Short answer. Recommendation answers are assembled from pages that already rank the options, not from your own website. Third-party lists took 63% of Google AI Overview citations in one study of "best software" queries, while the recommended product's own site took only 12%. Listicles alone earn 40% of commercial-intent AI citations across ChatGPT, Google AI Mode and Perplexity. Improving your presence means moving three parties: listicle publishers, review platforms and managed AEO/GEO providers.
Key takeaways
- Third-party lists took 63% of Google AI Overview citations on "best software" queries, while the recommended product's own site took just 12%.
- Listicles hold 21.9% of all AI citations and 40% of commercial-intent citations across ChatGPT, Google AI Mode and Perplexity, based on a 75,000-answer, 1-million-citation study by Wix Studio's AI Search Lab.
- Gemini-grounded "best in city" answers cited a directory or ranking site 78% of the time, and no business's own site reached the top cited domains.
- Self-ranked listicles were cited but a rival was recommended instead 69% of the time in a study of AI Overviews, and self-promotional listicles correlated with organic traffic declines.
- Three parties can move a recommendation answer: listicle publishers and editorial media, review platforms and directories, and managed AEO/GEO providers who reconcile facts across both.
How are AI recommendation answers actually built?
An AI engine answering a "which X should I buy" question does not reason from first principles about your product. It retrieves pages that already rank the options and compresses them into a shortlist, so whoever wrote the comparison effectively wrote the answer.
In June 2026, one analysis logged 1,259 citations behind Google AI Overviews for 100 "best [category] software" searches: third-party "best of" lists earned 63% of them, and the recommended product's own website earned 12%. Answer engine optimization (AEO) is the practice of shaping content so answer engines retrieve and quote it directly for a query; on recommendation queries, that content is rarely the brand's own page. Wix Studio's AI Search Lab analysed 75,000 AI answers and more than a million citations across ChatGPT, Google AI Mode and Perplexity and found listicles took 21.9% of all citations, the largest share of any page type, and 40% of commercial-intent citations. Articles took 16.7% overall and product pages 13.7%.
The pattern holds outside software. Acromatico ran 100 "best [vertical] in [city]" searches through Gemini with live Google Search grounding: the engine named an average of 12.5 businesses per query and cited directory and ranking sites in 78% of answers, with the most-cited domains including bestlawfirms.com, reddit.com, forbes.com, superlawyers.com and justia.com — no business's own website appeared on that list. Profound's citation data shows the same pattern on Perplexity, where G2, Gartner, NerdWallet, PCMag, TripAdvisor and Yelp lead on commercial intent, and Semrush's 2026 AI Visibility Index confirms it from the brand side: Patagonia held an AI visibility score around 79 to 80 throughout the study, supported by consistent descriptions on OutdoorGearLab, REI, Switchback Travel, GearJunkie and Reddit rather than by its own site.
Why self-ranked listicles backfire
A brand's own "why we are best" list looks like the same format, but it earns a different outcome. Lily Ray checked 100 B2B "best software" queries in Google AI Overviews three times between April and June 2026. Self-ranked listicles, where a brand ranks itself first, were cited 323 times, and in 224 of those cases — 69% — Google cited the brand's page and then recommended a rival from inside it. She also reported organic declines from around 20 January across dozens of sites that leaned on self-promotional listicles. The page earning the citation and the page losing traffic were the same page. The finding is narrower than "never publish comparisons": self-ranked listicles that put the brand first get cited and not recommended, while comparisons that state criteria, name competitors and concede where a rival wins are the format engines cite most, because the engine reads them as evidence rather than advertising.
Who controls the pages AI recommends from?
Three separate parties each control one layer of a recommendation answer, and no single one of them controls all of it.
Generative engine optimization (GEO) is the broader discipline of improving how a brand is represented across the third-party pages, structured data and language coverage that generative AI systems draw on — it is the umbrella the three parties below operate under.
