Every point in your score is traceable to a specific data signal from a specific source. There is no black box. Below is the exact formula — the same one our audit engine runs.
Google Business Profile (GBP)
25 ptsReviews & Reputation
22 ptsWebsite & Schema Markup
20 ptsLocal Search Presence
15 ptsAI Citations
18 ptsWe run three search queries — one per template — using your business type and city. Each query is sent to Google (via SerpApi) and Perplexity simultaneously. Your score is the best result across all queries (not an average), so a single strong result can unlock full points.
Example queries for a Milton physiotherapy clinic
best physiotherapy clinic in Milton, Ontario
physiotherapy clinic near Milton
top physiotherapy clinic Milton Ontario
Your GBP listing is the single most important signal for local AI search. We check whether Google's Knowledge Graph recognises your business by name, whether your category and contact details are present, and whether your rating appears.
| Signal | Points |
|---|---|
| Business found in Google Knowledge Graph | +10 |
| Star rating present | +5 |
| Business category / type present | +5 |
| Website or phone number on the listing | +5 |
Data source: SerpApi — Google Knowledge Graph (knowledge_graph)
AI engines use review volume and rating as trust signals. A business with 50+ reviews at 4.5★ is far more likely to be cited than one with 5 reviews at 3.9★.
| Signal | Points |
|---|---|
| 50 or more reviews | +12 |
| 10 – 49 reviews | +6 |
| Rating ≥ 4.5★ | +10 |
| Rating ≥ 4.0★ | +5 |
Review count and rating tiers do not stack beyond their category maximum.
Data source: SerpApi — Knowledge Graph (review_count, rating) with local pack fallback
AI engines parse websites for structured data (JSON-LD schema). A LocalBusiness schema tells AI exactly who you are, where you are, and what you do. FAQ schema provides ready-made Q&A pairs that AI cites verbatim.
| Signal | Points |
|---|---|
| Website is reachable (HTTP 200) | +8 |
| LocalBusiness JSON-LD schema detected | +6 |
| FAQ / HowTo JSON-LD schema detected | +6 |
Data source: Direct HTTP fetch of the business website + HTML parsing
The Google local pack (the map + top 3 results) is where most local searches begin. Being there means Google has indexed you as a credible local option. Appearing in organic results in addition to the local pack doubles the exposure.
| Signal | Points |
|---|---|
| Business appears in Google local pack | +10 |
| Business appears in organic search results | +5 |
Data source: SerpApi — Google local_results and organic_results
We test whether AI search engines mention your business by name when a potential customer asks about your category and city. We check five engines and weight them by reach and signal quality: ChatGPT and Google AI Overview have the widest local reach, Gemini is the #2 assistant (checked live via Google Search grounding), Perplexity is live but smaller, and Claude reflects model recall.
| Signal | Points |
|---|---|
| Mentioned by name in a ChatGPT answer (live web search) | +6 |
| Mentioned by name in a Google AI Overview answer | +5 |
| Mentioned by name in a Gemini answer (Google Search grounding) | +4 |
| Mentioned by name in a Perplexity AI answer | +2 |
| Mentioned by name in a Claude answer | +1 |
Data source: OpenAI (ChatGPT web search) + SerpApi Google AI Overview + Gemini (Google Search grounding) + Perplexity Sonar API + Anthropic Claude
Why is the max score 100 and not something else?
We calibrated the five pillars so that a business that does everything well — claimed GBP, 50+ reviews at 4.5★, reachable website with schema, in the local pack, cited by two AI engines — scores 100. Most businesses start between 20 and 50.
Does the score change if I do nothing?
It can — if a competitor gets more reviews, or if Google updates how it indexes your listing, your relative visibility changes. That's why monthly auto-audits matter.
Which AI engines do you check, and are they weighted equally?
Five: ChatGPT, Google AI Overview, Gemini, Perplexity and Claude. They are weighted by reach and signal quality — ChatGPT (6) and Google AI Overview (5) have the widest local reach, Gemini (4) is checked live via Google Search grounding, Perplexity (2) is live but smaller, and Claude (1) reflects model recall rather than live browsing.
Can I dispute my score?
Yes — every audit stores the raw data that produced it. Use the "Why this score?" button on your dashboard to see exactly what each engine returned for each query.
LeapOne AEO Methodology · Last updated May 2026