DEV Community

SureSignal
SureSignal

Posted on

How to check whether ChatGPT and Perplexity recommend your brand, and track changes weekly

#ai #llm #seo #marketing

How to check whether ChatGPT and Perplexity recommend your brand, and track changes weekly

More and more buyers ask an AI assistant "what's the best X for Y?" before they ever see a search results page. If your brand isn't in that answer, or a competitor is listed first, you'd want to know. You'd also want to know when it changes.

Step 1: Pick the prompts that matter

Start with 10–30 questions a real buyer would ask:

  • category: "best CRM for small business"
  • alternatives: "HubSpot alternatives"
  • comparison: "Pipedrive vs HubSpot"
  • problem: "how do I track sales emails automatically"

Use your own sales calls, support tickets and Search Console queries as sources. Avoid prompts that contain your brand name, unless you're checking how AI describes you.

Step 2: Do it by hand once

Ask each engine and note whether you're mentioned, at what position relative to competitors, and whether your site is cited. You'll quickly see the problem: answers vary from run to run, and engines differ from each other. One manual check is a sample, not a measurement.

Step 3: Automate the checks

The AI Search Visibility Tracker on Apify asks each prompt through each provider's official search-enabled model API: OpenAI search models, Perplexity Sonar, and Gemini and Claude with web search. It doesn't scrape chatgpt.com or perplexity.ai.

{
  "brand": "YourBrand",
  "brandDomain": "yourbrand.com",
  "competitors": ["competitor-one.com", "competitor-two.com", "Competitor Three"],
  "prompts": ["best CRM for small business", "HubSpot alternatives"],
  "engines": ["perplexity", "chatgpt"],
  "samplesPerPrompt": 2,
  "onlyChanges": true,
  "trackerName": "crm-weekly"
}
Enter fullscreen mode Exit fullscreen mode

For every prompt and engine you get brandMentioned, brandRank, brandCited, shareOfVoice, competitorsMentioned, citedDomains and the full answer text. You also get change fields: mentionChanged, rankChange, newCitedDomains, changed.

samplesPerPrompt asks the same question several times. AI answers vary, so averaging a few samples gives a steadier picture than one roll of the dice.

Step 4: Schedule weekly, and read only what moved

Schedule the task weekly and turn on Only output changes. Every check is a fresh, live answer, so the data is never a stale copy. Your dataset and Slack alert only show prompts where something moved: mention gained or lost, rank change, a new cited domain, a new competitor.

Pricing, plainly: $0.15 per answer checked, including unchanged checks that are filtered out of your output. Model usage is billed at cost to your Apify account through Apify's OpenRouter integration (the README estimates about $0.001–$0.01 per answer), or you can bring your own OpenRouter key. Errors aren't charged. Example: 20 prompts × 2 engines × 1 sample, weekly = 40 checks ≈ $6 a week plus model usage.

Step 5: Act on it

  • citedDomains tells you where AI engines get their information. If a competitor wins because of a G2 page or a comparison article, that's your content or PR to-do list.
  • Lost a mention? Check whether a newly cited domain replaced yours.
  • Re-run after publishing new content and watch rankChange.

Honest limits

  • No tool can see a specific user's personalized ChatGPT session. This measures the same question through official APIs, which is reproducible and terms-friendly but not identical to the consumer app.
  • Answers vary, so use samples and look at trends, not single checks.
  • No guarantees: tracking visibility doesn't create it.

A decision is only as good as the information behind it. "I think ChatGPT mentions us" isn't a strategy. A weekly record of what changed is.

SureSignal: only pay for new data.

Try it: AI Search Visibility Tracker on the Apify Store. $0.15 per AI answer check; every check is billed, and 'Only output changes' filters your report to what moved.

Top comments (0)