Listicle publishers and editorial media
These pages take 40% of commercial citations. An independent "10 best CRMs for small teams" on a publisher the engine already trusts is the single most retrieved document type for that question. Muck Rack's 2026 analysis of 25 million cited links across ChatGPT, Claude and Gemini found 84% of AI citations trace to earned media rather than owned content, and Stacker and Scrunch tracked 87 earned media stories across 30 clients and 2,600-plus prompts, documenting substantial median increases in brand citation rates within 30 days of distribution. What a publisher cannot move is the facts about you: it describes you from whatever public information exists, and an out-of-date pricing page or an ambiguous category gets repeated into the listicle and then into the engine's answer. This work sits with digital PR practices such as Go Fish Digital, Siege Media's earned media operation and Stacker, plus PR agencies running AI citation reporting.
Review platforms and directories
These control the structured, current, third-party opinion engines find easiest to read. A 2026 study of software category queries found every tool ChatGPT named had Capterra reviews and 99% had G2 reviews, and brands in the top 20 of their category on those platforms were cited roughly three times more often in "best software" answers. On local and professional services, the equivalent is Google Business Profile, Yelp, TripAdvisor and vertical directories; Google's own generative AI guidance points local businesses and merchants to Business Profiles and Merchant Center feeds for visibility in AI responses. A strong review profile keeps a brand eligible, but it does not make the engine prefer that brand over a competitor with an equally strong one. Review-generation and reputation programmes, usually run in-house or by a customer marketing team, plus local SEO specialists, work this layer.
Managed AEO/GEO providers
This is the layer underneath both of the above: whether the facts third parties are working from agree with each other and with the brand. Semrush found that on Gemini the overlap between brands mentioned and domains cited can be as low as 30% — the engine names a brand from third-party evidence without reading its site at all. A managed provider audits what those third parties say, reconciles entity facts across them, publishes honest comparison content in the formats engines cite, and, for multinational brands, does it in each market language rather than only in English. What a provider cannot move is a product reviewers do not like; it can make third-party judgement accurate and legible, not favourable. Lifewood runs managed AEO and GEO programmes with a six-stage workflow from Intake and Semantic Audit through Pillar Execution, QA, Deployment and Performance Reporting, with native-speaker review across 50+ languages from 40+ delivery centres across 30+ countries. Agencies including Omniscient Digital, First Page Sage and iPullRank do comparable work in English for their respective buyer types. For a side-by-side of providers working this layer, see Best AEO and GEO Agencies for AI Search Visibility and, for markets outside English, Top 10 Answer Engine Optimization Companies in Asia.
Why does multilingual coverage matter for recommendation answers?
Recommendation answers are assembled from third-party pages written in the language of the question, so a brand present only in English on a review platform is effectively absent from non-English recommendation queries.
In a market where the trusted review platforms, directories and forums are local and the engine grounds its answer in local search results, the same reconciliation work described above has to happen in that language, not just in English. Any provider working on non-English recommendation queries needs native-language reviewers who can check what local directories, forums and publishers say about a brand and correct it, or it is guessing. This is the same fact-consistency problem Why Third-Party Brand Mentions Matter for GEO and AI Search covers in more depth, and it is one reason brands run this work through a dedicated /geo programme rather than a single English-language listicle campaign.
Should you handle this in-house or hire a provider?
Doing it in-house works well for a narrower set of conditions; the rest is usually cheaper to buy than to build.
Handle it in-house when the brand operates in one language and one or two engines, the category's review platforms are obvious — G2 and Capterra for software, Google Business Profile and Yelp for local — and a content owner can publish an honest comparison and keep it current. The tooling is inexpensive: monitoring starts around $29 a month plus a fixed prompt list run manually, as covered in 7 Reasons AI Isn't Citing Your Brand (and the Fix for Each).
Bring in help when the facts about the brand disagree across third-party sources and nobody owns fixing that, when the brand needs earned inclusion on publisher lists and has no PR function, when it operates across languages or markets where the trusted third parties are unfamiliar, or when a wrong recommendation is expensive, as in regulated products or high-value B2B. 10 Questions to Ask Before Hiring AEO and GEO Help sets out what to check before signing a contract, and What Actually Gets You Cited by AI Answer Engines covers the content side of the same decision.
Either way, the measurement is the same: run a fixed set of "best X for Y" prompts per engine and log the cited sources. If the sources are third-party lists the brand is not on, that is a publisher problem. If they are review platforms where the brand is under-ranked, that is a reputation problem. If the engine names the brand but cites nothing of its own, the brand's facts are living on other people's pages and need reconciling. See /aeo-geo-providers for how to compare providers against that checklist.