A patient just asked ChatGPT which dentist near them takes their insurance. Did it get your answer right?
Patients ask AI before they call: who takes their plan, who is open Saturday, who sees kids, who can fit in a cracked tooth today. The engines answer location by location — and when one says your Round Rock office is closed weekends, the patient books somewhere else. Leolens checks what every engine says about each of your locations against the facts on your own pages, quotes the words back to you, and writes the corrections.
First report within 48 hours of setup · every question sampled 3× per engine · checked against the facts you declare
What we check, at every location
Insurance networks — which plans each location accepts, against what the engines tell a patient who asks "does this dentist take Delta Dental?"
Hours, per location — including Saturday and evening hours, office by office rather than one set for the group.
Providers and specialties — who practices where: pediatric dentistry, orthodontics, oral surgery, sedation, Invisalign.
Emergency and same-day care — whether AI says a location can see a patient in pain today.
New-patient acceptance — whether AI tells people you are taking new patients, and at which locations.
Parking — what a patient is told about getting to the office.
And the basics — address, phone number and languages spoken. Anything you have not declared is reported as not measured, never guessed.
What a finding looks like
Fernhollow Dental is a composite example — not a client.
Claude"does Fernhollow Dental in Lakeline see kids?"
"…Fernhollow Dental's Lakeline office treats adult patients only; for children, families usually look elsewhere…"
Your Lakeline page lists: pediatric dentistry, Tuesdays and Thursdays
1 of 4 engines had this wrong, in 2 of its 3 sampled answers. Checked against fernhollowdental.example/locations/lakeline.
What you get back
Every wrong fact, quoted — the engine's words verbatim, the question it was asked, the location, and the declared fact it contradicts.
Facts wrong at more than one location, first — an insurance plan misreported at four offices is one finding, ranked above a one-off.
The corrections, written for you — listing text, per-location structured data and an FAQ answer in the words patients use, each tied to the finding that prompted it.
Where you show up — how often each engine names each location, as a supporting section rather than the headline.
After the first report
Every cycle, your program re-checks every location against the facts you declared. Findings keep their dates, so "still wrong three months later" and "fixed and staying fixed" are both on the record. From $799 per month for a single location, cancel anytime. Not sure yet? Take a month, read the report, and cancel from the billing portal if it is not for you.