IT-Rating: ×2 citations in AI answers and 6 source AI platforms
A B2B directory of Ukrainian IT companies became the source ChatGPT, Gemini, Copilot, Perplexity and Google AI Overviews cite when a user asks "which company to choose". An AI visitor stays 3× longer and converts twice as often as a Google visitor.
About the project
| Parameter | Value |
|---|---|
| Client | it-rating.ua |
| Niche | B2B IT: contractor rating & directory |
| Market | Ukraine, UA / RU |
| Measurement period | January — September 2026 |
IT-Rating is an independent platform where companies look for a contractor: a web studio, an SEO or SMM agency, mobile development, design, hosting. It is a classic B2B choice with a long deal cycle: the buyer compares several options, reads reviews and only then contacts a contractor.
That kind of choice increasingly starts not with Google but with a question to an AI: "top SEO agencies in Kyiv", "who to hire for e-commerce development". The goal — for IT-Rating to appear as a source inside the AI answer.
Starting point
By January 2026 the site already had strong organic, but was only spottily present in AI search.
| Metric, January 2026 | Value |
|---|---|
| Citations across all AI platforms | 308 |
| Pages cited by AI | 85 |
| Platforms citing the site | 2 of 6 |
| Google AI Overviews + AI Mode | 0 |
| Gemini, Copilot | 0 |
Almost all visibility rested on ChatGPT and Perplexity. In Google AI Overviews, which sit above the organic results, the site did not appear at all.
What we did
Opened the site to AI crawlers
Checked robots.txt: GPTBot, PerplexityBot, ClaudeBot and Google-Extended have access to the ratings. Added an llms.txt with a section map so the model immediately understands which contractor categories the site has.
Repackaged ratings for the AI-answer format
Each category answers the choice question in the first paragraph: how many companies, by what criteria, when the rating was updated. Then tables and short cards the model can quote verbatim.
Covered prompts, not just keywords
Collected the questions buyers ask AIs: "how to choose", "how much it costs", "agency vs freelancer". Added FAQ blocks and dedicated materials for them. That is what unlocked Google AI Overviews and AI Mode.
Strengthened the brand as an entity
Consistent data about the platform and mentions on industry sites. AIs cite a source more readily when they recognize it across several independent sites.
Set up measurement
Citations per platform in Ahrefs AI Responses; AI referrals and conversions as a separate segment in GA4. Without this, AIEO becomes work "by feel".
Citation dynamics
Over nine months citations doubled, and the number of pages AIs link to grew from 85 to 136 at the peak. By late September the figure pulled back, but even after the pullback the base is 1.7× above January: AI results fluctuate more than organic.
| Snapshot | Citations | Cited pages |
|---|---|---|
| January | 308 | 85 |
| March | 306 | 79 |
| May | 463 | 106 |
| July | 642 | 136 |
| September | 627 | 123 |
Breakdown by AI platform
The main shift is not in volume but in reach. In January two platforms cited the site; in September all six. Google AI Overviews and AI Mode added the most: AI blocks right in Google's results, seen by most searchers.
| Platform | January | September |
|---|---|---|
| Google AI Overviews | 0 | 40 |
| Google AI Mode | 0 | 38 |
| Gemini | 0 | 27 |
| Copilot | 0 | 15 |
| Perplexity | 16 | 10 |
| ChatGPT | 11 | 4 |
Quality of AI traffic
From March to September 2026, ChatGPT, Gemini, Copilot, Claude and Perplexity sent 714 sessions. Small next to organic, but the hottest traffic on the site: the visitor arrives with a formed question and a short-list.
| Source | Sessions | Engagement | Avg. time | Conversions |
|---|---|---|---|---|
| chatgpt.com | 289 | 56% | 7:15 | 147 |
| gemini.google.com | 209 | 59% | 6:56 | 86 |
| copilot.com | 104 | 71% | 12:09 | 214 |
| claude.ai | 64 | 70% | 8:45 | 49 |
| perplexity.ai | 48 | 38% | 1:46 | 0 |
| AI total | 714 | 59% | 7:38 | 496 |
Results in numbers
| Metric | January 2026 | September 2026 | Change |
|---|---|---|---|
| Citations across all AI | 308 | 627 | ×2.0 |
| Cited pages | 85 | 123 | +45% |
| AI platforms with citations | 2 | 6 | +4 |
| Google AI Overviews | 0 | 40 | new channel |
| Google AI Mode | 0 | 38 | new channel |
| Organic traffic (Ahrefs) | 16,666 | 27,838 | +67% |
Takeaways
For a B2B choice the AI becomes the first filter: it forms the short-list before the person opens a site. IT-Rating made that short-list on six AI platforms, not one.
The same approach transfers to any niche with a complex choice: system integration, retail automation, logistics, industrial equipment. There too the buyer asks the AI "who to choose and why".
Want the same result?
Tell me about your product — I'll check whether ChatGPT, Gemini and Google AI Overviews name you, and propose a strategy.
Discuss a project →