Every figure below is read from the survey itself, so this page cannot drift from the product. Anything not answered here goes to hello@gaultandco.com.
Networks whose locations are run by different people — insurance brokerages with independent agents, franchise systems, dental and healthcare groups with mixed ownership, real-estate brokerages. Headquarters does not need write access to a single listing, because Gault never asks for one. Networks where corporate does own every listing are welcome and get the same product. Networks of every size are read the same way, and the whole rate schedule is published rather than quoted: the rate per location falls as the network grows. Past 600 locations the setup is done by hand, at the same rates.
Three areas, for every location, from public data only: 17 checks on the Google Business Profile, 29 on the website, and 13 on AI visibility — 59 in all. The three are weighted into the visibility score, and the weighting differs by trade — a dental practice does not weigh the same as an insurance office. Customers see the weighting for their trade inside Gault. An area that cannot be read, such as a site that blocks us or will not load, drops out of the score rather than counting as zero. A location we have confirmed has no website scores 0 for Website, the same as a missing Google profile. Things you might expect to see listed separately sit inside those three: reviews are read as part of the Google profile and again as an AI signal, and a Yelp or BBB entry is verified as corroboration that the location is real. Inside Gault, the method sets out the whole instrument, including where it stops.
Every network is read by Claude, ChatGPT and Meta AI. Gemini is paused until Google permits its search-backed answers to be measured. Gault does not ask an engine about your brand. It asks the question a customer would ask, carrying that location's own city and state, and reads the answer to see whether the location is among the five businesses named. Sixth is not named, and a national figure for a brand is not a reading of a branch.
Yes — 8 of the checks are about them, and they are read twice over. On the Google side: whether the rating clears 4.0, whether there are at least 20 reviews, the bar nearly half of shoppers look for, whether a review has arrived in the last three months, and whether the owner replies to them. On the AI side, rating and volume together are a signal in whether an engine recommends the location at all. What Gault does not do is ask for a review, post one, reply to one, or run a campaign to collect them. It reads what is there and tells you where it stands.
No, and that is the position rather than a limitation. Gault is never given an account, holds no password or access token, and never edits a profile, posts a review, or publishes anything on your behalf. It reads what a customer sees and puts it on the record. That is what lets it grade the work fairly no matter who did it — your own team, your agency, or another vendor — because nothing it finds is shaped by something it is trying to sell you. It runs alongside whatever you use today without touching it.
Automatic re-reading is switched on for your network during onboarding: every location is re-read once a month, and each one carries the date and time it was last read, on its own page, in the roster and on the map. A reading counts as current for 30 days, which is also how the survey decides what is due, so a stale location drops out of "read this month" rather than standing in for a current one.
Yes, on demand, with no per-scan fee. There is a pooled monthly allowance that scales with the network — 3 per location per month, never fewer than 25, so a 100-location network gets about 300 to share — and you will see a heads-up in the dashboard well before you approach it. Your scheduled readings are never affected by it, and if you are in the middle of an unusual push, ask and it goes up.
A live dashboard for the whole network, a per-location page with every finding and its source, and a branded PDF for any location on demand.
You send the change and it is imported. Inside the product you can confirm or reject the listing Gault matched to a location, and correct a wrong website yourself; opening and closing offices goes through a person, on purpose. Nothing moves on its own: a location you closed will not quietly vanish from the readings already taken of it, and a new one is not read until it is on your list. The record is meant to be something you can go back to a year later, which only works if it stops changing behind you.
Per location, per month, with the rate falling as the network grows. Month to month, cancel any time, no setup fee. The current rates are on the pricing page. Agencies and resellers are quoted individually.
Gault is never given an account. It reads what is public, and holds your location list, the readings taken from it, the name and email of each person you give access to, and your logo. Card details stay with Stripe. Your customers' data never enters it, because Gault has no reason to ask for any: it reads the same public surfaces your customers do. There is nothing of yours to hand back when you leave.
Every check carries a graded evidence tier and names the research behind it — Google's own guidance, peer-reviewed work, or large-scale analysis with a disclosed method — and the source sits one click from the finding inside the product. The grades are published rather than implied, including the weak ones: the number of the 59 checks resting on Gault's own reasoning rather than an outside source is 1, and each one is labelled that way wherever it appears. Where a threshold is Gault's calibration rather than a published number, the check says so in the same breath as the advice.
Four things, and they are written down rather than buried. An AI answer changes from one ask to the next, so Gault asks each assistant the main question three times and scores the share of answers that name the location; even so, a move of a few points between monthly readings can be chance, not a trend. Being named in an AI answer and ranking in Google's local results are different things, and neither predicts the other. Answers vary by who is asking: Gault asks on each assistant's current public model, from the location's city, with no account or history, and a real customer signed in to their own account may see something else. And Gault reports — it does not act.
Aidan Bunch, who built Gault, Monday to Friday, 9 to 6 Central. Questions reach the person who wrote the thing rather than a queue, which means the awkward ones get a real answer instead of a deflection.