You can ask ChatGPT a question and look for your company name. That gives you one result.
It does not tell you whether the brand is consistently visible, which competitors appear instead, whether the answer cites your website, or how the result changes across different customer questions.
To measure brand visibility usefully, you need a controlled prompt set, a repeatable process, and a record of the answers.
Quick answer
To check whether ChatGPT mentions your brand:
- Define the exact brand name and common variations.
- Select real competitors.
- Create a balanced set of branded and non-branded customer questions.
- Run the prompts in a consistent ChatGPT surface and record the context.
- Capture mentions, citations, recommendations, competitors, and sources.
- Repeat the same prompt set over time.
- Connect missing visibility to relevant pages on your website.
A single manual search is useful for exploration. It is not a reliable visibility report.
What does a ChatGPT brand mention look like?
A brand can appear in several ways.
Direct mention
The answer names the brand, but does not necessarily link to it or recommend it.
Citation
The answer links to the brand’s website or uses one of its pages as a source.
The brand may also be mentioned while the citation points to a third-party comparison, directory, review, or article.
Recommendation
The answer presents the brand as a suitable option for a specific use case, audience, location, or constraint.
Comparison
The answer places the brand alongside competitors and describes differences.
Misrepresentation or stale information
The answer mentions the brand but describes an old feature set, incorrect audience, outdated price, or unsupported capability.
Visibility is therefore more than a yes-or-no mention count. The context matters.
Start with the measurement goal
Before writing prompts, decide what you want to learn.
Possible goals include:
- Check whether ChatGPT recognizes the brand name.
- Measure discovery for non-branded category questions.
- Compare visibility with specific competitors.
- Find questions where competitors appear and the brand does not.
- Track whether the official website is cited.
- Monitor how the brand is described.
- Verify whether a new page or product launch changes the observed answers.
The goal determines the prompt set.
If you only ask branded questions such as “What is Company X?”, you are measuring recognition. You are not measuring whether a new customer would discover Company X without already knowing its name.
Build a useful ChatGPT prompt set
A balanced project should include several prompt groups.
1. Branded prompts
Use these to check recognition and description.
- “What is [Brand]?”
- “What does [Brand] do?”
- “Who is [Brand] for?”
- “What are the main features of [Brand]?”
- “What are the alternatives to [Brand]?”
These prompts are useful for accuracy, but they should not dominate the project.
2. Category prompts
Use these to test discovery without naming the brand.
- “What are the best [category] tools?”
- “Which [category] services are suitable for small businesses?”
- “What tools help with [category problem]?”
- “Which platforms offer [core capability]?”
Category prompts reveal whether the brand is associated with the market it wants to serve.
3. Problem and use-case prompts
These often reflect stronger real-world intent.
- “How can I solve [specific problem]?”
- “What should I use if [constraint]?”
- “How can an agency manage [problem] across client websites?”
- “How can I fix [problem] without rebuilding my website?”
For Prerender Buddy, an example would be:
“How can I make an already-live React website readable to AI crawlers without migrating to SSR?”
4. Comparison prompts
Use these to observe competitor framing.
- “What are the best alternatives to [Competitor]?”
- “[Brand] vs [Competitor]: which is better for [audience]?”
- “What is a simpler option than [established product]?”
- “Which tool is better for a small agency?”
5. Audience and location prompts
These reveal whether visibility changes with context.
- “What is the best [category] tool for a startup?”
- “Which [service] works for agencies?”
- “Which companies provide [service] in [location]?”
- “What should a local business use for [problem]?”
Only use audience and location variations that match the actual product and market.
Avoid leading prompts when measuring discovery
A prompt such as “Why is [Brand] the best solution?” includes the conclusion you are trying to measure.
It may be useful for testing how the system handles a claim, but it is not a neutral discovery prompt.
Prefer questions that allow multiple answers:
- “Which tools solve this problem?”
- “What are the options for this audience?”
- “How do these products compare?”
Keep branded and non-branded prompt results separate.
Record the context of every run
An answer without context is difficult to compare later.
Record:
- exact prompt
- date and time
- provider and product surface
- model, when shown
- whether web search or retrieval was active
- location or language context, when relevant
- complete answer
- brand mention
- recommendation context
- citations and linked sources
- competitors mentioned
Screenshots are useful evidence, but structured data is easier to compare across many prompts and dates.
Create a simple tracking table
| Prompt | Brand mentioned | Brand cited | Recommended | Competitors | Sources | Date |
|---|---|---|---|---|---|---|
| Best [category] for small agencies | Yes/No | Yes/No | Yes/No | Names | URLs | Date |
| Alternatives to [competitor] | Yes/No | Yes/No | Yes/No | Names | URLs | Date |
| How to solve [problem] | Yes/No | Yes/No | Yes/No | Names | URLs | Date |
This is enough for a small manual baseline.
Once the project contains dozens of prompts, multiple providers, and historical runs, automated tracking becomes more practical.
Check more than the company name
Brand matching should account for:
- official brand name
- common spacing or punctuation variations
- product name versus company name
- domain name
- known abbreviations
Avoid overly broad aliases that create false positives.
For example, a short common word should not automatically count as a brand mention unless the answer context clearly identifies the company.
Distinguish a mention from a citation
If ChatGPT names the brand, inspect the linked sources.
Possible outcomes include:
- The official website is cited.
- A third-party article is cited.
- A directory or marketplace is cited.
- A competitor comparison is cited.
- The brand is mentioned without a visible citation.
- The official website is cited without a prominent brand recommendation.
Each outcome suggests a different investigation.
