An AI is about to recommend a real-estate expert in your neighbourhood. It already reads your profile.

Yesterday, ten blue links. Today, one name.

That is the whole change. A buyer used to search, compare, click. Now they ask a question and get an answer. Often a single name. Delivered without hesitation.

That name comes from somewhere. It comes from pages the machine has already read and judged reliable: third-party pages, dated, citing their sources. A Payotte profile is exactly that — and you did nothing to be there.

Here is what our server logs show, across two 24-hour windows:

payotte.com server logs, two continuous 24-hour windows (June 30, 5 a.m. → July 1, 5 a.m., then July 7, 5 a.m. → July 8, 5 a.m.). Facebook link previews and Applebot excluded. See methodology.

AI crawlers read Payotte 4.9× more than Google does. That is an order of magnitude, not a law: from one window to the next the gap runs from 2.9× to 16.4×, depending on whether Google is working or sleeping. But the direction is clear. What is missing from your profile no longer produces an incomplete listing. It produces an incomplete answer, spoken aloud to your future client.

One name per neighbourhood. You are one of 495.

Payotte publishes one expert per neighbourhood and per profession. Never two. There are 640 slots nationwide — 128 neighbourhoods, 5 real-estate professions — and none is for sale. No advertising, no placement, no commission.

We assessed 637 professionals. 631 are published. 495 carry the green seal, reserved for the best-documented files. It is not an award you apply for: it is a place you hold, until a fuller file takes it.

What the machine sees of you

It sees your Google reviews: the rating, the count, and the day we captured them.

It sees your licence. We check it against your regulator's register — we consult 24 of them, from OACIQ to each province's Law Society. 562 profiles out of 631 carry a licence confirmed in a register. Not claimed: confirmed. An AI repeats that kind of detail, because it is dated and attributed.

It sees your years in the trade… when they exist somewhere. On July 8, 2026, we queried Quebec's four professional registers — OACIQ, OEAQ, AIBQ, AMF. None publishes your licence date. To a machine, your twenty years exist only if someone wrote them down. 93 of our profiles have no verifiable start year.

What it does not see

It sees your address. It does not know you sold three houses on the same street last year.

That is local presence: 15 points out of 100, answering the only question a buyer really asks — does this expert actually work here, or serve my neighbourhood from the other end of the province? A Laval broker who "also covers" Rosemont is not a Rosemont broker. The machine sees two addresses and a distance.

Across our 631 published profiles, 195 experts have a confirmed local presence. The other 436 — 69% — lose points. Not because they are absent from their neighbourhood: because we could not prove it.

Zero out of 631

Our profile has a field for the one proof that settles it: how many files you closed in that neighbourhood. Across 631 profiles, that field is filled in 0 times.

Not once. And no one is at fault: this information exists nowhere but with you. No register publishes it. No crawler will find it. We can count your reviews, verify your licence, date your registration. We cannot know how many houses you sold on Beaubien Street.

That is the paradox: machines are about to say your name to a buyer, and the very thing that makes you the expert of this neighbourhood, none of them can find. It fits in a sentence only you can write.

Why now, and not in two years

Because your place is not yours. It belongs to the best-documented file in your neighbourhood. The day a colleague sends us their figures and you do not, they have not "beaten" you: they have simply become, on paper, the better-verified of the two. And it is their name the machine will read.

Because your profile shows your score in public, criterion by criterion. Not in an internal file: on the page. "Reviews 34/35 · Experience 0/30 · Licence 15/15 · Presence 12/15". 629 of our 631 published profiles show at least one criterion below its maximum; 29 display a flat "Experience 0/30". Anyone — a competitor, say — can read exactly where you are leaving points behind.

And because 37 professionals have already understood. They replied, we verified, and 50 profiles now carry the "Recommended" status: confirmed by the person concerned. What no machine could guess about them, they wrote down. It cost them nothing. It took them five minutes.

Four sentences, and the machine sees you

Four pieces of information no register holds. They are enough:

• How many files you closed in that neighbourhood over the past twelve months.
• The address of the office that truly serves it — not the head office.
• The year you began practising.
• Your local ties: chamber of commerce, association, neighbourhood council.

Your profile already shows your score, criterion by criterion, for all to see. You reply, we verify, we date it, we publish. It is free. And it buys nothing: moving from 12 to 15 presence points gains you three points, not a place. We promise no clients, no traffic, no position in an AI answer. One thing only: that what is written about you is accurate.

By the way: how many reviews does it take to be the No.1?

It depends entirely on your profession. Here are Payotte's 495 No.1s, trade by trade.

Site figures computed on July 8, 2026 across 637 assessed experts.

A home inspector becomes No.1 with 215 reviews at the median. A certified appraiser gets there with 32. Not because they work less well: their clients are banks and courts, who leave no Google reviews. We measured it in June, in the Google reviews blind spot.

Look at the last column. It shows how many No.1s have a confirmed local presence. No profession exceeds 41%. That is what this study is about.

What this says about every directory

A ranking built on what is public favours visible professions and buries the rest. Inspectors collect reviews; appraisers have none. Brokers advertise their seniority; the registers withhold it. In the end, the ranking measures visibility, not competence.

We do not escape it. So we publish our blind spots: the review imbalance, corrected by a floor for confirmed B2B professions; local presence, unverifiable without you; and our own coverage gaps. An honest directory does not merely grade others: it publishes what it does not know.

Methodology

• The AI crawler figures. Read by hand from payotte.com's server logs. GoDaddy keeps only a rolling window: no calendar day is complete. Stitching the pieces together gives two genuinely continuous 24-hour windows (traffic in all 24 hours): June 30, 5 a.m. → July 1, 5 a.m., and July 7, 5 a.m. → July 8, 5 a.m. Total: 48 hours. This is not continuous measurement.
• The ratio is volatile — and we say so. Window A: 869 AI requests against 53 from Googlebot (16.4×). Window B: 887 against 305 (2.9×). Googlebot works in bursts. The 4.9× figure is the 48-hour average, not a constant.
• What counts as an "AI crawler." meta-externalagent, ClaudeBot, GPTBot, OAI-SearchBot, ChatGPT-User, PerplexityBot, Amazonbot, Bytespider. Excluded: facebookexternalhit (link preview, no AI) and Applebot (search engine) — including them would add roughly 200 requests. A crawler declares itself through a user-agent string it can falsify.
• What these figures do not say. They show who reads the site. They show neither how often an AI cites an expert, nor what it earns them. Nobody can measure that today, and we do not pretend to.
• Site sample. Payotte publishes only one verified expert per market (city × neighbourhood × profession), the highest-scoring on its 100-point scale. These figures describe "the best-verified expert in each market," never all professionals. Scope: 631 published profiles, 128 neighbourhoods, 54 cities, 8 provinces.
• Scale. Google reviews 35 · experience 30 · active licence 15 · local presence 15 · verifiable bonus 5 = 100. There is no automatic formula: each criterion is a documented judgement whose source and date appear on the profile.
• Reading local presence. 15/15 = confirmed · 12–14 = plausible but qualified (adjacent office, remote service, head-office address) · below 12 = nothing found in public sources.
• Stated limitation. 93 profiles have no verifiable start year. Among them, 38 carry a year of zero while their score reflects thirteen or more years of practice: a data-entry inconsistency, treated here as "unknown" rather than "beginner." Median experience excludes those values.
• Sources. Google reviews and provincial licence registers, compiled and verified by Payotte. Unverifiable field = left blank, never fabricated.

Frequently asked questions