Live case study · measured in real time

We sell AI visibility.
And we started with ourselves.

insync.com.ua is the first site to go through our GEO framework. A technical audit of 100/100, and in a clean session ChatGPT already quotes our positioning verbatim — citing “InSync Agency” as the source. Below are real numbers you can replicate on your own site.

0AI Visibility Score
insync.com.ua · 8-factor audit · 8 of 8 passed
The main proof

ChatGPT is already retelling us — verbatim

This isn't rendering or editing. In a “clean” ChatGPT session (Temporary Chat, no history or cookies), when you ask about InSync, the AI quotes the positioning straight from our site and labels the source as “InSync Agency.” That's how the GEO loop works: the site becomes what AI reads and repeats.

You → What kind of agency is InSync? What do they do, how much does it cost?
AIChatGPT · Temporary Chat · with web search
“InSync is a performance marketing agency that combines paid advertising and AI. It works with businesses in UA, EU, US, Canada, Switzerland, Kazakhstan, and the UAE… Service cost is $450–700/mo — for the Ukrainian market, that's not overpriced.”
Source cited by ChatGPT: InSync Agency
You → Who is InSync's CEO and what do they specialize in?
AIChatGPT · Temporary Chat · with web search
“The CEO is Anton, the agency focuses on targeted advertising and AI solutions for international markets…” A remark from the model itself: “still few independent sources” — that's exactly the gap our subscription closes (external citations, reviews on Clutch/GoodFirms).
Source cited by ChatGPT: InSync Agency
The gap with the market

The same scanner — on real websites

We ran sites from our working niches through the scanner. The average business site scores 15–30 out of 100 — AI models barely see it and recommend whoever they can see instead.

insync.com.uaour site after the GEO framework
0
A “furniture” niche sitean ad landing page, no GEO
0
A “HoReCa” niche sitea restaurant site with bookings
0
What's inside the 100 points

8 AI visibility factors — all covered

Each factor is a specific reason ChatGPT, Perplexity, or Gemini either cites a business or stays silent about it.

llms.txt

A “passport” for AI models: services, pricing, case studies, geography — in a format the LLM reads first.

20 POINTS

AI crawlers in robots.txt

GPTBot, ClaudeBot, PerplexityBot, and 7 more bots are explicitly allowed to index and cite.

15 POINTS

Schema.org (JSON-LD)

6 types of markup: Organization, Person, Service, FAQPage, Breadcrumb, ContactPoint — machine-readable facts.

15 POINTS

Sitemap.xml

28 pages in the sitemap: services, niches, comparisons, blog — everything crawlers need to find.

15 POINTS

FAQ markup

Question-and-answer is the LLM's favorite format for citation. On every key page.

10 POINTS

Meta description

Unique descriptions on every page — engines pull answer snippets from them.

10 POINTS

The site responds fast

Accessibility and speed: a crawler that doesn't get a response in time won't cite anything.

10 POINTS

Title tags

Headlines built around real niche questions, not “Home — Company N.”

5 POINTS
A live measurement, not a promise

We show the scoreboard — including our own

A 100/100 technical framework is readiness. What decides things next are mentions in AI answers. Here's our honest baseline at the start and an open plan: we document our own before/after every month. No one else on the market shows you the score this transparently.

0%
Share of Voice in ChatGPT at the start
(20 benchmark niche queries)
0/4branded queries — won
0/16category queries — the 90-day target

A public 90-day plan

How we turn 20% → growth in category share. The same plan — for clients.
  1. Content for category queries. Answer pages for “who to choose / how much does it cost / how to pick a contractor” — in a format engines cite.
  2. External citations. Reviews on Clutch/GoodFirms and industry mentions — the same gap ChatGPT pointed out itself: “few independent sources.”
  3. A 20×4 measurement every month. ChatGPT, Perplexity, Gemini, Copilot: did they name us or a competitor (Netpeak, Inweb, Elit-Web…). The dynamics — in the report.
  4. Data-driven strategy rebuild. We see who gets cited instead of us, and go there. We fix things faster than the market can notice.
Methodology · AI Attribution

Three layers of proof in every report

We don't sell "faith" in a single score. Every month — three independent layers, each with a screenshot and an export. No number taken on faith.

LAYER L1

Referral traffic from AI

Visits from chatgpt.com, perplexity.ai, gemini.google.com, copilot — the "AI Sources" segment in your GA4. Proves: real people came from AI answers.

LAYER L2

Visits from AI crawlers

GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot from server logs. Proves: AI models are physically reading the site and updating what they know about it.

LAYER L3

Share of Voice 20×4

20 control queries for the niche × 4 engines: whether the client is mentioned, who's mentioned instead, month-over-month dynamics against competitors.

To be honest: 100/100 means technical readiness for AI search, not "#1 in ChatGPT." Mentions build up over 2–3 months through content and external sources. Anyone promising guaranteed rankings in ChatGPT is lying. What we guarantee is a system of work and a transparent measurement — including our own, published openly.

The setup in numbers

What's built under the hood

0pages in the GEO setup: services, niches, comparisons, blog
0AI crawlers explicitly allowed in robots.txt
0types of schema.org markup on the site
0language versions: UA / RU / EN with hreflang
Your move

How many points does your site score?

The same scanner we used to check ourselves: 8 AI visibility factors, a /100 score, and a list of exactly what to fix. Free, no sign-up, right here — 60 seconds.

We check publicly available data on your site: the homepage, robots.txt, llms.txt, sitemap.xml, and markup. We don't change or store anything.
Opening the homepage
Reading title, description, and headings
Looking for schema.org and FAQ markup
Checking llms.txt — the passport for AI models
Checking robots.txt: are AI crawlers let in
Calculating your AI Visibility Score
0out of 100
—

What to do with this

The score is a diagnosis, not a prescription. Leave your contact and we'll send you a breakdown of your specific case — what to fix first, what it'll get you in your niche, and how much it costs to close the gap.

Or message our strategist directly: @anton_tysiachnik

Want this handled every month without lifting a finger? GEO Autopilot — a subscription from $199: the technical setup, 4 new pages a month, and tracking across AI engines.