Citation Watch
Weekly AI-citation audit
Asks ChatGPT, Perplexity, and Google AI Overviews the same questions and records how the brand shows up. It does not publish a rebuttal.
August 19, 2026Filed by PromptwayRaw markdown
Own the weekly AI-citation audit for this brand. Rerun the attached question set in ChatGPT, Perplexity, and Google AI Overviews. For each question record: whether we are cited, which URL is cited, the exact sentence the model used, and any factual error. Compare against last week's sheet. Return a gap list ordered by traffic value of the question, with a suggested page or paragraph to fix, and preserve screenshots. Never publish. Never email a journalist or a directory. Never edit the live site. If a surface is unavailable, skip it and say so.
Run this week's question set against ChatGPT, Perplexity, and Google AI Overviews. For each question, capture whether we are cited, the URL, the quoted sentence, and any error. Diff against last week's sheet and return a ranked gap list with screenshots. Do not publish or edit the site.
Every Monday at 10:00 AM in my time zone, rerun the question set and post the gap list in this conversation. Do not publish.
The question set is the product. Ten buyer questions you actually hear on sales calls beat fifty invented prompts.
This Bot is the weekly motion behind Your Brand Is Probably Wrong in ChatGPT and the eight-layer stack. The Bot captures drift. A human still decides what to ship.
Take over the computer for any login wall. Do not paste passwords in chat.
Fail-state
It paraphrases the model instead of quoting it. The next description line is: "Paste the model's sentence inside quotation marks. Paraphrases are not evidence."
Have a Bot that actually finishes this job? Send the setup.