Infrasity
GEO Dashboard

Prompt Detail

A full-page view of a single prompt: who's cited, why you're winning or losing, and what to do next.

Click any prompt in All Prompts to open its detail page. It brings together everything known about one prompt: your citation rate, the domains that get cited instead of you, a diagnosis, a suggested fix, and the raw answers from each AI engine.

Prompt detail page with the prompt title, four stat tiles, why-you-are-losing and suggested-fix panels, and a who-gets-cited table
Prompt Detail: header actions, key stats, diagnosis, and who gets cited.

The header shows the breadcrumb (GEO Prompts › cluster), tags for the prompt's status (such as Owned (#1)) and type (such as List), and the full prompt text. Three actions sit on the right:

  • Copy prompt: copies the prompt text.
  • Add to sprint: creates a topic from this prompt in a sprint you choose. After it's added, an Added to sprint tag appears.
  • Create brief: starts a content brief for this prompt.

Key stats

  • Citation rate: the share of answers that cite your domain (e.g. "33 of 33 answers cite infrasity.com").
  • Average AI position: your average rank across the tracked engines, with change versus the last scan.
  • Share of citations: your share of all citations for this prompt.
  • Answer leader: the domain cited most often for this prompt.

Why you are losing this prompt, and the suggested fix

Two panels diagnose the prompt. Why you are losing this prompt lists the issues found. Suggested fix recommends what to change. When you already lead the answer, both say so ("nothing to fix here").

Who gets cited in this answer

A ranked table of the domains cited for this prompt in the selected window, with a bar, citation count, and share. Your domain is tagged You. Toggle between Domain and URLs to see either the sites or the specific pages.

Position, last two runs

Your position in the previous run and the latest one, so you can see at a glance whether you moved.

Same topic, same problem

Other prompts in the same cluster with their citation rate, position, and rank change, with the current prompt marked This prompt. Use it to spot whether a problem is isolated or cluster-wide.

Model results

Lower on the page, Model results lists each engine's run history and the AI agent's answer.

  • Select a run on the left to read the full answer and its citations.
  • Copy answer copies the response text.
  • Web Search & Grounding Analysis shows the sources the engine used; View sources expands them.
  • Brand Inspector highlights where your brand appears in the answer.
  • Runs where you weren't cited are marked Not cited.

Sample data is labelled

If a section is showing placeholder data instead of real runs, it carries a Sample data tag. Don't use those numbers for reporting.

LLM usage: llms.txt

On this page