Most AI SEO case studies prove nothing. This template sets out the six proof points a credible case file must contain, from baseline data to an attribution method.
Ranking in ChatGPT means being cited inside its answers, not holding a link. Here are the seven proven GEO tactics that lift AI citations, with the research.
Most ChatGPT SEO advice treats every tactic as equal. This playbook ranks 12 tactics by the effect size each one showed in a published study or controlled test.
AI engines reason over entities, not keywords. This 7-step audit compares your content to a Wikidata or Google Knowledge Graph snapshot to find the entities you are missing.
AI citation rates swing day to day, so naive before-and-after tests mislead. Here is a 6-step statistical framework for sample size, control groups and significance.
We took 10 UK pages, applied one isolated change each, and measured citation rate after 4 weeks. Direct-answer block lifted +19pp; meta-description rewrites produced 0pp. Full ranked list with implementation time.
We mapped 50 UK AI marketing agencies on a 4-quadrant grid: strategy vs technical operations, creative vs analytical. 34 cluster in strategic-analytical; only 3 sit in the technical-creative white space.
We surveyed 100 self-described AI SEO experts on LinkedIn and scored each against 8 capabilities. The median respondent demonstrated 3.7 of 8. Top decile averaged 6.4; bottom decile 1.8. Full distribution.
We assessed 23 UK AI SEO and GEO agencies against a 5-question RFP filter testing non-fakeable capabilities. Only 6 passed all five. The filter cuts a typical long-list by two-thirds.
We synthesised public ChatGPT prompt-leak reporting, peer-reviewed retrieval research and our 50-query AiBoost behaviour test into a single 14-factor map with confidence scores. 5 high-confidence, 6 medium, 3 low.