DEV CommunitySunday · July 5, 2026FREE

Checking whether ChatGPT actually recommends your product

The article describes a shift in how users discover products: asking AI chatbots like ChatGPT or Perplexity 'what's the best note-taking app' yields a shortlist of three to five names, making brand presence in those answers critical. The author notes that manually pasting questions into chat UIs does not scale beyond about ten queries and cannot be trended week over week. Instead, they advocate using official APIs with web search capabilities: Perplexity's sonar models, OpenAI's gpt-4o-search-preview, and Gemini's gemini-2.5-flash with Google Search grounding. A single OpenRouter key can cover all three endpoints, simplifying code. For each buyer question, the author records four metrics per engine: whether the brand was mentioned, how early in the answer, whether the domain was cited as a source, and how often competitors appeared (share of voice). They built this as an Apify actor called AI Brand Visibility Tracker, which takes a brand name, domain, competitors, and topics, then generates realistic questions and runs the analysis.