Prompt-Level Tracking
Prompt-level tracking is the practice of monitoring how AI engines answer specific, fixed prompts over time. Every Rankry metric is built from tracked prompts, so the prompt is the unit of measurement in AI visibility. Holding the prompts constant is what turns the natural variation in AI answers into a usable trend.
What this page covers
Section titled “What this page covers”- Why prompts are the unit of measurement
- Why fixed prompts matter
- What Rankry records per prompt
- Core prompts and Lab prompts
Why fixed prompts matter
Section titled “Why fixed prompts matter”AI answers are probabilistic and shift as models and sources change. A one-off check tells you little. Tracking the same prompts on a schedule makes results comparable and lets you tie a change in visibility to a specific page you shipped.
What Rankry records per prompt
Section titled “What Rankry records per prompt”For each prompt and model, Rankry records whether your brand appears, its position, the sentiment, the competitors named, and any cited sources, along with the full response as evidence.
Core prompts and Lab prompts
Section titled “Core prompts and Lab prompts”Core prompts make up your Semantic Core and run with each report. Lab prompts are ad-hoc questions you run on demand or monitor daily or weekly. Use core prompts for your baseline and Lab prompts to test a new page or angle.
Frequently asked questions
Section titled “Frequently asked questions”Why track the same prompts repeatedly?
Section titled “Why track the same prompts repeatedly?”AI answers vary run to run. Repeating fixed prompts turns that variability into a reliable trend.
Which prompts should I track?
Section titled “Which prompts should I track?”The commercial, buyer-intent prompts a prospect would ask before choosing in your category, such as best-of, alternatives, head-to-head, and which-should-I-use questions.
How is prompt tracking related to AI visibility?
Section titled “How is prompt tracking related to AI visibility?”AI visibility is the aggregate. Prompt tracking is the unit. Every visibility, position, and share-of-voice number is computed from individual prompts.