl fixes](https://www.mersel.ai/blog/why-monitoring-tools-not-enough). I have watched teams pay for the meter and then stare at it while the competitor kept the citation. The number changed; the work did not.

Do not spend this week commissioning a homepage rewrite, either. A rewrite can consume the time you need to inspect the pages and prompts behind the screenshot. You may eventually need to change the homepage. You do not know that from one answer. **Measure the gap first, then buy help for the work you cannot do yourself.** The vendor email can wait until next Monday.

## Your seven days: build a prompt log, then diagnose the misses

You need a sheet, the screenshot’s exact prompt, and about an hour a day. Keep one row per prompt and record what you see rather than what you hoped to see. Save the wording as asked: a small change can turn a broad information request into a buying question, and then your rows no longer describe the same test. This is the week I would run at your desk, in order:

1. **Day 1: list 20 buyer prompts in your customers’ words.** Pull them from sales calls, support tickets, and search terms that already convert. Include the CEO’s prompt word for word; do not tidy its phrasing.
2. **Day 2: run the prompts across ChatGPT, Perplexity, and AI Overviews.** Save the answer, every brand named, and every citation link. [Logging 20 questions monthly for share of voice](https://mediaofficers.com/insights/how-to-get-cited-by-chatgpt-perplexity-ai-overviews) gives you a baseline; this is your first pass without a subscription.
3. **Day 3: check whether bots can reach your key pages.** Open `/robots.txt`, check for blocks on OAI-SearchBot, Claude-SearchBot, and PerplexityBot, and inspect what a bot can actually fetch. A page that loads in your browser may still be a poor source if its answer only appears after JavaScript runs.
4. **Day 4: map each missed prompt to the page that should answer it.** Mark the row if the page is absent, hard to fetch, or present but missing the direct answer. Blank page cells reveal work for your content queue; [this AI-readability swipe file](https://groas.com/post/the-ai-readability-swipe-file-robots-txt) gives you blocks to check when you make those pages citable.
5. **Day 5: list the third-party pages cited instead of you.** Copy 3 to 5 cited links for each miss and tag the type: Reddit thread, YouTube video, comparison site, publisher, or competitor page. Watch for repeats. As [_Stop Counting AI Brand Mentions. Fix the Pages That Never Get Cited_](https://groas.com/post/stop-counting-prompts-why-tracking-your) argues, the tracker records the miss; the page map tells you what needs attention.
6. **Days 6–7: choose three pages to write or fix, and ship what you can.** Pick one per cluster of missed prompts, not ten scattered ideas. Put a direct answer in the first 100 words, state a price range or limitation plainly where the question calls for it, and include a sentence that can stand alone as a citation. Log each publication date. **Book a retest of the same 20 prompts two weeks later.** Teams skip that last step, then cannot tell which fix changed an answer and start next month from another screenshot.

The order is deliberate. [Cloudflare began blocking AI crawlers by default, and major AI crawlers in one test did not run JavaScript](https://blog.hubspot.com/marketing/ai-search-optimization-tools). If a bot cannot fetch your page’s answer, polishing that answer is premature. If the page loads but never addresses the buyer’s question, changing bot settings will not write the missing paragraph. **Check the fetch, then judge the copy.**

By the end of Day 2, you have two piles: prompts where you appear and prompts where you do not. The rest of the week is about the second pile. Some misses will point to a page you need to write. Others will point to access, an answer buried in the wrong place, or a third-party source you cannot edit. Put those labels beside the prompts. A single visibility score cannot tell you whether to call a developer, brief a writer, or leave a third-party citation alone. Your sheet can.

## Tell your CEO what you found, not what a rank promises

Bring the sheet on Friday, not a subscription receipt. Start with the prompt count, then show where your brand appeared, where it did not, and what you found behind the misses. If your sheet shows six appearances and fourteen misses, say so. If nine misses map to pages you have not written, three to pages bots cannot load cleanly, and the rest to third-party pages quoting a competitor, say that too. **Use your counts, not mine.** Those numbers are an example of the report’s shape, not a result you can promise before you run it.

You do not have to explain away every blank row. You do have to distinguish a problem you can act on from one you have merely observed. That is a better answer to _why not us_ than a chart with a tidy line and no account of what sits behind it.

Then name the work and the retest date. “We have three pages to fix, and we will rerun the same prompts in two weeks” gives your CEO something the screenshot cannot: a cause to investigate and a way to see whether the fix helped. Do not promise a rank. You are checking whether the answers change.

If the question turns to tools, give the price context without pretending price is the decision. [Recurring monitoring ranges from $29/month for Otterly and $50/month for HubSpot AEO to $99/month for Profound Starter and $399/month for Growth](https://blog.hubspot.com/marketing/ai-search-optimization-tools). A tracker may make sense once you know which prompts matter. Before that, you would be paying each month to watch a gap you can diagnose this week. Promise the retest, not the dashboard.

