How to Check What ChatGPT and Perplexity Say About Your Business

By Editorial Team · Updated September 2, 2026 · 11 min read

Check what ChatGPT and Perplexity say about your business by running a small prompt library in a fresh session at least three times, then scoring named versus cited versus wrong. One screenshot is noise, because these answers shift between identical runs, sometimes a lot.

An AI visibility check is a repeatable prompt test, not a casual chat in a logged-in thread. RedFork has helped more than 800 businesses grow and built 500-plus websites since 2014, and the pattern is consistent: Partners who treat one branded ChatGPT tab as a report come back surprised the next week, while Partners who keep a sheet see the swing and stop overreacting to any single answer.

Key Takeaways

• Score named and cited as two columns. A shop can be linked and never named, or named and never linked.

• Run each prompt at least three times in a fresh, logged-out session with memory off, because a single run isn't a reading.

• Note whether ChatGPT actually searched. A no-search answer from memory is data, not a void, so record it as "search: no."

• Recognition isn't recommendation: SOCi found ChatGPT surfaces about 1.2% of local locations on unbranded prompts, versus 35.9% in Google's 3-pack.

• BrightLocal found 45% of 1,002 US adults had asked AI for a local recommendation, and most verify against real reviews before acting.

The Short Version

1. Copy the 12 prompts. 2. Open a fresh ChatGPT session and a fresh Perplexity session, logged out, no memory. 3. Run each prompt three times. 4. Fill the split sheet: named, cited, sources, wrong facts. 5. Fix the pages and listings the answers already used.

Field	What you write down
Engine	ChatGPT, Perplexity, Gemini, Copilot
Prompt	The exact sentence you pasted
Named?	Yes or no, in the prose
Cited?	Yes or no, as a source link
Sources	Every URL or site the answer showed
Wrong?	Hours, address, services, or city errors, verbatim

In This Article

• The Split Sheet

• Prompt Library for a Local Shop

• How Do You Run ChatGPT Without Contaminating the Test?

• How to Read a Perplexity Answer

• Are Named and Cited the Same Result?

• How Many Runs Before You Trust a Screenshot?

• What to Do With a Bad Week of Answers

The Split Sheet

The split sheet is a five-column log you fill on every run: engine, prompt, named, cited, and the sources the answer showed. Mentions and citations diverge, so a shop can be named in the prose with no link, or cited as a source and never named.

Print the sheet or copy it into a notebook, and date every row. Canonicalize URLs before you count, because example.com/page and www.example.com/page/ are one citation, not two. The point of the sheet is to catch the swing that a single tab hides: one branded ChatGPT answer looks like a verdict, and a week of runs looks like a rate. The Partners who keep the sheet stop overreacting to a good screenshot and a bad one in the same week, because both are just rows.

Prompt Library for a Local Shop

The prompt library is twelve sentences a neighbor would type, split into branded checks, unbranded city prompts, and problem-led asks. Rewrite the brackets for your shop, and don't add your brand to the unbranded set.

Branded (3). These test recognition and fact accuracy, not recommendation.

1. What do you know about [Business Name] in [City]?

2. Is [Business Name] a good [trade] in [City]? What are the hours and the address?

3. How does [Business Name] compare to other [trade] options in [City]?

Unbranded city (6). These are the product.

1. Who should I call for [job] in [City]?

2. Best [trade] near [neighborhood or city]?

3. I need [service] this week in [City]. Who do you recommend?

4. Which [trade] in [City] is open Saturday and takes walk-ins?

5. Affordable [service] in [City] for a small job?

6. Who does [specific job] in [City], and what should I budget?

Problem-led (3). These test whether the model reaches for you without a category noun.

1. My [thing] broke in [City]. What do I do, and who can help today?

2. I just moved to [City] and need a regular [trade]. Where should I start?

3. How do I tell a good [trade] from a bad one in [City]?

Cross-engine overlap is low, so one engine's row tells you little about the next. Run the same library on ChatGPT, Perplexity, and, if you have time, Gemini and Copilot, and keep the results in separate rows rather than averaging them into one "AI score."

How Do You Run ChatGPT Without Contaminating the Test?

You run ChatGPT in a fresh, logged-out session with memory off and no custom instructions, then paste one prompt and stop there. Follow-ups train the thread, and that's a conversation, not a test.

ChatGPT's search draws on Bing-indexed results plus its own logic, and it doesn't always run a live search, so note whether each run actually searched and record a memory-only answer as "search: no" rather than scoring it as a retrieval miss. Keep the location consistent: if you live in one city and serve another, name the city in the prompt, because a "near me" with no city is a lottery. If you repeat the audit, use the same day of the month and the same window, since personalization and timing are both contamination you can control.

How to Read a Perplexity Answer

A Perplexity answer is a short list plus visible sources, so you read the prose for the name and the source list for the URL. Those two lines are your named and cited columns for that run.

Perplexity almost always retrieves, which makes it the easiest engine to score and the most likely to show you the exact page it trusted. If your shop is named but the sources are Yelp, Reddit, and a competitor's blog, the fix isn't another homepage paragraph, it's presence on the pages it actually cited. If your URL is cited but the prose names someone else, you won the footnote and lost the recommendation. Write both rows down and don't average them into a single score, because a clean branded answer sitting next to a silent city prompt is the common local pattern, and collapsing them hides the exact gap you're trying to fix.

Are Named and Cited the Same Result?

Named and cited are two different outcomes on the same answer, and a shop can win one while completely losing the other. Named means the shop appears in the sentence a customer reads, cited means a source link points at you, and wrong means the answer invented or stale-dated a fact.

