Every gap between you and an AI recommendation.
Ranked.
We read your site the way an AI engine does, score what we find across 11 categories, and rank every gap by how much closing it moves your number.
Curious whether your name is already on AI’s shortlist?
See if AI already recommends youSearch existing citation records. Free. No account needed.
This is the report.
Not a picture of one. Every panel below is rendered by the same code your report runs on, against a worked example. The notes explain what each part tells you.
A number you can act on
Zero to 100, weighted across 11 categories. Under 25 reads poor, 25–49 needs work, 50–74 good, 75 and up excellent. It is a prediction from public data, not a verdict — we say so on the report itself.
Your public footprint.
The same thing AI reads.
Where the score came from
We read your site, your Google Business Profile, and your Foursquare listing — the same public footprint the engines look at. Every number above is traceable to one of the three, so nothing in the score is a black box.
Every signal,
every category.
4 more off-site categories — AI Engine Visibility, Knowledge Depth, Transaction Authority, Review Strength — worth 30 pts. They need an account to measure, so they aren't scored here.
Every finding, in fix order
Expand any category to see exactly what we found and why it matters — named problems on your site, not a generic checklist. Red is broken, amber is weak, green is already working.
What fixing it is worth
The target is arithmetic, not a promise: we add up the points still available in every category holding a fixable red or amber finding. Close them all and the sample moves 24 → 62.
Whether that number is bad
A score on its own means nothing. We plot you against other agents scanned in your metro, so you know whether a 24 is normal or a problem. The comparison appears once enough agents in your area have been scanned.
Where the engines get their answers.
Ranked from 48 real AI-engine probes in your metro — these are the sources the models actually cited. Where you're marked not detected, their answers are being built without you.
| # | Source | Type | Cited by | Citations | You |
|---|---|---|---|---|---|
| 1 | zillow.com | Portal | GPT · CLD · GEM | 61 | verified |
| 2 | realtor.com | Portal | GPT · GRK | 44 | not detected |
| 3 | yelp.com | Directory | GPT · GEM | 29 | not detected |
| 4 | redfin.com | Portal | CLD | 17 | couldn't check |
| 5 | linkedin.com | Social | GPT | 9 | present |
The sources your market runs on
We probe the engines with real buyer questions for the sample metro and rank the sources they actually cite. Every row marked not detected is a place their answers are being assembled without you — and when two sources disagree about your contact details, that conflict is shown, not smoothed over.
Who the engines recommend in your market.
Mined from real AI-engine answers about your market — these are the names the models actually give out, with their public Google Business numbers.
| # | Agent | Named by | Mentions | Google Business |
|---|---|---|---|---|
| 1 | Jordan Example Example Realty Co | GPT · CLD · GEM | 23 | 4.9★ · 87 reviews |
| 2 | Casey Sample Sample & Partners | GPT · GRK | 17 | 4.7★ · 42 reviews |
| 3 | Riley Placeholder | GEM | 11 | — |
| 4 | Avery Demo Demo Homes Group | CLD | 8 | 4.5★ · 19 reviews |
| 5 | Quinn Fictional Fictional Properties | — | 6 | — |
| → | You | not yet measured | — | 4.6★ · 11 reviews |
Who the engines already recommend
The names the engines return for the sample market, ranked by how often they come up. (These are fictional names — your report shows the real ones.) An em-dash means we could not verify a matching Google Business listing, and we say so rather than guessing.
The part that needs an account
What each engine actually says about you — the verbatim answer and the sources it cited — stays locked on the free report. Those probes cost real money to run and only work once you have told us how buyers in your niche ask.
11 categories. Every weight published.
A score you can’t audit is just an opinion with a number on it. Here is the entire scoring model — what each category is worth, and what it is looking for.
Earned by changes to your own site. The Site Fix Pack gives you the exact change for each.
Whether your site publishes machine-readable facts about you — your name, service area, credentials, and reviews — in a format engines can actually parse.
Whether AI crawlers can read your site at all. A site that renders entirely in the browser looks like a blank page to them.
Whether your writing is shaped the way engines quote from — direct answers, concrete numbers, real depth.
Whether your name, address, and phone agree everywhere they appear. Disagreement makes an engine less confident it is looking at one person.
The basics that signal a site worth citing — HTTPS, mobile support, load speed, reachable contact details.
Earned everywhere else — listings, profiles, reviews, and the engines themselves. The Site Fix Pack covers the profile-side fixes here too; the marked categories need an account before we can measure them.
How many of the four engines name you when asked for an agent in your market. Scored from a live probe run, so it needs an account.
How much corroboration exists for you elsewhere on the web — profiles, listings, and links an engine can cross-reference.
