Observed Signal · Aug 1, 2026 · Market Research · Source: DEV Community · Impact: 3/5 · Sentiment: Neutral
AI Recommendations in Romania's Book Market
A reproducible study tested 18 purchase questions in Romanian across five AI engines (ChatGPT, Google Gemini, Perplexity, Google AI Mode and Google AI Overviews) for two weekly runs (18–31 July 2026). The researchers analysed 180 answers and found 247 brand mentions, 21 tracked brands and 422 distinct cited domains. 58% of answers named no retailer; when brands were named they were often multi-branded rather than single recommendations. Libris led by raw mentions and first-position visibility. The study distinguishes 'ghost brands' (sites cited as sources but rarely named) and 'memory brands' (brands named by the model without being cited). Engines behave like distinct channels with large per-engine differences and weekly volatility. The dataset and CSV files are published CC BY 4.0 and available on Zenodo, GitHub, Kaggle and Hugging Face.
Demonstrates how conversational AI engines act as distinct discovery channels that reshape retailer visibility and generative search optimization (GEO); provides reproducible dataset useful to marketers and publishers but is not a major platform policy or technical infra change.
Track Perplexity Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Study queried 18 purchase questions across five AI engines, producing 180 analysed answers.
- There were 247 brand mentions, 21 tracked brands, and 422 distinct cited domains in total.
- 58% of answers (105 of 180) named no retailer or brand.
- Libris had 37 mentions (≈20.6% of all answers) and often appeared in first position.
- Study window: engine behaviour described between 18 and 31 July 2026; data published CC BY 4.0 with DOI 10.5281/zenodo.21736211.
Connected Companies & Entities
4 Entities mapped“We put 18 real purchase questions to five AI engines — ChatGPT, Google Gemini, Perplexity, Google AI Mode and Google AI Overviews — in Roman...”
“We put 18 real purchase questions to five AI engines — ChatGPT, Google Gemini, Perplexity, Google AI Mode and Google AI Overviews — in Roman...”
“Mention rate per brand, as a share of each engine's 36 answers: (table row includes Litera: 13.9% ChatGPT, 25.0% Gemini, 5.6% Perplexity, 27...”
“When someone asks ChatGPT "which online bookstore should I use for children's books?", they get one answer, not ten blue links....”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
AI Engines Disagree by 2× on Which Brands to Recommend
Treyci, an AI visibility intelligence platform, released measurement data showing that major AI engines (ChatGPT, Perplexity, Gemini, Grok) sharply disagree on which brands to recommend in buying-intent queries. In one B2B software category, one engine referenced tracked brands in 81% of answers while another did so in only 43%, a nearly two-fold visibility gap. The study scored over 1,200 AI answers from 100 buying-intent questions per category, repeated three times per engine monthly. Key findings: AI answers are probabilistic (same prompt yields different vendor lists across sessions), engines cite third-party reviews and publications rather than vendor websites, and adoption is ahead of measurement—only 41% of 100 B2B SaaS companies published llms.txt files. Founder Keith Schilling emphasizes that brands must measure the distribution of answers across engines, not rely on single screenshots.
B2B Brands Risk Obscurity on AI Platforms, Study Finds
SEMAI analyzed 25,540 URLs cited by ChatGPT, Google Gemini, and Perplexity over a 60-day period and found each AI platform cites materially different content types. Blogs (41–55%) and webpages (38–47%) together account for more than 90% of AI citations, but platforms diverge on niche sources: ChatGPT cites LinkedIn, Wikipedia, and academic/research-backed content more than Gemini or Perplexity; Perplexity emphasizes comparison and solution pages; Gemini favors brand-owned blog content and avoids community sources. SEMAI argues that optimizing for a single channel (e.g., Google) can leave B2B brands invisible to buyers who begin research on AI platforms. SEMAI’s visibility platform tracks ChatGPT, Perplexity, Gemini, and Claude and offers LLM search volume metrics, cluster classifications, and AI-generated content calendars; a free AI visibility audit is available on its site.
AI Answer Engines Challenge Brands
An Adweek article (published May 1, 2026) summarizes a fireside conversation at ADWEEK House Possible co-hosted with Brandwatch in which ADWEEK Editor in Chief Ryan Joe and Brandwatch VP Eric deLima Rubb discussed how AI answer engines (e.g., ChatGPT) are reshaping search and the customer journey. Rubb cited an approximate 26% year‑on‑year growth in search activity and warned that AI models are being trained on social evaluation behavior from platforms like Reddit and forums. He advised marketers to treat AI as a stakeholder — influencing what AI cites by contributing authentic content on social platforms, leveraging visually rich sites like Pinterest (which he called under-indexed today), and monitoring post‑content conversations such as YouTube comment threads. The piece notes current measurement limitations for AI citation behavior and recommends an omnichannel approach while brands experiment to influence AI-driven discovery. The article was created in partnership with Brandwatch.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.