When someone asks AI where to eat tonight, what does it say about you?
Diners have started asking assistants the questions they used to type into a search bar — whether you are open now, whether you take reservations, whether you have anything their vegetarian friend can eat. The assistant answers in a sentence, with no list of links to scroll past and no second opinion. If that sentence has your Sunday hours wrong, the table stays empty and nobody tells you why.
What your customers are asking
Is there anywhere good open near me right now?
Best brunch spot in [neighborhood] that takes reservations
Somewhere quiet for dinner tonight with vegetarian options
Is [your restaurant] open on Mondays?
Every scan asks questions like these on ChatGPT, Gemini and Claude, then compares each answer with your real details.
What goes wrong for restaurants and cafés
Hours are wrong more often than anything else
Restaurant hours change for holidays, for seasons, for a slow Tuesday you decided to close. Assistants tend to be working from whatever was published when their sources were gathered, which can be a year or more stale. A customer told you close at 9 when you now close at 10 is a table you never knew you lost — and unlike a bad review, there is no notification when it happens.
You get compared to whoever the assistant names next
Ask an assistant for dinner nearby and it rarely names one place. It names three. Verity shows you which restaurants come up alongside you, on which assistants, and how often — which is a far more direct read on your local competition than a ranking report, because it is the actual list a diner sees before choosing.
Menu and dietary claims get invented
Assistants summarize, and summaries drift. A café described as having vegan options when the vegan supplier changed last spring sends someone to your door for something you cannot serve them. Every factual claim in a Verity report is checked per assistant, so you see exactly which one is saying it.
The detail that matters most here: hours
Across trades, hours are the single most commonly wrong fact in AI answers, and in food service they are also the most expensive to get wrong — a customer deciding where to eat is deciding for the next thirty minutes, not next week. Verity compares the hours each assistant reports against the ones on your Google Business Profile and flags every mismatch separately, so a correct answer on Gemini does not hide a stale one on ChatGPT.
Illustrative example.
Questions from restaurants and cafés
Read the full FAQDoes this work for a café with no reservations system?
Yes. The check is against your Google Business Profile — hours, address, phone, the attributes you have set — so it works the same whether you take bookings or run entirely on walk-ins. If anything, walk-in businesses are more exposed to wrong hours, because the customer is deciding in the moment.
We have two locations. Is that two scans?
Yes, one credit per location per scan, because they are genuinely different businesses to an assistant: different addresses, different hours, and often different answers. There are no per-location tiers or seats — you buy scan credits and spend them on whichever location you want checked.