Score them apart, because a mention without a citation is still a recommendation, and a citation without a mention is a footnote. Wrong hours, a closed location, or a service you don't offer is a trust problem even when the name is right, and BrightLocal found most AI users fact-check a recommendation against a real review or source, so the customer will see the error. Copy the wrong sentence verbatim, because that's the rewrite brief for the page or listing that fed it. Don't argue with the model in the same thread; fix the source, then re-run next month.

How Many Runs Before You Trust a Screenshot?

You shouldn't trust one ChatGPT or Perplexity screenshot as a visibility score, because answers shift between identical runs, sometimes a lot. Three runs is the floor for a shop-owner audit, and more is better if you have the time.

Look at the pattern, not the hero run. If you're named on one of three ChatGPT runs and two of three Perplexity runs, you have a rate, not a verdict. If run one names you and run four names two others, both rows stay on the sheet. Cited sources also churn as models update, so measure monthly and react quarterly rather than after every model headline. Paid AI-visibility tools can automate the sheet if you later need share of voice across many prompts or locations, but manual is enough to start, because the sheet is the habit and the tool is optional.

What to Do With a Bad Week of Answers

A bad week of answers is a source list, not a reason to buy a third acronym, so fix the pages those answers already used. Read the sources the engines showed, then make those pages and listings tell the same story about your shop.

If branded prompts fail, the entity is thin, so start with the About page, matching listings, and a consistent name-address-phone string; the mechanics of that live in Google Business Profile in AI answers. If branded prompts pass and unbranded city prompts fail, you have recognition without recommendation, which is the harder gap, covered in how to get ChatGPT to recommend your business. If the sources are directories, go there; if the sources are a competitor's city page, write the extractable sentence you're missing. Express websites on curated Wix templates cover a one-location site plus ads-only marketing, while Pro custom Wix work with the full marketing mix fits when the page set and content need to change. Unserious Marketing won't replace a split sheet.

If the Sheet Is Full, Fix the Sources It Named

Keep the library, date the rows, and change the pages and listings the answers already trusted. That's the whole audit, and it beats arguing with a chatbot in a live thread.

If you want that work scoped on a Wix site, we'll read the sheet with you and fix the sources it named. Express if you need a template and ads, Pro if you need the full mix, and we won't treat one screenshot as a launch date.

GET PRICING at redforkmarketing.com/sales-contact.

Frequently Asked Questions

Can I run this inside my usual ChatGPT account?

You can, and you'll contaminate it. Memory, custom instructions, and prior turns all change the answer, so use a fresh session, and logged out is cleaner still. The whole point is to see what a stranger in your city sees, not what a model tuned to your own history serves back to you, so strip away everything that personalizes the result.

What does a branded prompt prove?

A branded prompt proves recognition and fact accuracy, not that a neighbor will hear your name. Use it to catch wrong hours, a stale address, or a service you don't offer, and use the unbranded set to score actual recommendation. Passing the branded check while failing the unbranded one is the most common local result, which is why you run both.

Should I include Gemini and Copilot?

Yes, if you have time, because cross-engine overlap is low. Copilot leans on Bing and Gemini leans on Google's ecosystem, so a ChatGPT-only sheet is really just a ChatGPT report. If time is tight, start with ChatGPT and Perplexity, which cover the two most common behaviors, then add the others once the habit is in place.

How often should I repeat the audit?

Monthly is enough for a one-location shop, using the same week, the same window, and the same prompt set each time. Don't refresh the score after every model headline, because the run-to-run noise will swamp any real trend at that cadence. A steady monthly rhythm shows you the direction; a daily one just shows you the variance.

Do I need a paid AI visibility tool?

No. The sheet is free, and manual runs are enough to start. Add a tool only when you need share of voice across many prompts or many locations, where automation genuinely saves time. For a single shop, the discipline of running the library and dating the rows matters far more than any dashboard, so build the habit before you buy software.

How We Researched This

We checked the live sources for this topic on September 2, 2026.

We read BrightLocal's 2026 Local Consumer Review Survey (45% of consumers using AI for local recommendations, and most verifying against real reviews), SOCi's 2026 Local Visibility Index (ChatGPT surfacing about 1.2% of locations versus 35.9% in Google's 3-pack), and OpenAI's documentation that ChatGPT's search may draw on Bing-indexed results.

The prompt library and split sheet are original to this page and sized for a local shop, not a large enterprise study. Company figures (500-plus websites, more than 800 Partners, Wix Certified, founded 2014) are RedFork-reported.

Latest Updates

• September 2, 2026: Refreshed. Supporting figures re-verified against sources read September 2, 2026, unverifiable per-study numbers removed, and a byline, read time, and standard section order added.

References

1. BrightLocal, "Local Consumer Review Survey 2026." https://www.brightlocal.com/research/lcrs-ai-trust/

2. SOCi, "In AI-Driven Discovery, Few Brands Are Chosen, Most Disappear." https://www.soci.ai/news/in-ai-driven-discovery-few-brands-are-chosen-most-disappear/

3. OpenAI, "ChatGPT Search." https://help.openai.com/en/articles/9237897-chatgpt-search

4. Wix Marketplace, "RedFork Marketing." https://www.wix.com/marketplace/wix-partner/redfork-marketing

Disclaimer

This article is educational. RedFork doesn't guarantee rankings, citations, traffic, or revenue. Statistics were current as of September 2, 2026 and should be re-checked on refresh.