Can you get us the first spot in ChatGPT?

No. Anyone promising that is selling something they cannot control, since there is no ranking to buy and no position to hold. What can be done is making your record readable, then measuring whether your name comes back. Two thirds of the companies the assistants named had that record in place, which is a correlation worth acting on and nothing more than that.

What about llms.txt? Everything we read says to add one.

We measured that specifically and it does not hold up. Among companies the assistants cited, 24.7 percent had an llms.txt file, against 24.4 percent across the registry as a whole. That gap is noise. The file is also generated automatically by some site builders on 94 percent of their sites, which is why it looks common. Adding one costs nothing and changes nothing.

How is any of this different from SEO?

Search ranks pages and sends a click. An assistant reads records and returns a sentence, and that sentence names one or two companies with no list underneath. The work overlaps, the outcome does not: here nobody clicks, they simply hear your name or somebody else's. If what you want is the ranking work itself, that is the site and its SEO agent.

How long before anything changes?

The record is published within days, but the assistants re-read on their own schedule and nobody controls it. Count several weeks before the re-test says anything meaningful. Anyone quoting you a date for this is inventing it.

Task · AI visibility

When an assistant answers, it names somebody. The only question is who.

We ran 18 buying prompts through the assistants your customers use. Every one came back with named companies, phone numbers, even licence numbers. None of those companies got there by asking, and none of them are paying for the spot.

Book a 20 minute call

18 / 18

buying prompts came back naming specific companies rather than categories. Perplexity cited its sources in 6 cases out of 6, fifteen to twenty references each time. The assistant is already answering the question. What is still open is whose record it reads to do it.

LeadLoup measurement, 18 buyer-intent prompts, Quebec construction market, 2026-08-26

Who this page is for, and who it is not for.

Yes, if

People research what you sell before they call you, and that research now starts with an assistant.
Your business already exists online. It is simply not written in a form a machine can read back.
You want the record fixed once and left correct, not a monthly report about how it is doing.

No, if

You want first place in ChatGPT. Nobody can sell you that, and the rest of this page does not pretend otherwise.
There is no site yet for an assistant to read. Build the site first, then this becomes worth doing.
Your customers walk in off the street and never look you up beforehand. None of this changes anything for you.

What an assistant actually reads about you

Not your homepage
An assistant does not browse your site the way a visitor does. It assembles an answer from records it can parse quickly: structured data, directory listings, what other sites state about you as fact. Your homepage wording is close to the last thing it looks at.
The business record
Name, trade, area served, hours, contact, licence. Written as prose, a machine has to guess at all of it. Written as structured data, it reads back exactly. Among the companies the assistants named, 40 percent carried that markup, against 13.7 percent across the full registry we checked.
The prompts, run again
The same buying prompts, before and after. That is the only honest measurement here: not a position in a ranking, but whether your name comes back in the answer your buyer is actually reading.

What you get, and what stays yours

06 deliverables

A record machines read backName, trade, area served, hours, contact and licence, in the format assistants parse.
Service pages marked upEach service stated in the same structured form, not only in prose a machine has to interpret.
Crawler access verifiedrobots.txt read line by line. Sites routinely block the very crawlers they want reading them.
Your buyers' real promptsMeasured from how people actually ask, not guessed from how you would phrase it.
A before and after you can readWhat the assistants said about you on day one, next to what they say once the record is fixed.
The record stays on your domainNothing switches off if we stop working together. It is your property, not an access we rent you.

Two steps, and the first one costs nothing

The audit is free on purpose: somebody in your position buys results, not a diagnosis. You keep the measurement whether or not you go any further with us.

The audit
$0
We run your buying prompts and read your record as a machine reads it. You get what the assistants say about you today.
- Your buying prompts, run across the assistants
- What gets named today, and what gets named instead of you
- Your record read the way a machine reads it
- Yours to keep, with no obligation attached
The fixrecommended
$1,497 CAD
One time. The record gets written, the service pages get marked up, crawler access gets cleared, and the prompts get run again.
- The business record, written and published
- Structured markup across your service pages
- robots.txt corrected where it blocks the wrong crawlers
- The same prompts re-run, before against after
- Everything published on your own domain
The re-test
$197 CADper month
Optional, and only worth it if your services or your area change. The prompts get re-run monthly and the record follows.
- Buying prompts re-run every month
- The record updated as your services change
- What moved, and what did not

Prices in Canadian dollars, taxes extra. No ranking is promised here, because none can honestly be sold: what we measured is a correlation between a readable record and getting cited, not a demonstrated cause.

What people ask us about this

Can you get us the first spot in ChatGPT?: No. Anyone promising that is selling something they cannot control, since there is no ranking to buy and no position to hold. What can be done is making your record readable, then measuring whether your name comes back. Two thirds of the companies the assistants named had that record in place, which is a correlation worth acting on and nothing more than that.
What about llms.txt? Everything we read says to add one.: We measured that specifically and it does not hold up. Among companies the assistants cited, 24.7 percent had an llms.txt file, against 24.4 percent across the registry as a whole. That gap is noise. The file is also generated automatically by some site builders on 94 percent of their sites, which is why it looks common. Adding one costs nothing and changes nothing.
How is any of this different from SEO?: Search ranks pages and sends a click. An assistant reads records and returns a sentence, and that sentence names one or two companies with no list underneath. The work overlaps, the outcome does not: here nobody clicks, they simply hear your name or somebody else's. If what you want is the ranking work itself, that is the site and its SEO agent.
How long before anything changes?: The record is published within days, but the assistants re-read on their own schedule and nobody controls it. Count several weeks before the re-test says anything meaningful. Anyone quoting you a date for this is inventing it.

We run your prompts first. You read the answers before deciding anything.

Book a 20 minute call

Twenty minutes, no commitment. You leave knowing what the assistants say about you today, even if we never work together.