Prompt volume
Definition: Prompt volume is an estimate of how often people ask AI assistants about a topic or phrase, offered by some AI visibility tools as the AI-search counterpart to keyword search volume.
Also called: AI search volume, prompt search volume, AI prompt volume
Updated
Where the numbers come from
OpenAI, Anthropic, Google and Perplexity do not publish query-level data about what people ask their assistants, in the way Google Ads publishes keyword volumes. Any prompt volume you see is therefore modelled by the tool that shows it. The common methods are:
- Panel data: browser extensions or apps that record a sample of users’ AI conversations, then scale the sample up.
- Keyword translation: taking search volume for a topic and adjusting it by an assumed share of people who ask AI instead.
- First-party usage: counting prompts run inside the vendor’s own product, which mostly reflects its customers’ tracking, not public demand.
Each method has blind spots, and estimates for the same topic can differ several-fold between vendors. Treat any figure as a vendor claim unless the tool documents its panel size and method.
How to use it anyway
- Compare topics against each other within one tool rather than reading absolute numbers.
- Cross-check against keyword volumes for the same topic. If a tool claims a topic is huge in AI but nobody searches for it on Google, ask why.
- Map prompts to topics, not exact phrases. Prompts are long and conversational, and almost no two are identical.
One prompt, many searches
A single prompt can trigger several searches behind the scenes through query fan-out. That means prompt counts and search counts are not interchangeable, and a topic with modest prompt volume can still send many retrieval requests to the pages that cover its sub-questions.
The AI visibility guide explains how to choose a prompt set when volume data is unreliable.