When someone asks an AI engine for the best headphones, TV or vacuum cleaner, the brand doesn't write the shortlist. Reviewers do.
These findings come from the consumer studies in the GEO Compass benchmark: about 158,000 AI answers across the US, India, the UK, Germany, France, Spain, the Netherlands, Sweden, Australia, Singapore and other markets.
Earned media dominates
Earned media supplies 43% of citations in consumer answers, more than twice the benchmark average and the highest of any sector. Brands' own websites are cited in only 6% of unbranded consumer answers, half the benchmark rate.
Share of answers with sources. GEO Compass benchmark, consumer, all markets.
Product reviewers outrank every brand website in category questions. Google's AI surfaces add video: AI Mode cites YouTube in 44% of consumer answers, against 7% for ChatGPT.
The good news: consumer answers are positive
When AI engines discuss consumer brands, they are positive 60% of the time in unbranded answers, the highest of any sector, against 39% across the benchmark. Being on the shortlist is valuable. The work is getting onto it.
Every market trusts its own reviewers
| Market | Most-cited consumer sources (excluding brand sites) |
|---|---|
| Australia | YouTube, TechRadar, Reddit, Tom's Guide, CNET, Wired |
| India | YouTube, Indiatimes, Reddit, Amazon, Instagram, Quora |
| Singapore | Wikipedia, Facebook, CNA, Instagram, Reddit, The Straits Times |
| United Kingdom | TechRadar, YouTube, Reddit, Tom's Guide, The Independent |
| Germany | CHIP, HIFI.DE, COMPUTER BILD, YouTube, TechRadar, FAZ |
In Germany, local-language answers lean even harder on earned media, and the leading sources are German titles. In India, marketplaces such as Amazon and community platforms such as Quora join the review titles. In Singapore, brand sites carry more weight than elsewhere, with national news and social platforms behind them.
What consumer brands should do
- Map the reviewers AI engines cite in each market and prioritise them for product seeding and PR. They may not be your traditional media list.
- Make sure reviewers have what they need: specifications, test units, comparison data and clear differentiators, so the review describes you accurately.
- Invest in YouTube for Google's AI surfaces, through your own channel and creator reviews with chapters, transcripts and question-led phrasing.
- Keep marketplace listings complete in markets such as India where Amazon and aggregators are cited.
- Publish honest comparison and "best for" pages on your own site, with structured Product data, so engines have a quotable source from you.
- Localise. An English-only media plan misses the sources that shape answers in German, French, Spanish and other local-language questions.
See consumer figures on the industry page, or compare engines in the explorer.
Frequently asked questions
- Which sources do AI engines use for product recommendations?
- Mostly review publishers and video. In the GEO Compass benchmark, earned media supplies 43% of citations in consumer categories, the highest of any sector. TechRadar, RTINGS, What Hi-Fi?, Tom's Guide, YouTube and Reddit are among the most-cited sites.
- Does AI search favour local review sites?
- Yes. Answers to questions asked in local languages lean on local reviewers. In Germany, CHIP, HIFI.DE and COMPUTER BILD are cited more than international titles.
Figures marked as GEO Compass data come from the anonymised GEO Compass benchmark of AI answers collected by We AI & Data Advisors. Related: AI search in Australia · AI search in Singapore · AI search in India · AI search in Europe · Consumer