How to Find Topics Where Competitors Outperform You in ChatGPT

What does "a competitor outperforms you in ChatGPT" actually mean?

It is not one ranking. It is a pattern across many questions and runs, in three layers:

• Mentioned: the competitor's name appears in the answer.
• Recommended: the answer suggests it as a good choice for the question.
• Cited: one of its pages is linked as a source.

A competitor can win one layer and lose another. It may be recommended often but rarely cited, because third-party reviews carry it. Or it may be cited for how-to questions but rarely recommended. The useful unit is the topic, meaning a group of related questions, because that is what you can act on with content and outreach.

How do you track competitor mentions in AI answers and find visibility gaps?

1. Write 40–80 buyer questions in the words customers use, from sales calls, support tickets, reviews and search data.
2. Group them by topic and intent: for example "best for small teams", "integrations", "pricing", "alternatives to X", "how to do Y".
3. Run each question three to five times in ChatGPT with search on, in the country you care about. Use an unpersonalized temporary chat, or turn off memory and personalisation, so earlier conversations do not influence later answers.
4. Record for each answer: every brand named, its role (recommended, listed, mentioned), its position, and all cited URLs.
5. Normalise brand names, merging variants such as "HubSpot Sales Hub" and "HubSpot".
6. Count per topic: in how many runs each brand appears and is recommended.
7. Flag gaps: topics where a competitor appears in most runs and you appear in few or none.
8. Repeat monthly with the same questions to see movement.

Repetition matters because answers vary a lot between runs. Counts across runs show real patterns; single answers do not.

How do you find which brands AI mentions for category questions?

Start with broad category questions, such as "What are the best [category] tools for [audience]?" and "Which [category] should a [company type] use?" Run them repeatedly and list every brand named. This list is the engine's view of your category. It often includes:

• the market leaders you expect;
• smaller brands that are strong in reviews or comparison articles;
• adjacent products from neighbouring categories;
• occasionally, discontinued or renamed products from older training data.

Then narrow the questions by audience, budget, integration or use case. Brands that appear only for specific questions are often the niche competitors that beat you where you should be strongest.

Which competitors does AI mention instead of my brand?

For each question where you do not appear, list the brands that do, and count how often each appears across your gap questions. The two or three brands that recur most are the ones taking your place. Check their role too: a competitor that is recommended instead of you is a bigger problem than one merely listed. Why this happens for service firms is covered in why ChatGPT recommends competitors instead of your agency or consulting firm.

How should you log answers so the analysis is repeatable?

A simple spreadsheet with one row per brand per answer is enough:

Keep the full answer text in a separate sheet or folder, keyed by date, question and run. With this format, the topic grid is a pivot table. Month-to-month comparisons use the same questions, and anyone can check a number by opening the answers behind it.

How do you turn answers into a topic gap map?

Summarise the counts in a grid, one row per topic:

Prioritise large gaps on topics that bring revenue. Then open the answers for those topics and list the sources cited. That tells you whether to fix your own pages, get into a third-party article, or build reviews.

How do content gaps affect competitive positioning in AI search?

When your site does not answer a question, ChatGPT has nothing from you to cite and less evidence to recommend you. It fills the gap with competitors' pages and third-party sources that describe them. Over many questions, this shapes how the engine positions each brand. The competitor with pages on integrations, migration and pricing becomes "the one that integrates with Salesforce". You become "an alternative", or disappear.

Typical content gaps behind lost topics:

• no page for a use case or audience you serve well;
• integration details buried in documentation or missing;
• no comparison or alternatives page, so others' comparisons define you;
• prices not stated, so budget questions go to brands that state them;
• no migration or setup guidance, so "switching" questions go to competitors.

Content alone rarely closes a gap on recommendation questions, which also depend on reviews and third-party coverage. But without it, engines have nothing of yours to cite, and you depend entirely on others describing you correctly.

Which competitors should agencies track in AI visibility comparisons?

Track two groups:

1. Business competitors: the companies the client loses deals to, from sales records and win/loss analysis.
2. AI competitors: the brands ChatGPT and other engines actually name for the client's questions, found through the category questions above.

Where the two lists overlap, you have the core set. Where they differ, both matter: business competitors show where revenue is at stake, and AI competitors show who occupies the answers.

How many competitors should you compare your visibility against?

