AI-engine visibility of Romania's electronics and IT retail market: 18 purchase questions across 5 AI engines, August 2026

Raw data from an original study measuring how 13 retailers in Romania's electronics and IT market appear inside the answers of five AI engines (ChatGPT, Google Gemini, Perplexity, Google AI Mode, Google AI Overviews). Eighteen brand-free purchase questions were asked once from Romania in Romanian, producing 87 analysed answers and 378 brand mentions, with the position of each mention, the domains cited as sources and sentiment recorded throughout. Unlike the book market measured in the previous episode of this series, this market is highly concentrated: 86 of the 87 answers name at least one tracked retailer, and the competitive difference lies almost entirely in the position a brand occupies inside the answer rather than in whether it appears at all.

Method. Eighteen real purchase questions were put to ChatGPT, Google Gemini, Perplexity, Google AI Mode and Google AI Overviews, in Romanian, from Romania, in one complete run on 3 August 2026. No question contained a brand name — every retailer that appears was named by the engine on its own. Of the 90 queries issued, 87 produced an analysed answer; Google did not display an AI Overview for three questions. The result is 378 mentions across 13 tracked retailers, with the position of each mention, the domains cited as sources, and sentiment recorded throughout. Monitoring was automated via the LLM Pulse platform, with no session history and no personalisation.

Key findings.

• This is a concentrated market: 86 of the 87 answers name at least one of the 13 tracked retailers. Only one answer names none — the opposite of the book market, where 58% of answers named no major brand.
• Position, not presence, separates the players. eMAG takes first place in 66 of its 82 mentions (80.5%). evoMAG, fourth by visibility, has 37 of its 47 mentions in positions 4-7 and only 3 in first place.
• The market ranking is inherited, not computed. A single listicle — shopilo.ro — is cited in 21 answers, more than the websites of PC Garage or Media Galaxy, and one Wikipedia list page is cited in 7.
• Specialists own their territory and it pays. F64 appears in all five answers about professional cameras and almost nowhere else: 9 mentions produce 40% of the weighted visibility of a retailer with 47 mentions.
• Google withheld AI Overviews on 3 of the 18 questions — gaming laptop, large TV and cheapest phones, three of the most commercial queries in the set.
• Being named and being cited are different things. 21 of the 47 answers naming evoMAG cite no evomag.ro page at all, while 4 answers cite evomag.ro without recommending the brand.
• Reddit is the second most-cited non-retailer source in the market (17 answers), after the listicle shopilo.ro (21).
• Sentiment is almost uniformly positive: one negative mention across the whole tracked set.

Contents. Three CSV files (one row per retailer; one row per question-brand pair; the 60 most-cited domains out of 249), four charts as SVG, and a data dictionary.

Denominators. Overall mention rates are computed against 87 — the analysed answers. Per-engine rates are computed against that engine's own answer count: 18 for ChatGPT, Google AI Mode, Perplexity and Gemini, but 15 for Google AI Overviews, which did not trigger on three questions. The monitoring platform reports AI Overview rates against 18; this dataset corrects them to 15. Per-question rates use 5 answers, or 4 for the three questions without an AI Overview. Mixing these denominators produces wrong numbers.

Limits. A single run on one day — a precise snapshot, not a trend. Weekly volatility, measured in the previous episodes of this series, cannot be observed here. 18 questions rather than the full purchase universe. 13 tracked retailers rather than the whole market; the answers also name specialists such as Flip.ro, Decathlon, TrotiBike and ReBoxed that are not tracked. About 5 answers per question, so single-question differences are noise. Mentions are not conversions. Models update continuously. A single unpersonalised user profile, one country and one language.

Responsibility note. All brands named in this dataset were named by the AI engines, not by the authors. The material is market research and does not evaluate the quality, services or commercial performance of any brand. Trademarks belong to their rightful holders and are used for identification within a factual analysis. Corrections: office@websem.ro

Canonical version of the analysis: https://websem.ro/resurse/aeo/studiu-electronice-it-romania

This is episode 3 of an ongoing series measuring how Romanian markets appear in AI answers. Episode 1 covered luxury jewelry (DOI 10.5281/zenodo.21724399). Episode 2 covered the book market (DOI 10.5281/zenodo.21736211).