Signal 01
Your server logs
Which AI crawler read which page, and when. First-party, on your infrastructure, and impossible to fake.
ServicesWhat we are best at
Every question your buyers ask is an enquiry going somewhere.
When someone asks an assistant who to call in your town, it names a handful of firms it can read, extract and trust, and ignores everyone else. There is no page two, so every one of those questions is an enquiry that goes to somebody, and right now it is not going to you. This is the work that puts you in that handful, and it is the thing we can prove we are good at.
What this is worth, and who measured it
+8.4%
A 0.1 second improvement in load speed lifted retail conversion by 8.4% and average order value by 9.2%. Same traffic, same spend, more revenue.
These are findings from published research on the wider market, measured by the people named beside each one. They are not results we are promising you, and we do not have client results of our own to show yet, which we would rather say plainly than dress up.
Three sizes, and what moves the number
Small
most firms$1,800 to $4,200
usually two to three weeks
The foundation, and it is the half that decides whether anything else matters. Crawler access unblocked, structured data, llms.txt, indexation, and the pages that earn you money rewritten so a machine can lift a clean answer from them.
Medium
$4,200 to $9,500
usually four to six weeks
The foundation plus the thing worth quoting: the dataset, comparison or tool in your category that nobody has published. That is what earns links from sources the models already trust, and it is what turns you into the answer rather than a candidate.
Large
from $9,500
quoted after a paid scoping week
A large or multi-location estate, several practice areas each needing its own answer, or a migration where the existing search equity is worth real money. Scoped properly first, because guessing at these is how a firm loses rankings it already had.
What moves the price
- What state the site is in. A modern server-rendered site is quick; one that assembles itself in the browser may need the build instead
- How many pages genuinely matter, because restructuring twelve money pages is the work and the other two hundred usually are not
- Whether a citation asset already exists or has to be created from raw data
- How competitive your category already is in AI answers, which we can tell you for free before you commit
If you are not sure which size you are, the free twenty-minute call will place you within about a thousand dollars, and it costs nothing to find out.
How the work runs
Week 1
Find out what the machines actually see
Your site gets fetched as GPTBot and PerplexityBot, so the report describes what those crawlers are really served rather than what your CMS claims. Your real questions go to the real assistants and we record who gets named instead of you. Almost nobody tests the first half, and it is where the ugliest surprises live.
- Blocked crawlers are the single most common finding, and usually nobody at the firm knows it is happening.
Weeks 2 to 3
Fix the foundation
Crawler access, structured data on every page, llms.txt, a dynamic sitemap, indexation, and clean server-rendered HTML where the page currently assembles itself in the browser. This is the unglamorous half and it is the half that decides whether anything else matters.
Weeks 3 to 4
Make the pages quotable
The pages that matter get restructured answer-first, so an assistant can lift a clean, complete answer and attribute it to you. Then the citation assets: the dataset, the table or the tool in your category that nobody has published, which is what earns a link from something the models already trust.
Every month, if ongoing
Measure, and say what changed
Citation checks across the assistants, dated and reproducible. Crawler activity from your own server logs. Referral sessions from the assistants in your analytics. Three independent signals, and none of them is a score we invented.
How we prove it, and where we will not
This field is full of invented scores. Everything here is either something a machine actually served, something an assistant actually said, or a line in your own log file.
No proprietary visibility score. Three independent signals you can check yourself.
Signal 02
The assistants, asked directly
Your real questions put to ChatGPT, Perplexity, Claude, Gemini and Google, with the date and the exact prompt recorded so you can repeat it.
Signal 03
Referral sessions
Visits arriving from chatgpt.com, perplexity.ai and the rest, in your own analytics.
Signal 04
What we will not claim
We will never promise a ranking or a citation, because nobody controls those. We promise the work, and we show you the three signals so you can judge it.
What you get
№ 01
The crawler report
What GPTBot and PerplexityBot are really served by your site today, fetched as them rather than guessed.
№ 02
The citation ledger
Question by question: which assistant named you, which named a competitor, and which source it cited. Dated and reproducible.
№ 03
The foundation, installed
Access, structured data, llms.txt, sitemap, indexation and answer-first structure on the pages that matter.
№ 04
The evidence, monthly
If ongoing: crawler logs, referral sessions and fresh citation checks, with a plain note on what changed and what we did about it.
What this is not
What this is not:
- Not a guarantee of a ranking or a citation. Nobody controls those and anyone promising one is telling you something useful about themselves.
- Not link buying, and not content at volume. Generated filler is the fastest way to become invisible to the assistants we are trying to get you into.
- Not a proprietary score. Three checkable signals instead.
- Not a fight with a platform that cannot take the work. If yours cannot, we will say so and quote the build instead.
Fair questions
Do you handle Google Business Profile and local search too?
Yes, and it is included rather than sold separately. For a firm that serves a city, the profile, consistent name, address and phone across the directories, and LocalBusiness markup are what make you eligible to be the answer to a "near me" question at all, in Google and increasingly in an assistant that assembles local answers from those same sources. It is checklist work rather than engineering, which is exactly why it gets skipped, and why we do not skip it.
How are AI SEO services different from ordinary SEO?
It is SEO plus two things. The underlying discipline (crawlable, fast, structured, trustworthy) is the same, and anyone telling you otherwise is selling novelty. What is added is being readable and quotable by assistants that do not run JavaScript and do not click, and measuring citation rather than only ranking.
How long before anything happens?
Crawler access problems fix in days and you see it in your logs immediately. Structure and indexation move in weeks. Citations move over months, and anyone who tells you otherwise is guessing. We measure from week one so you watch it rather than waiting for a quarterly report.
Can we check any of this ourselves?
All of it, which is the point. Run our free scan on your own site right now with no signup. Everything we report afterwards is either a line in your own server log, a dated prompt you can repeat, or a number in your own analytics.
Do we need this if you are building our site?
No, it is already in there. Every site we build ships with the full foundation, so this SKU exists for firms whose site is fine and invisible, not for the ones we are building.
Start here
Find out what the machines see.
Run the free scan first, it takes ten seconds and needs no access. If the result warrants a conversation, we will have one.
John Weeks
The person who replies is the person who does the work.
Next on the path: The care plans (Visibility decays as the engines change. This is the half that needs to keep happening.)