![Small investigators mapping buyer prompts and competitor citations across a spreadsheet](https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/a922fc6a-dda9-4681-9fa5-5c9742c6ec14.png)

## Keep the sheet if you can ship the fixes; get help if you cannot

Keep this in-house if someone on your team can change `robots.txt` and publish a page without a ticket queue, a subject expert can write plain answers that include prices and limits, and your prompt set is small enough to retest by hand every two weeks. I ran accounts that size for years with a spreadsheet and a calendar reminder. If that is your situation, keep the sheet and the retest date. Skip the retainer.

That approach does not fit every team. If you have no dev access, no writer, or a CEO who expects a new screenshot answered every morning, the sheet can identify the work without getting it done. The same is true when misses spread across ChatGPT, Perplexity, and AI Overviews, third-party pages do much of the quoting, and you need new citable pages each month while holding down paid search. You need someone to own the fixes, not just display the misses.

A dashboard meters that work. [groas](https://groas.com/) executes it with specialized models running checks and fixes continuously and a human strategist owning direction and accountability. I like that split for the same reason I stopped treating repetitive account work as a mark of strategic genius: machines can handle the repeated checks; a person still has to decide what matters and answer for the result. If your week ends with fourteen misses and three pages you cannot ship alone, that is when you hand it off.

Run the week before you buy anything.

— Alexander

## Frequently Asked Questions

### My CEO got a ChatGPT screenshot showing a competitor instead of us. Should I panic about our AI visibility?

No. A single screenshot shows one buyer question and one answer, so it is a sample of one rather than a verdict on your brand. Before recommending spend, spend about a week finding which buyer prompts leave your brand out and why.

### How should I test whether my brand shows up in AI answers?

Test with buyer prompts written in your customers' words, not with your brand name. Searching your own brand name does not reflect how buyers actually ask. Include your CEO's prompt word for word, but it cannot be the whole test.

### Does the wording of a prompt change how many brands ChatGPT mentions?

Yes. One analysis of thousands of ChatGPT prompts found 2.4 brand mentions on average, with 44% of prompts producing zero mentions. Prompts containing 'best' averaged 4.8 mentions, and commercial wording such as 'where', 'deals', and 'affordable' drew 4 to 8 times more mentions than informational wording such as 'how', 'what', or 'cost'.

### How do I check whether AI crawlers can actually reach my pages?

Open /robots.txt and check for blocks on OAI-SearchBot, Claude-SearchBot, and PerplexityBot, then inspect what a bot can actually fetch. A page that loads in your browser may still be a poor source if its answer only appears after JavaScript runs, since major AI crawlers in one test did not run JavaScript.

### What should I do if competitor citations point to Reddit threads or other third-party pages?

Note which third-party pages name your competitor, what question those pages answer, and whether your brand belongs in that conversation. Keep the problems separate: you can fix a missing answer on your own site, but you cannot edit someone else's thread. A missing mention and a missing page are not always the same problem.

### What should I do in my first week after the ChatGPT screenshot?

Build a prompt log in a sheet, one row per prompt, taking about an hour a day. List 20 buyer prompts, run them across ChatGPT, Perplexity, and AI Overviews, check crawler access to your pages, map missed prompts to pages that should answer them, list third-party pages cited instead of you, and then write or fix a few pages and book a retest two weeks later.

### Should I buy an AI visibility monitoring dashboard right away?

Not this week. Monitoring tools track mentions and prompt triggers, but they do not ship the content and crawl fixes that actually change answers. Diagnose the gap first, then buy help for work you cannot do yourself; a tracker may make sense once you know which prompts matter.

### What should I tell my CEO at the end of the week instead of buying a tool?

Bring the sheet and report what you found: the prompt count, where your brand appeared, where it did not, and what sits behind each miss, such as pages not written, pages bots cannot load cleanly, or third-party pages quoting a competitor. Name the work and book a retest of the same prompts two weeks later, and do not promise a rank.

## Pay For Results, Not For Hours

Businesses buy the outcome, agencies resell it, and groas answers for it either way.

[See If You Qualify](https://groas.typeform.com/to/xC1bQNUT)