Three to five named competitors per report is usually enough to give context without clutter. Record every brand in the raw data, since a newcomer may start appearing. But report on the few that matter, and add a line for "other brands" so readers see the whole share.

How do I define competitors when ChatGPT recommends different companies for each buyer question?

Define competitors by topic, not globally. A project management tool may compete with one set of brands for "agile teams", another for "construction projects" and another for "free plans". Keep a small overall set for the headline report, and per-topic competitor lists for the detailed analysis. That reflects how buyers ask and how engines answer.

What should an agency do when a client's real competitors differ from its SEO competitors?

SEO competitors are often publishers, directories and review sites that rank for the same keywords, not companies selling the same thing. In AI answers, those publishers matter as sources, not as competitors. So:

• report business competitors and AI-named competitors as competitors;
• report publishers, review sites and directories as sources to be present in;
• explain the difference to the client, since it changes the plan from "outrank them" to "get included by them".

Why does AI name competitors that are not actual market rivals?

• Co-occurrence: brands often mentioned alongside yours in articles get linked in the model's associations.
• Category ambiguity: a vague category name mixes different markets, such as "CRM" for sales teams and for property managers.
• Name similarity: similarly named companies get confused.
• Old information: discontinued or merged products persist from training data.
• Adjacent products: tools that solve part of the problem get listed as alternatives.

When this happens, clarify your category and audience on your site and profiles, and see whether third-party sources group you with the wrong companies.

How do I find competitors that AI recommends when Google search results do not?

Compare the two lists directly. For each question, record the brands named in Google's top results, including those listed inside ranking articles, and the brands ChatGPT names. Brands that appear only in ChatGPT answers usually have one of three advantages. They may be strong in Bing or in other sources ChatGPT draws on, well covered in reviews and comparison articles, or well established in older training data. Studies have found that most pages ChatGPT cites do not rank on Google's first page, so this divergence is normal. How ChatGPT's sources differ from Google's is covered in whether 90% of ChatGPT citations come from pages ranked 21 or lower.

What does this look like in specific industries?

How do I find the sales technology competitors recommended instead of my platform by AI?

Build questions around how sales leaders shop: "best sales engagement platform for a 20-person SDR team", "sales tool that integrates with HubSpot and Gong", "cheaper alternative to [leader]", "sales software with built-in dialer for the UK". Group by team size, integrations, channel and budget. In this category, answers lean heavily on review platforms and on comparison articles from sales publications and consultants. Check which of those cite the competitors that win, and whether your integrations, pricing and team-size fit are stated anywhere ChatGPT can read.

How do I find which healthcare software vendor ChatGPT recommends instead of ours?

Use questions that reflect clinical and compliance needs: "best EHR for a three-physician paediatric practice", "practice management software that integrates with [lab or billing system]", "HIPAA-compliant telehealth platform for therapists". Group by specialty, practice size, integration and compliance. Check whether healthcare answers cite review platforms, healthcare IT publications or research firms such as KLAS. Check whether the answers that win mention compliance or certification, such as HIPAA or ONC Health IT certification, and whether your own pages state yours clearly. Gaps often come from missing specialty pages or unclear compliance details.

How can I see which headless CMS competitors AI assistants recommend instead of us?

Developers ask technical questions: "best headless CMS for a Next.js ecommerce site", "headless CMS with visual editing for marketers", "open-source headless CMS I can self-host", "headless CMS with good localisation". Expect well-known vendors such as Contentful, Sanity, Strapi and Storyblok to appear often. Answers draw on documentation, framework integration guides, GitHub activity, developer community lists and comparison posts. Gaps usually trace to missing framework guides, unclear pricing for scale, or absence from the lists developers cite.

What should you do with the gaps?

Pick three to five gap topics with the highest revenue value and work through them before expanding. How to study the competitor pages that ChatGPT cites is covered in how to reverse-engineer competitor pages cited by ChatGPT.

What do people ask about competitor gaps in ChatGPT?

Can I just ask ChatGPT who my competitors are?

You can, and it is a useful starting point, but one answer reflects one run. Build the list from repeated category questions instead.

Should I use the API or the ChatGPT app?

For reflecting what buyers see, the app with search on is closer. APIs are useful for scale but may behave differently.

How often should I redo the gap analysis?

Rerun the question grid monthly and review competitor lists quarterly, or after major launches and news.

Do I need a tool for this?

Not for a first analysis of 40–80 questions. Tools help when you track many questions, engines or markets over time.