How to check if ChatGPT, Perplexity and Google AI Overviews mention your brand

Short answer: you can check by hand for free in about an hour, and you should do that once before paying for anything. A tool only becomes worth it when you need the same questions asked every day, across several engines, for you and your competitors. That is the part humans are bad at.

This guide covers the manual method step by step, the mistakes that make manual checks misleading, and the tools in our directory that automate the job, with prices we verified on each vendor's pricing page on 21 October 2026.

Step 1: Write the questions your buyers actually ask

AI engines don't have keywords. They have prompts. So start with a list of 10 to 20 questions a real buyer would type, split into three groups:

• Category questions, where you are not named: "What's the best invoicing software for freelancers in Europe?"
• Comparison questions: "Is [your brand] or [competitor] better for small agencies?"
• Brand questions: "What is [your brand] and how much does it cost?"

Category questions matter most. If you only show up when someone already types your name, AI search isn't bringing you new customers. Brand questions are still worth checking, because they show whether the engines describe your product, pricing and positioning correctly.

Step 2: Ask ChatGPT with web search on

ChatGPT answers some questions from its training data and some by searching the web. For a visibility check you want the search version, because that is where links and sources appear. OpenAI's help center says you can force it: open a conversation, choose Search from the tools menu (or type / and pick Search), then ask your question. Answers that used the web can include citations, and a Sources button lists the pages it used.

For each question, write down three things:

1. Were you mentioned at all, and in what position?
2. Which competitors were mentioned?
3. Which pages were cited as sources?

The third column is the one people skip, and it's the most useful. If ChatGPT keeps citing the same review site or comparison article, that page is shaping the answer. Getting listed or corrected there is often faster than rewriting your own website.

One technical check while you're here: OpenAI says sites become eligible for ChatGPT search results when they allow its crawler, OAI-SearchBot, and when the host or CDN doesn't block OpenAI's search IP addresses. Open your robots.txt and your firewall or CDN bot settings. A blocked crawler is an easy-to-miss, easy-to-fix reason for being absent from its search answers.

Step 3: Ask Perplexity the same questions

Perplexity lists its sources alongside its answers, which makes it the easiest engine to audit. Use the exact same wording as in ChatGPT, and record the same three columns. Differences between engines are normal. What you're looking for are patterns: a competitor who appears everywhere, or one source domain that keeps coming back.

Step 4: Check Google AI Overviews and AI Mode

Search your questions on Google in a private window. Two things from Google's own documentation are worth knowing before you read too much into the result:

• AI Overviews often don't appear. Google says they show only when its systems decide they add something beyond classic results. No AI Overview for a query isn't a failure on your side.
• There's no special trick to get in. To be shown as a supporting link, a page has to be indexed and eligible for a snippet in normal Search. Google says there are no extra technical requirements, no special schema and no "AI text files" needed.

Google also explains that AI Overviews and AI Mode can run a "query fan-out", meaning several related searches behind the scenes. That's why a page can be cited for a question it doesn't literally answer.

For measurement, Google counts clicks and impressions from AI Overviews and AI Mode inside the normal Performance report in Search Console, under the "Web" search type. You can't filter AI traffic out separately there, but you can see whether pages you expect to be cited are getting impressions at all.

Step 5: Put it in a spreadsheet and repeat it

A simple sheet with one row per question per engine is enough: date, engine, question, mentioned (yes/no), position, competitors named, cited sources. Then run the same list again in two to four weeks.

Why manual checks mislead

Doing this by hand is the right first step, but it has real limits:

• Answers vary between runs. Ask the same question twice and you can get a different list of brands. One check is a sample, not a measurement.
• Your account changes the answer. ChatGPT may use memory and approximate location when it rewrites your query into searches, according to OpenAI's help center. Use a logged-out or clean session, and note your location.
• It doesn't scale. 20 questions × 3 engines × weekly checks is 240 answers a month to read and record. Add three competitors and the spreadsheet stops being useful.

That's the point where a tool pays for itself: it runs a fixed prompt list on a schedule, stores every answer, and turns it into share of voice and citation reports.

Tools that automate it, and what they cost

Every price below comes from our pricing table, checked on the vendor's own pricing page on 21 October 2026.

Free: get a baseline first

• HubSpot AI Search Grader is a free one-off check across ChatGPT, Perplexity and Gemini. No account needed. Good for a first number, not for tracking.
• Contentstack Canoe has a free-forever plan: one brand, up to 10 questions, a monthly report on ChatGPT and Gemini. It's the only free option we've found that repeats automatically.
• AthenaHQ has a free Essential plan with 300 credits, where one credit is one AI response. That's enough to test its 11-engine coverage before its $295/mo Starter plan.

Cheapest paid tracking

• Rankscale Essentials is listed at $20/mo, billed yearly, and covers the widest list of engines among the budget tools (ChatGPT, Gemini, Perplexity, Claude, AI Overviews, AI Mode and more). It runs on credits rather than a fixed prompt count. See our Rankscale review.
• Otterly.AI Lite costs $29/mo for 15 prompts across ChatGPT, Google AI Overviews, Perplexity and Copilot. It's the easiest to understand: one prompt, checked on every included engine.
• Ahrefs Brand Radar custom prompts start at $50/mo. Its real strength is the AI Visibility Index ($199/mo), which shows the market without you writing prompts. That suits competitor research more than daily tracking.

Rankscale vs Otterly.AI is the most common budget decision, so we compared them directly: Otterly.AI vs Rankscale.

Our recommendation

1. Run the manual check first. One afternoon, 15 questions, three engines. It tells you whether you have a visibility problem at all, and it often reveals that a single blocked crawler or one outdated review page is the real issue.
2. Set up a free repeat. Contentstack Canoe's free plan or AthenaHQ's free credits give you a second data point without a card.
3. Pay only when you need trends or competitors. For most small teams, $20–$29 a month (Rankscale or Otterly.AI) is enough. The budget tools list and startup picks go further.

The tool doesn't make you visible. It only tells you, reliably, whether the work you do elsewhere is changing the answers.