## Structured data

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://graph.groas.com/entity/groas","name":"groas","url":"https://groas.com/","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/an-open-letter-to-the-marketing-manager#webpage","url":"https://groas.com/post/an-open-letter-to-the-marketing-manager","name":"To the Marketing Manager Whose CEO Sent a Competitor’s ChatGPT Screenshot","description":"A competitor appeared in your CEO’s ChatGPT screenshot. Before buying a monitoring dashboard, spend seven days finding which buyer prompts miss you and why.","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/8c81f354-e1f8-45f7-ad04-ef58a9f453bf.png"},"breadcrumb":{"@id":"https://groas.com/post/an-open-letter-to-the-marketing-manager#breadcrumb"},"dateModified":"2026-10-18T05:25:52.946Z","mainEntity":{"@id":"https://groas.com/post/an-open-letter-to-the-marketing-manager#article"}},{"@type":"BreadcrumbList","@id":"https://groas.com/post/an-open-letter-to-the-marketing-manager#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":"To the Marketing Manager Whose CEO Sent a Competitor’s ChatGPT Screenshot","item":"https://groas.com/post/an-open-letter-to-the-marketing-manager"}]},{"@type":"BlogPosting","@id":"https://groas.com/post/an-open-letter-to-the-marketing-manager#article","url":"https://groas.com/post/an-open-letter-to-the-marketing-manager","mainEntityOfPage":{"@id":"https://groas.com/post/an-open-letter-to-the-marketing-manager#webpage"},"isPartOf":{"@id":"https://groas.com/blog#blog"},"headline":"To the Marketing Manager Whose CEO Sent a Competitor’s ChatGPT Screenshot","description":"A competitor appeared in your CEO’s ChatGPT screenshot. Before buying a monitoring dashboard, spend seven days finding which buyer prompts miss you and why.","datePublished":"2026-10-18T05:25:52.882Z","dateModified":"2026-10-18T05:25:52.946Z","image":{"@type":"ImageObject","url":"https://pub-87da24ecbbfc4c3bad6875f3aa013712.r2.dev/generated-images/8c81f354-e1f8-45f7-ad04-ef58a9f453bf.png"},"author":{"@id":"https://groas.com/author/alexander-perelman#person"},"publisher":{"@id":"https://graph.groas.com/entity/groas"},"keywords":"answer, prompts, screenshot, page, pages, chatgpt, chatgpt screenshot, ceo","wordCount":1906,"inLanguage":"en-US"},{"@type":"Person","@id":"https://groas.com/author/alexander-perelman#person","name":"Alexander Perelman","jobTitle":"Head Of Product @ groas","description":"Ex Goldman Sachs and Ex 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/an-open-letter-to-the-marketing-manager#faq","mainEntity":[{"@type":"Question","name":"My CEO got a ChatGPT screenshot showing a competitor instead of us. Should I panic about our AI visibility?","acceptedAnswer":{"@type":"Answer","text":"No. A single screenshot shows one buyer question and one answer, so it is a sample of one rather than a verdict on your brand. Before recommending spend, spend about a week finding which buyer prompts leave your brand out and why."}},{"@type":"Question","name":"How should I test whether my brand shows up in AI answers?","acceptedAnswer":{"@type":"Answer","text":"Test with buyer prompts written in your customers' words, not with your brand name. Searching your own brand name does not reflect how buyers actually ask. Include your CEO's prompt word for word, but it cannot be the whole test."}},{"@type":"Question","name":"Does the wording of a prompt change how many brands ChatGPT mentions?","acceptedAnswer":{"@type":"Answer","text":"Yes. One analysis of thousands of ChatGPT prompts found 2.4 brand mentions on average, with 44% of prompts producing zero mentions. Prompts containing 'best' averaged 4.8 mentions, and commercial wording such as 'where', 'deals', and 'affordable' drew 4 to 8 times more mentions than informational wording such as 'how', 'what', or 'cost'."}},{"@type":"Question","name":"How do I check whether AI crawlers can actually reach my pages?","acceptedAnswer":{"@type":"Answer","text":"Open /robots.txt and check for blocks on OAI-SearchBot, Claude-SearchBot, and PerplexityBot, then inspect what a bot can actually fetch. A page that loads in your browser may still be a poor source if its answer only appears after JavaScript runs, since major AI crawlers in one test did not run JavaScript."}},{"@type":"Question","name":"What should I do if competitor citations point to Reddit threads or other third-party pages?","acceptedAnswer":{"@type":"Answer","text":"Note which third-party pages name your competitor, what question those pages answer, and whether your brand belongs in that conversation. Keep the problems separate: you can fix a missing answer on your own site, but you cannot edit someone else's thread. A missing mention and a missing page are not always the same problem."}},{"@type":"Question","name":"What should I do in my first week after the ChatGPT screenshot?","acceptedAnswer":{"@type":"Answer","text":"Build a prompt log in a sheet, one row per prompt, taking about an hour a day. List 20 buyer prompts, run them across ChatGPT, Perplexity, and AI Overviews, check crawler access to your pages, map missed prompts to pages that should answer them, list third-party pages cited instead of you, and then write or fix a few pages and book a retest two weeks later."}},{"@type":"Question","name":"Should I buy an AI visibility monitoring dashboard right away?","acceptedAnswer":{"@type":"Answer","text":"Not this week. Monitoring tools track mentions and prompt triggers, but they do not ship the content and crawl fixes that actually change answers. Diagnose the gap first, then buy help for work you cannot do yourself; a tracker may make sense once you know which prompts matter."}},{"@type":"Question","name":"What should I tell my CEO at the end of the week instead of buying a tool?","acceptedAnswer":{"@type":"Answer","text":"Bring the sheet and report what you found: the prompt count, where your brand appeared, where it did not, and what sits behind each miss, such as pages not written, pages bots cannot load cleanly, or third-party pages quoting a competitor. Name the work and book a retest of the same prompts two weeks later, and do not promise a rank."}}]}]}
```