Prompt Volume

Prompt volume refers to how many times you ask a given buying question, say, who installs tankless water heaters near me, across an AI engine when you're trying to measure how that engine answers. One prompt is an anecdote. A meaningful measurement asks the same question repeatedly, and often several close variations of it, because the same prompt run twice can come back with a different answer.

Volume also has to be split across engines rather than pooled into one number. A prompt run 50 times on ChatGPT tells you about ChatGPT. It tells you nothing reliable about Perplexity or Gemini, because Profound's research found ChatGPT and Perplexity overlap on only 11% of the domains they cite. Real measurement means running volume per engine, not once across a blended average.

This is also why a lot of the AI-visibility numbers you'll see quoted online (including in some of the vendor research this glossary cites) are more like a directional signal than a scientific result. Prompt design, sample size, and which cities or trades get asked about all shape the outcome heavily, and few published studies disclose their full methodology.

Tibly's approach is to generate the buying questions for your specific trade and service-area cities, then ask each one repeatedly across ChatGPT, Gemini, Perplexity, Grok, and Google's AI Overviews on a daily cadence, so the prompt volume behind your visibility score is high enough, and repeated often enough, to show a real pattern rather than a single lucky or unlucky answer.