The numbers and methodology behind AI search.
Original research, definitive guides, and primary methodology on Answer Engine Optimization, AI search behavior, and the era after Google. Open methodology. Free to cite, link, or quote with attribution. Updated as the engines update.
The 2026 State of Answer Engine Optimization Across Six Industries
An open-methodology study on how ChatGPT, Claude, Perplexity, and Gemini name brands across six high-value verticals — and what separates the brands that get cited from the ones that get ignored. 1,200 prompts tested. Six headline findings. Vertical-by-vertical breakdowns. Full methodology disclosed.
Read the full studyThe reference material.
The Complete Guide to AEO
Answer Engine Optimization explained from first principles — what AEO is, why it's distinct from SEO, the eight technical factors that matter, and how to deploy them this quarter.
MethodologyHow to Map Buyer Prompts for AI Search
Keyword research is dead for AI search. Buyer prompt mapping replaces it. The four prompt tiers, the verification step, the scoring formula, and what a finished prompt map looks like.
Technical PlaybookSchema Saturation for AEO — The Playbook
Which schema.org types matter for AEO, which are vanity, how to wire @id references correctly, and how to validate without losing a week to debugging.
Comparative AnalysisHow ChatGPT, Claude, Perplexity, and Gemini Differ
The four major AI engines have meaningfully different selection biases. Claude favors editorial press. Perplexity favors original publishers and Reddit. ChatGPT favors structural clarity. Gemini follows Google.
Open methodology. Free to cite. Designed to be challenged.
Most agency research is marketing dressed as research. Numbers cherry-picked, methodologies hidden, claims unfalsifiable. KAK Growth research goes in the opposite direction.
Every study we publish discloses the full methodology — sample sizes, prompt lists, test conditions, limitations. The raw data is available on request to journalists, researchers, and competitors. We invite challenge and replication.
This is partly principled and partly strategic. Open methodology builds the kind of authority AI engines actually weight. Brands that publish replicable, citable research become the brands AI engines treat as primary sources. The compound matters.
Our research is published under a permissive license: quote, cite, link, or reference with attribution. We encourage other researchers to extend or correct what we've published. The goal is collective progress on a discipline that's still being mapped.
What we're working on next.
Quarterly: 2026 Q3 State of AEO
The next iteration of our flagship study. Expanded prompt sample (n ≈ 2,400), three additional verticals, paired before/after testing on the same brands. Targeting publication July 2026.
The Citation Drift Study
Tracking how AI engine citations change month-over-month for the same prompts. Documenting how often cited brands hold their position, and what triggers brand displacement. Targeting first-edition publication August 2026.
The AEO Implementation Cost Study
Honest economics of running a full AEO program — from in-house team configurations to retainer-agency benchmarks. Sample of 30+ brands across our six verticals. Targeting publication October 2026.
The fastest way through the research is the audit.
Reading is useful. Running the diagnostic on your actual brand is faster. The Free AI Visibility Audit gives you industry-specific data within five business days using the same methodology as our research.