Original research · 2026-07 edition · dated snapshot

AI SEO Statistics 2026

We measured three exact AI API model versions against 547 frozen buyer-intent benchmark questions across 14 industries, then coded each response against a fixed 12-behavior rubric. The results describe this edition's controlled measurement window.

Explore the measured findingsBrowse the 14 industriesOpen Data Explorer

Cross-industry rankings

Where AI behavior diverges the most.

Service-level · 413 services measured

The services AI is most (and least) likely to send to a professional.

We ran the same measurement on 413 individual services. The spread is enormous: some services see AI recommend a professional almost every time, others almost never. Each links to its own measured study.

Measured service register12 evidence rows
Citable dataset
ServiceIndustryHire-a-pro rateSampleQuestion-level disagreement
Charity Nonprofitstudy →Directional panelprofessional88.9%15 questions / 45 responses23.7%
Counselorstudy →Directional panelhealth86.7%15 questions / 45 responses20.4%
Dui Lawyerstudy →Directional panellegal86.7%5 questions / 15 responses21.1%
Restoration Companystudy →Directional panelhome86.6%15 questions / 45 responses18.5%
Personal Trainerstudy →Directional panelfitness84.5%15 questions / 45 responses22.2%
Tailorsstudy →home82.5%40 questions / 120 responses17.8%
Amusement Parksstudy →professional82.5%40 questions / 120 responses26.8%
Dermatologiststudy →Directional panelhealth82.2%15 questions / 45 responses17.4%
Locksmithstudy →Directional panelhome82.2%15 questions / 45 responses25.6%
Yacht Brokerstudy →Directional panelhospitality82.2%15 questions / 45 responses25.2%
Gas Engineersstudy →home81.7%40 questions / 120 responses19.9%
Real Estate Lawstudy →legal81.7%40 questions / 120 responses21.7%

Exact API model versions are listed in each study. Panels below 40 questions are marked directional. Rates describe the measured edition, not a population estimate. Free to cite with attribution.

Where AI least often recommends a professional.

Measured service register12 evidence rows
Citable dataset
ServiceIndustryHire-a-pro rateSampleQuestion-level disagreement
Vocational Schoolstudy →Directional paneleducation2.2%15 questions / 45 responses29.3%
Expert SEO Saasstudy →technology2.5%40 questions / 120 responses15%
Fast Food Restaurantsstudy →hospitality4.2%40 questions / 120 responses18.2%
Hotelstudy →Directional panelhospitality6.7%15 questions / 45 responses24.4%
Cafestudy →Directional panelhospitality6.7%15 questions / 45 responses17.4%
Brewerystudy →Directional panelhospitality6.7%15 questions / 45 responses17%
Private Schoolstudy →Directional paneleducation6.7%15 questions / 45 responses19.6%
Toy Storesstudy →Directional panelecommerce6.7%15 questions / 45 responses23.7%
Community Banksstudy →Directional panelfinancial8.9%15 questions / 45 responses23.7%
Best SEO for Travel Nursing Companystudy →health10%40 questions / 120 responses14%
Internet Cafesstudy →hospitality10.8%40 questions / 120 responses17.8%
Winerystudy →Directional panelhospitality11.1%15 questions / 45 responses24.1%

Exact API model versions are listed in each study. Panels below 40 questions are marked directional. Rates describe the measured edition, not a population estimate. Free to cite with attribution.

The 14 industry studies

Every industry, measured with the same protocol.

AutomotiveChatGPT tells car owners to hire a professional 77% of the time, more than double Gemini's 38%Beauty2% of AI responses name specific beauty providers, indicating models act as educational advisors rather than local directories.Ecommerce65% of Gemini responses name specific ecommerce providers, nearly double the 33% rate seen in ChatGPT.EducationGemini names specific education providers in 58% of answers, versus 0% asking clarifying questions firstFinancial Services65% of Claude responses ask users clarifying questions about their financial situation, compared to 0% from Gemini.FitnessGemini names specific fitness providers 2.3x more often than ChatGPT — 58% of responses versus 25%HealthcareAI models rarely name specific healthcare providers, doing so in only 17% of responses on average.Home ServicesAI models almost never name a specific home services provider, even though a named-provider answer would occur 97.5% of the time under pure consensus modelingHospitalityGemini names specific hospitality providers in 63% of answers, more than triple ChatGPT's rate the model doesn't consistently matchLegalChatGPT tells users to hire a lawyer 92.5% of the time, while Gemini does so just 35% of the time — a 58-point gap on the same legal questionsManufacturingAI models name a specific manufacturing provider in only 0.1 to 0.2 responses per answer on averageProfessional ServicesAI models name a specific professional services provider in only 7% of answers on averageReal EstateChatGPT tells buyers to hire a real estate professional 78% of the time, while Gemini does so just 25% of the timeTechnologyClaude names specific tech providers in 48% of answers, nearly double ChatGPT's 28%
Martial Notarangelo
Research leadMartial NotarangeloFounder, Authority Specialist

This benchmark is researched and maintained in-house. For the underlying data, methodology detail, or an on-record comment for your story, reach the research lead directly.

Request expert commentAbout the research lead