A third-party citation is not automatically bad. It may show which external sources influence category discovery. But if the official page should be the clearest source for a fact and it is never cited, review crawler access, page clarity, specificity, freshness, and supporting evidence.
Check whether ChatGPT can access the relevant pages
OpenAI’s current publisher guidance says public websites can appear in ChatGPT search and advises publishers not to block OAI-SearchBot if they want content available for summaries and snippets.
That access is useful, but it does not guarantee a mention.
For every important missing prompt, identify the page you expected ChatGPT to use and check:
- whether the URL is public
- whether relevant crawler access is allowed
- whether the raw response contains meaningful content
- whether the page answers the prompt directly
- whether the title, heading, and copy describe the product clearly
- whether the page is current
AI Crawler Access vs AI Visibility explains why these are separate measurements.
Compare your brand with real competitors
Do not count every company named in an answer as a direct competitor.
Track competitors that overlap on:
- customer problem
- target audience
- geography
- product category
- deployment model
- price or company size
Then inspect prompt-level patterns:
- Which competitor appears most often?
- Does it appear for one topic or across several?
- Is it cited directly or through third-party sources?
- Is it recommended for an audience your brand also serves?
- Which wording or evidence is clearer on its relevant pages?
The objective is not to copy competitor content. It is to understand why the available evidence supports one answer more easily than another.
Repeat the same prompts
AI-generated answers can vary. One result can be favorable or unfavorable by chance.
Use the first complete run as a baseline. Repeat the same prompt set on a consistent schedule and compare completed batches.
Weekly monitoring is a practical starting point for an active project. Also run a new batch after:
- a major content publication
- a product or positioning change
- a website migration
- an important technical fix
- a significant competitor launch
Do not change half the prompt set and then present the result as a direct trend.
Track ChatGPT referral traffic separately
Answer visibility and website visits are different metrics.
OpenAI’s publisher guidance says referral links from ChatGPT include utm_source=chatgpt.com, allowing publishers to inspect inbound ChatGPT traffic in analytics tools.
Use referral traffic as an additional outcome, not as a replacement for prompt-level monitoring. A brand can be mentioned without receiving a click, and a cited page can receive traffic from a prompt outside your tracked set.
What should you do when the brand is missing?
Do not immediately generate a generic article.
Classify the gap:
- Access gap: the relevant page is blocked, unavailable, or returns thin content.
- Prompt gap: competitors appear for a specific high-value question.
- Topic gap: the site does not adequately cover the underlying subject.
- Citation gap: other sources are cited while the expected brand page is not.
- Positioning gap: the brand’s page does not clearly connect the product with the use case or audience.
- External evidence gap: the answer relies on third-party sources where the brand has little presence.
Then choose whether to fix, update, create, consolidate, or pursue external coverage.
Who this is for
- SaaS founders with already-shipped JavaScript websites
- React, Vite, Vue, Lovable, Bolt, or Base44 users
- SEO freelancers checking crawler-readable HTML
- Agencies maintaining client sites without rebuilding them
Where Prerender Buddy fits
Prerender Buddy turns this manual workflow into a repeatable project.
The AI Visibility layer is designed around:
- project-specific brands and competitors
- controlled prompt groups
- provider-level runs
- mentions, citations, recommendations, and sources
- completed historical batch comparisons
- prompt-level evidence
The broader workflow then checks whether relevant pages are accessible and healthy before recommending what to improve.
That prevents AI Visibility from becoming a collection of screenshots or a score with no next action.
What this process cannot prove
Prompt tracking cannot reveal every question a real user may ask.
It cannot guarantee that two users will receive identical answers.
It cannot prove the internal reason a provider selected a source.
It provides controlled observations across the prompts, providers, dates, and contexts you selected.
Use that evidence to make better decisions without presenting it as universal certainty.
You may not need Prerender Buddy if
- Server HTML is already complete.
- Static pages crawl correctly.
- You are already rebuilding with SSR or static generation.
- You only need an audit, not a rendering fix.
Final recap
To see whether ChatGPT mentions your brand:
- Define the brand and real competitors.
- Build branded and non-branded prompt groups.
- Record the complete context and answer.
- Separate mentions, citations, recommendations, and competitor appearances.
- Inspect the cited sources.
- Test the website pages connected to missing prompts.
- Repeat the same prompt set over time.
- Classify the gap before choosing an action.
One manual search can start the investigation. A controlled project turns it into evidence.
If the results expose missing mentions, competitors, or citations, continue with Why ChatGPT Mentions Competitors and How to Find AI Visibility Gaps to classify the evidence and choose the next action.
Start from a healthy website
Check crawler-readable content before treating every missing mention as a content problem.
Start free to connect the website and begin the Prerender Buddy visibility workflow.
Frequently asked questions
Does ChatGPT know my business exists?
Test branded recognition prompts first, then use non-branded category, problem, audience, and comparison prompts to see whether the business is discoverable without naming it. Recognition and non-branded visibility are different measurements.
How many prompts should I track?
Begin with a small, balanced set that covers the most important customer intents. Twenty well-chosen prompts are more useful than hundreds of minor variations. Expand when the first results reveal distinct topics, audiences, or comparison contexts.
How often should I check ChatGPT mentions?
Weekly checks are a practical starting point for an active project. Use the same controlled prompt set so completed batches remain comparable.
Is a brand mention the same as a recommendation?
No. A mention only confirms that the brand name appeared. A recommendation presents it as a suitable option for the requested context. A citation indicates that a source or page was linked or used in the answer.
Can I track visits from ChatGPT?
Yes. OpenAI’s current publisher guidance says ChatGPT referral URLs include utm_source=chatgpt.com. Referral analytics should be measured separately from prompt-level visibility.