Businesses whose customers now ask AI first.
Use of AI for local business recommendations rose from 6% to 45% in a year, according to BrightLocal's 2026 consumer survey. The answers your customers get were not submitted to you for approval.
Your patients ask AI about hours, insurance, and what a visit costs.
In a Yext survey of 1,120 US adults this March, 47% had used AI to find a local business in the past month — and among households earning $150k+, AI has overtaken Google as the starting point. For a practice, the exposure is concrete: quoted exam prices, accepted insurance, and hours, answered by a system you don't operate.
The four dental practices in our July 13 sweep came back clean — the AI's stated hours and services matched their sites exactly. That's the honest report, and it's also the point: a clean answer today is not a promise about next month, which is what monitoring is for.
The AI answers for your firm whether or not it knows your firm.
Both of our first documented law findings are stranger than a wrong phone number. Asked about a two-partner immigration practice by name, Google's AI recommended a hospice, a family doctor, and an astrologer instead. Asked about an estate-planning attorney, it advertised — twice, stably — a free 30-minute consultation the attorney has not offered, for a practice that isn't taking new clients.
Courts have already decided whose words these are. In Moffatt v. Air Canada, the airline argued its chatbot was a separate legal entity responsible for its own actions; the tribunal called that a remarkable submission and held the airline liable. The hallucination case record runs to 1,751 court decisions as of July 13, 2026, in Damien Charlotin's public database.
Price-sensitive procedures, quoted by a machine that mixes up your menu.
The clearest finding in our log is a med spa's: Google's AI quoted a 6-session laser package at $399. The spa's own offer page prices it at $1,899 — the AI had cross-wired a laser-facial promotion onto the package. A customer books at one price and hears the real one at the front desk, and your receptionist wears the difference.
This failure is reproducible by anyone — which is also why it's fixable and checkable. The finding, the source page, and the re-scan method are all in the log.
Schedules, service areas, and specials you didn't publish.
The pattern we keep documenting is the AI supplying operational facts nobody published — one law firm in our sweep had not posted a schedule anywhere, and the AI stated one anyway. For a plumber or an HVAC company, the same failure looks like invented service windows, coverage areas, or specials. One Missouri restaurant owner put it publicly: Google AI is telling people specials that do not exist, and the staff wears the anger.
The insurance market has noticed before most owners have. Berkshire Hathaway, Chubb and Travelers sought state approval to exclude AI-related damages from general-liability policies — regulators approved more than 80% of those requests — while HSB, a Munich Re company, now sells small-business AI liability coverage. Doing nothing is quietly getting more expensive.
The numbers above, cited.
Survey figures are vendor research and attributed by name — BrightLocal and Yext sell tools in this market, which is why we name them rather than launder the statistic. Court and insurance items are primary or corroborated records.
- BrightLocal, Local Consumer Review Survey 2026
- Yext, consumer research, March 2026
- First Alert 4 — Stefanina's Pizzeria, Wentzville MO
- McCarthy Tétrault — Moffatt v. Air Canada
- Damien Charlotin, AI Hallucination Cases database
- pymnts — insurers and AI risk
- HSB (Munich Re) — AI liability insurance for small businesses