To track your brand in ChatGPT, Perplexity and Gemini, write a fixed set of questions your customers really ask, put them to each assistant every month, and log four things: whether your brand is mentioned, whether your site is cited as a source, which competitors are named, and which sources the assistant relied on. The trend over several months matters more than any single answer.
Why can't you just check once?
AI answers change between runs, users, locations and model versions. Asking the same question twice can give a different list of names. One check is a snapshot; a fixed question set repeated monthly turns noise into a trend you can act on.
Assistants with web search, such as Perplexity, ChatGPT with search and Gemini, also change as the web changes. A new comparison article or review thread can shift answers within weeks.
How do you build a good question set?
Start with 15 to 30 questions grouped by intent, written the way a real person would type or speak them. Avoid putting your brand name in most of them, because you want to know if the assistant brings you up unprompted.
| Group | Example for a study-abroad consultant in Delhi |
|---|---|
| Category | Which are the best study abroad consultants in Delhi for Canada? |
| Problem | How do I get a student visa for Germany after B.Tech? |
| Comparison | Is it better to apply to UK universities directly or through an agent? |
| Price | How much do study abroad consultants charge in India? |
| Brand | Is [your brand] good for UK admissions? |
Take the questions from enquiry calls, counselling sessions, Search Console queries and the "People also ask" boxes on Google. Keep the wording fixed once chosen so month-to-month results are comparable.
What should you record for each answer?
Record the same fields every time, in a spreadsheet or tracking tool. Consistent fields are what make the data useful.
- Mentioned. Does your brand name appear in the answer text?
- Cited. Is a page on your site linked as a source? Note which page.
- Position. Are you named first, in the middle or as an afterthought?
- Accuracy. Are your fees, locations, courses and contact details correct?
- Competitors named. List every other brand mentioned.
- Sources used. List the domains the assistant linked, such as review sites, directories, news or forums.
- Sentiment. Is the description positive, neutral or negative?
What is the difference between a mention and a citation?
A mention is your brand name in the answer. A citation is a link to your site as a source. You can have one without the other.
Being mentioned without being cited usually means the assistant learned about you from other sites, such as reviews or listicles. Being cited without being mentioned usually means your page answered a factual sub-question but the assistant did not recommend you. Both are worth tracking separately because they are fixed in different ways: citations through clearer on-site content, mentions through reputation elsewhere.
How do you run the checks fairly?
Reduce the things that make answers vary, and write down what you could not control.
- Use a logged-out or fresh session where possible, so past chats do not personalise answers.
- Note the date, the assistant and the model or mode used.
- Ask each question in a new chat, not as a follow-up.
- Run important questions two or three times and record how often you appear.
- Keep the same location settings each month, since answers to "near me" or city questions depend on location.
How do you act on the results?
Look at the sources column first. If assistants keep citing the same three review sites or a particular comparison article when naming competitors, those are the places where your brand needs to be present and accurately described.
If answers state wrong facts about you, fix them at the source: your own site, your Google Business Profile and the directories that carry your details. If your pages are never cited for factual questions, rewrite them with clear answers under question headings, as covered in generative engine optimization.
Common mistakes
- Only asking questions that include your brand name, which flatters results.
- Changing the question wording every month.
- Treating one bad answer as a crisis instead of watching the trend.
- Ignoring the sources list, which is the most actionable data you collect.
How to do this with Spider
Spider can run your chosen questions across ChatGPT, Perplexity and Gemini on a schedule, recording mentions, citations, competitors and sources so you do not have to copy answers by hand. It also tracks which sites Google AI Overviews cite for your keywords. Report these alongside your normal SEO KPIs so AI visibility sits next to rankings and traffic.
Check this on your own site
Spider's free check tests this and about 100 other things on any page, and links each problem to a guide like this one.
