The six-pillar framework that measures how well AI assistants can find, understand, trust and recommend a private clinic. Built by a doctor, for regulated healthcare.
The GROWTH Score™ is a diagnostic created by Dr Andrew Greenland of The AI Growth Clinic. It scores a private clinic from 0 to 100 on how clearly AI assistants (ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews) can identify, cite and recommend it. The score is built from six pillars: Ground Truth, Retrievability, On-page Answers, Wider Web Authority, Trust Signals and Head-to-Head Visibility.
Patients increasingly ask an AI assistant for a recommendation before they ever open Google. The assistant either names your clinic or it names someone else. The GROWTH Score exists to measure that, in a way that is repeatable, evidence-backed and specific to the realities of regulated medicine: register-verified credentials, medical schema, liability-safe content and the directories patients actually trust.
Five pillars measure the inputs AI systems depend on. The sixth measures the outcome. Each is scored 0-100 against a fixed checklist.
| Pillar | The question it answers | Examples of what we check |
|---|---|---|
| G - Ground Truth | Does the clinic exist as one clean, unambiguous entity that an AI can resolve? | MedicalClinic, Physician and Organization schema; NAP (Name, Address, Phone) consistency; Google Business Profile; sameAs links to professional registers |
| R - Retrievability | Can AI crawlers actually reach and read the site? | robots.txt rules for GPTBot, ClaudeBot, PerplexityBot and Google-Extended; sitemap and llms.txt; content rendered without JavaScript; no consent walls blocking crawlers |
| O - On-page Answers | Can an AI lift a clean, quotable answer from your pages? | Answer-shaped opening paragraphs; one page per core service; real FAQ content with FAQPage schema; MedicalProcedure markup; location and service on the same page |
| W - Wider Web Authority | Does the rest of the internet agree about who you are? | Healthcare directories (Doctify, Top Doctors, WhatClinic in the UK; Healthgrades, Zocdoc, Vitals in the US; Jameda, DocPlanner in Europe); independent reviews; editorial and podcast mentions |
| T - Trust Signals | Would a liability-conscious AI feel safe recommending you? | Register-verified practitioner credentials (GMC, GDC, CQC, US state medical boards); named clinical authorship; E-E-A-T markers; review volume and rating |
| H - Head-to-Head Visibility | When real patients ask real questions, are you named? | Patient-language prompts tested live across AI assistants; visibility rate; position; share of voice against named competitors; accuracy of how the AI describes you |
Head-to-Head Visibility is the proof pillar. The other five explain why it is high or low.
Pillar scores are combined into one overall score out of 100 using a fixed weighting that gives most weight to the pillars AI systems lean on hardest for regulated healthcare. The overall score falls into one of five bands.
Two illustrative scorecards showing patterns we commonly see. They are composites for explanation, not individual client results.
The read: a technically sound website that AI can crawl, but thin service pages, a patchy directory footprint and doctors’ credentials that are not linked to the GMC register. Named in fewer than 1 in 5 relevant AI answers.
The read: strong content and credentials, but a cookie wall and JavaScript-only rendering hide key pages from AI crawlers. Fixing Retrievability alone typically unlocks the next band.
| Traditional SEO audit | Generic AEO / GEO audit | GROWTH Score | |
|---|---|---|---|
| Measures | Google rankings and site health | AI mentions, usually sector-agnostic | AI discoverability, trust and recommendation, specific to healthcare |
| Medical credentials | Rarely checked | Rarely checked | Scored: register links, named clinicians, E-E-A-T |
| Medical schema | Sometimes | Generic Organization / LocalBusiness | MedicalClinic, Physician, MedicalProcedure |
| Live AI evidence | No | Sometimes | Always, with verbatim answers |
| Built by | Marketers | Marketers | A medical doctor |
AI assistants are cautious about healthcare. They look for signals that a clinic is real, regulated and competent before they will recommend it, because a bad recommendation carries real risk. Most AI visibility frameworks were built for restaurants, software or trades and treat those signals as optional. In private medicine they are the deciding factor. Dr Andrew Greenland built the GROWTH Score to measure what actually moves AI recommendations in regulated healthcare. Read about his background and credentials.
Free six-pillar Snapshot with live AI answers for your services. Delivered the same working day.