Prompt Tracking

Prompt tracking is the foundation of every Rankry metric. Rankry holds a fixed set of buyer-intent prompts, runs them through each AI engine on a schedule, and structures the answers into data, so your visibility, position, and share of voice are all built from the same repeatable measurements.

What Rankry measures

For each tracked prompt and each model, Rankry records presence, position, sentiment, competitors named, and any cited sources, then stores the exact response text. Holding the prompts constant is what turns the inherent run-to-run variability of AI answers into a meaningful trend.

Fields & metrics

FieldWhat it holds
PromptThe exact buyer-intent question being tracked.
CategoryThe buyer-intent bucket the prompt belongs to.
Per-model resultPresence, position, and sentiment for each engine.
Competitors namedRival brands that appeared in the answer.
Raw responseThe full AI answer kept as evidence.

Example use case

Before a launch, a team adds the prompt "best [category] tool with API access" to tracking. Two weeks after publishing a developer page, that prompt flips from naming a competitor to naming them first on three of five engines, and the raw responses show why.

Related

FAQ

Prompt tracking questions

Can I add my own prompts?
Yes. Rankry generates a tailored set for your category and lets you add and edit your own in Prompt Lab, so you can track the exact questions your buyers ask.
Why keep the raw AI responses?
Because the response is the evidence. Every score links back to the actual answer it came from, so you can verify and understand any number rather than trusting a black box.
How many prompts should I track?
Enough to cover your main buyer-intent categories. Rankry starts you with a generated set and you expand from there based on the decisions you care about.