Whether your Google Business Profile exists, is complete, and carries the reviews and photos engines read.
How much specific local knowledge you have published — the neighborhood detail no engine can infer from a homepage.
Your closed transaction record, published in a form an engine can verify rather than a claim on an About page.
The count and rating of previously recorded testimonials. Google review samples are displayed separately and do not recalculate these saved points.
A work list, not a grade.
Nobody needs a tool to tell them they scored badly. The report is worth running because it turns the number into a queue of specific work, ordered by what each item is actually worth.
- 01
Every check returns a finding, not a grade
Each category reports what we actually found on your site — the missing schema type, the page that took five seconds to load, the listing nobody claimed. Named problems, with the reason each one matters to an engine.
- 02
Findings sort by what they are worth
A missing structured-data block and a missing sitemap link are not the same size of problem. The list leads with whatever is holding the most points, so the first thing you fix is the thing that moves you most.
- 03
The target is arithmetic you can check
We add up the points still available in every category holding a fixable red or amber finding — diagnostics we could not verify don’t count. That sum is the target — not a projection, not a promise about leads. If you close everything on the list, that is the score you get.
A score means nothing on its own.
Is a 24 bad? You cannot answer that without knowing what everyone else in your market scored. The report puts you in the distribution — the median, your percentile, and where you actually sit on the curve.
That is also the difference between a number that stings and a number you can use. Being twelve points under your metro’s median tells you how much ground there is to make up.
Your metro first
We compare you against other agents scanned in your market, because a score that beats the national average can still be last on your street.
It needs 30 scans to appear
Until 30 agents in your area have been scanned there is no distribution to sit you in, so we widen to your state, then nationally, rather than invent one.
Nobody can see you in it
The comparison stores a score and a market. No name, no site, nothing that identifies you — and you are never counted in your own distribution.
And this is the file that fixes it.
The report names what is wrong. The Site Fix Pack is one file that says what to do about each one — written for your site, with the exact change, and a way to check it worked. Hand it to your developer, work through it with an AI coding tool, or paste it into ChatGPT.
Problem. Your site builds its content in the browser. AI crawlers do not run JavaScript, so what they receive is an empty page — whatever your visitors see, the engines see nothing.
What the scan saw. yoursite.com · 4 pages read · 0 characters of readable text
Do this.
- 1.Confirm what a crawler actually receives: curl -sS https://yoursite.com ⋯
- 2.Pick a rendering strategy that emits real HTML ⋯
Verify. ⋯
Abbreviated for the web. The file you get is complete, written for your site, and names every issue we found on it.
Ordered by what it costs you
The most damaging problems come first. A site AI cannot read at all outranks a missing meta description, and the file is sequenced so the first thing you do is the thing that matters most.
The finding, in full
Not a category name and a score — the actual problem, in plain language, with what we saw on your site when we looked. This part is in your free report too.
The exact change, step by step
What to do, in order, with the code where there is code. Written so it can be handed to a developer, worked through with an AI coding tool, or followed by you.
How to check it worked
Every fix ends with a way to confirm it landed that does not involve us. You should not have to take our word for it.
Your report is free and tells you everything that is wrong. The Site Fix Pack comes with Visibility.
What costs nothing, and what doesn’t.
Worth knowing before you start, not after. The whole diagnosis is free — every finding, ranked, with nothing held back. What costs money is the part that fixes it, and the part where we ask the engines about you directly.
- Your score, 0–100
Weighted across every category we can read without your help.
- Every finding, ranked
What is broken, what is weak, what already works — in fix order.
- Your target
The points still on the table, added up.
- Your metro cohort
The median, your percentile, and where you sit on the curve.
- What we read
Your site, your Google Business Profile, your Foursquare listing.
- A link you can keep
The report lives at its own URL. Share it or come back to it.
We ask ChatGPT, Claude, Gemini, and Grok for an agent in your market and show you the answer word for word, with every source they cited and whether you were one of them.
One file with every fixable problem we found — on your site and your public profiles — the exact change for each, the code to paste, where it goes on your platform, and how to check it worked. Hand it to a developer, drop it into a coding tool, or paste it into ChatGPT. $49 per month.
All of them live off your site. They need either a live probe or facts about you that aren’t published anywhere we can read.
- AI Engine Visibility · 5 pts
- Knowledge Depth · 10 pts
- Transaction Authority · 10 pts
- Review Strength · 5 pts
Why this part isn’t free: every probe costs real money to run, and the answers only mean something when the questions match how buyers in your niche actually ask — which is the part only you can tell us. The Site Fix Pack is the work of turning a list of problems into something you can actually hand to someone.
Find out what’s actually in your way.
One URL, and you get the score, the findings, and the fix order back. Nothing to install, nothing to cancel.