Turn an AI-visibility gap into a cited page

A gap prompt is a question your buyers ask where AI engines mentioned you in zero measured runs — not low visibility, none. Generic content advice won't close it, because every AI answer is already citing someone else for a reason. This playbook uses CrunchJunkie's Content Brief to read that reason out of your own scan evidence, write the page the answer engines are missing, and prove the result.

1. 1

   Find your gap prompts

   Open the project's Prompts tab. A gap prompt has runs but 0% visibility — the brand absent from every measured answer. The Brief button appears exactly on those rows, and the gap count in the header tells you the size of the opportunity. Start with the gap closest to buying intent, not the biggest one.

2. 2

   Generate the brief from evidence

   Click Brief. Crunch assembles the evidence first: the real AI answers for this prompt, the competitor pages those answers cited, and each winner's extracted GEO signals — direct quotations, statistics with named sources, authoritative outbound links. The brief is hard-constrained to that evidence: nothing invented, and an honest "not enough signal" when the data is thin.

3. 3

   Read why the winners win

   Before writing anything, look at the winner pages' signals. Peer-reviewed GEO research (Aggarwal et al., KDD 2024) measured that quotations, sourced statistics and authoritative citations each lift a page's chance of being cited by roughly a quarter — and the winners usually have them. Your brief tells you which of those signals the winning pages carry that yours don't.

4. 4

   Write the page the brief describes

   Follow the brief's structure: answer the prompt's question directly, add real quotations and statistics with named sources, link to genuinely authoritative references, and mark it up with the schema the brief recommends. Never fabricate a quote or a figure to tick a box — stuffed or fake signals are the one tactic the research shows actively hurts.

5. 5

   Publish, then measure the impact

   Once the page is live and indexable, open the brief again and click Measure impact. CrunchJunkie re-scans that single prompt and compares only fresh runs against your recorded 0% baseline — so the lift you see is real, current evidence, not an average diluted by history. No lift yet? Engines re-crawl on their own schedule; re-measure in a week or two.

6. 6

   Scale it with batch briefing

   When one gap is closed, the same loop applies to all of them: batch briefing generates a grounded brief for every gap prompt in one pass, each saved next to its prompt. Work through them in intent order, and let the per-prompt measurements build your case study as you go.