Czech AI E-commerce Visibility Index

The first comprehensive study of how Czech e-commerce brands appear in AI-generated answers versus traditional search results. We analyzed 10,303 domains across 19 product categories using 1,150 test prompts sent to GPT-5, Gemini, and Perplexity, cross-referenced against Google.cz and Bing. Data collected March 2026.

Key Findings — March 2026

  • 39.6%AI↔SERP overlap — 60%+ of AI-recommended domains don't appear in SERP for the same query
  • 1,840ghost mentions — AI models recommend defunct e-shops; 75% driven by GPT-5's stale training data
  • 25.7%Jaccard index — low set-similarity between AI and SERP domain sets across all 19 categories
  • 10,303unique Czech e-commerce domains tracked across electronics, fashion, home, sports, beauty, and 14 other categories
  • 8defunct e-shops still recommended by AI — sending shoppers to dead pages and eroding trust in AI answers
  • Power-lawconcentration — a small number of domains captures a disproportionate share of all AI recommendations
Free AI Visibility Audit

Is your brand visible in AI search?

We analyzed thousands of Czech e-commerce domains across multiple AI engines. Find out where your brand stands — and what you're missing.

Get Your Visibility Score
Analysis conducted: March 2026 1,150 prompts · 19 categories · 3 AI models · 2 search engines
10,303
Domains Analyzed
39.6%
AI↔SERP Overlap
1,840
Ghost Mentions
25.7%
Jaccard Index
1,150
Prompts Tested
19
Categories
Recommendation Concentration Analysis ⓘShows what percentage of all AI brand mentions are captured by the top N domains. A steep curve indicates high concentration — a few brands dominate AI recommendations.
Top 10 — AI Mentions ⓘAI Visibility is based on position-weighted mentions across GPT-5, Gemini, and Perplexity.
Top 10 — SERP Mentions ⓘSERP Visibility is based on Google.cz and Bing rankings across 1,150 test queries.
Model Disagreement ⓘDomains where AI models disagree most. High spread = one model rates it highly, others ignore it.
Category AI↔SERP Alignment ⓘEach chart shows how AI-recommended domains compare with traditional search results for that category. Dots below the diagonal = AI ranks them higher than SERP.
Each dot is a Czech e-shop. The diagonal line marks AI↔SERP parity — dots below mean AI recommends them more, dots above mean traditional search favors them.
Status
Sort
Category
What do the columns mean?
#
Rank
Overall position by score
Score
Score
AI + SERP mentions combined
AI
AI Mentions
Times recommended by ChatGPT, Gemini, Perplexity
SERP
Search Mentions
Times appearing in Google / Bing results
POS
Avg. Position
Average rank when mentioned — lower is better
SRC
Sources
Which platforms mention this domain (colored dots)
▦
Rank Spread
Position distribution across positions 1–10
CAT
Categories
Number of product categories where it appears
# Domain Score AI Mentions Search Mentions Avg. Position Sources Rank Spread Categories
Position Breakdown ⓘHow appearances spread across positions #1–10. The prime zone (top 3) captures most visibility. Click any domain row to inspect its breakdown.
AI Rank vs SERP Rank — All Domains ⓘEach chart shows how AI-recommended domains compare with traditional search results for that category. Dots below the diagonal = AI ranks them higher than SERP.
Each dot is a Czech e-shop. The diagonal line marks AI↔SERP parity — dots below mean AI recommends them more, dots above mean traditional search favors them.
Category AI↔SERP Alignment
AI vs SERP — Full Domain Landscape ⓘEvery tracked domain plotted by AI mentions vs SERP mentions. Four quadrants reveal different competitive positions.
Each dot is a Czech e-commerce domain. The four quadrants show where opportunities and blind spots live.
Why Do These Gaps Exist?
Filter by category:
Source Correlation Matrix ⓘJaccard similarity between all 5 sources (3 AI models + 2 search engines). Low values mean the sources recommend almost entirely different domains. Each cell = % of shared domains out of the combined set.
Each cell shows the Jaccard similarity — the percentage of shared domains between two sources. Lower values indicate more divergent recommendations.
Top 15 per Source ⓘSide-by-side top 15 for each AI model and search engine. Dead brands marked ☠ in red.
AI Mention Distribution ⓘHow many domains fall into each mention-count bucket. The vast majority of domains have very few AI mentions — a classic long-tail distribution.
Most domains live in the long tail — mentioned only once or twice by AI. This is where unexpected opportunities hide.
The Domain Universe ⓘHow the 10,303 tracked domains split between AI-only, SERP-only, and both. Large AI-only and SERP-only pools show how different the two worlds are.
Single-Model Domains ⓘDomains that appear in only one AI model. Each model has its own "private" knowledge — brands invisible to other models.
Long-Tail Surprises ⓘDomains that AI recommends but Google/Bing never show. These are invisible in traditional SEO yet surface in AI answers — the biggest opportunity for niche e-shops.
AI-only domains that don't appear in any search engine results. Unknown to SEO, visible to AI.