# Prompts: the questions you choose to be measured on

> For the complete site index for AI agents, see [llms.txt](https://airpulse.ai/llms.txt). Docs index: [docs.md](https://airpulse.ai/docs.md).

> In the product: Analytics → Visibility → Prompt. Canonical: https://airpulse.ai/docs/prompts

Prompts are the unit of tracking: real questions buyers ask AI engines, grouped by topic, scored per engine and region.

Each prompt row shows its visibility score with a blunt Bad / Excellent label, the regions it runs in, and its execution count. Prompts can be added four ways: manually, by bulk upload, AI-generated from your brand context, or derived from tracked keywords (search volume via a DataForSEO partnership). Once added, analysis runs on the daily schedule without further babysitting. A 0% row is a prompt you are absent from entirely — the sharpest kind of gap.

![Prompt management tab listing prompts under a topic with visibility scores, region flags and execution counts](https://airpulse.ai/docs/visibility-prompts.jpg)

*Visibility → Prompt. Topics group prompts; scores and regions are per prompt.*

> Mix chat-style prompts (how someone talks to ChatGPT) with search-style prompts (how someone triggers a Google AI Overview). And note: no AI engine exposes true per-prompt search volume — keyword volume is the proxy. (Onboarding guidance)

## Terms on this screen

- **Prompt** — A buyer question tracked against the engines, e.g. “Enterprise webinar platform pricing comparison”.
- **Execution** — One prompt asked once, on one engine, in one region. Scores aggregate over executions.
- **Fan-out query** — The sub-questions an engine expands a query into before composing its answer. Content that covers the fan-out set is what survives the engine’s dedup and answer-planning step.