Updated on 7 November 2026, by João Leite, Founder of murmur.
Reviews and Reclame Aqui influence AI: they enter as a source for the answer, which repeats the company’s rating and reputation on those sites. How much they enter changes with the AI model, with the subject of the question and with the country. Up to 6 October 2026, we did not find a study that separates the weight of reviews in the choice of the company the AI recommends. Those that crossed reviews with citation found association, not cause.
- In the 51 sector studies of September 2026, Claude cited or named a review site in 12.0% of 1,273 answers. ChatGPT did so in 1.3% of 1,275, and Google AIO, Google’s AI summary in search, in 4.3% of 1,258.
- Reclame Aqui in ChatGPT: in March 2026, it was among the sources of 6.5% of the conversations Promptado measured, about 204 Brazilian brands. Promptado sells a monitoring tool for AI models.
- The sample changes the result: G2 is the third most cited by ChatGPT in Surfer’s September 2026 list. In the 233 software recommendations DerivateX counted, it had zero citations.
- No study we read isolated the weight: in G2’s own calculation, from October 2025, review volume explained less than 1% of the difference in citations.
- What to do at your company: ask every customer for a review, with nothing in return, and answer and resolve complaints.
Reviews and Reclame Aqui appear more when the question is about buying
Reviews and Reclame Aqui appear more in AI answers when the person asking is close to buying. Trustpilot, a review site that sells plans to the companies reviewed, commissioned a study from the agency Seer Interactive published in May 2026. It analyzed 804,491 answers from four AI models, without Claude. In questions from people getting to know the subject, review sites were 1.51% of citations. In questions from people about to buy, they were 24.27%.
murmur, a GEO agency in Brazil that puts brands in the answers of ChatGPT, Claude, Google AIO and other AI agents, writes the pages that answer the buyer. It monitors the brand week by week. It also publishes sector studies, with 25 category questions in each one, without brand names. In the 51 studies of September 2026, we counted the answers that cite or name one of eight review sites. They are Reclame Aqui, where the consumer complains about a company, Trustpilot, Clutch, for services, and five for software: G2, Capterra, GetApp, Software Advice and TrustRadius. Citing is putting the site’s link among the sources, and naming is writing its name.
In these studies, Claude cited or named one of the eight sites in 153 of 1,273 answers, or 12.0%. ChatGPT did so in 17 of 1,275 (1.3%). The collection misses some of ChatGPT’s sources in list answers, so its number is at least that. Google AIO, the summary written by AI at the top of Google search, did so in 54 of 1,258 (4.3%), among the summary’s sources or by name in the text. In total, that is 224 of 3,806 answers. Four sectors accounted for 37% of them: European citizenship, student exchange, digital accounting and consortium.
The most recent news from each AI model, the good and the bad for review sites:
| AI model | The good | The bad | In Brazil |
|---|---|---|---|
| ChatGPT | Surfer, September 2026: Clutch 2nd and G2 3rd among the most cited domains, in questions from Surfer users | DerivateX, May 2026: zero citations to G2 and Capterra in 233 software recommendations | Promptado, March 2026: Reclame Aqui among the sources of 6.5% of conversations about 204 brands |
| Claude | Orbit Media, September 2026: Clutch as the most cited third-party site in business-to-business buying questions | Otterly, June 2026: no category for review sites; 64% of citations in software and technology went to company and product sites | murmur, September 2026: review site in 12.0% of answers, Reclame Aqui in 9.4% |
| Google AIO | SE Ranking, December 2025, US: 34.5% of summaries of commercial searches with at least one review site | Ahrefs, September 2026, US: no software review site among the 50 most cited domains | murmur, September 2026: review site in 4.3% of answers, among the sources or by name |
| Perplexity | Peec AI, updated in September 2026: G2 among the five most cited sources; Ahrefs: Trustpilot 10 places higher, in 26th | Orbit Media, September 2026: LinkedIn, YouTube and Reddit as the most cited | murmur, September 2026: 41 of 372 answers with a review site, almost always Reclame Aqui |
In ChatGPT, Google AIO and Perplexity, recent measurements give opposite results. In Claude, the measurements we read count different things. The review site rises or disappears depending on the subject of the question, the country and who built the sample.
How much reviews weigh in the AI recommendation: association, not cause
The weight of reviews in the AI recommendation does not appear in any study we read, and those that came close found association. G2, a software review site that sells services to the vendors reviewed, cross-referenced reviews and citations in AI models across 500 software categories. In the calculation published by G2 in October 2025, the categories with 10% more reviews had 2% more citations. Review volume explained less than 1% of the difference between them, and G2 writes the limitation:
“These limitations suggest our estimates should be interpreted as suggestive associations rather than causal effects.”
In the Seer study, answers about brands with a Trustpilot profile cited their page there in more than half of the cases, at the median, and answers about brands without an active profile, in 1%. Without a profile, there is little to cite, and Seer acknowledges that other factors may have played a part:
“Of course it’s possible that other factors like marketing spend and audience size influenced these rates, but our data shows a measurable lift for brands who invest in a fully optimized Trustpilot profile.”
The closest to measuring the weight was a laboratory test, without search. In June 2026, Xi Chu and Yupeng Hou put invented skincare brands against a known brand, in three AI models. The known brand’s advantage disappeared when the competitor had less than 0.1 star more. The rating was in the text given to the model, not on a page found by search.
In ChatGPT, review sites appear a lot in one sample and almost nothing in another
In ChatGPT, the presence of review sites varies depending on the subject and on who assembles the sample. Surfer, which sells an SEO tool, counted ChatGPT citations from 1 to 28 September 2026, in questions from its users, many of them in marketing and software. Clutch, which gathers reviews of service vendors, came second among the most cited domains, and G2 third. Trustpilot came 47th.
In software buying, DerivateX, an SEO and GEO agency for software companies, found the opposite. In 233 ChatGPT recommendations with search on, G2 and Capterra had zero citations, and review sites had 0.9% of citations. The report is from May 2026. DerivateX writes that review profiles are still worth it, but in other channels: they shape the buyer’s perception and conversion.
When it cites a review site, ChatGPT often brings the rating along. In the Seer study, it used the Trustpilot rating or stars in 59% of the answers that cited it. For brands without a profile, it even wrote that there were no independent reviews of them on Trustpilot. The search ChatGPT does before answering is explained in the guide on how ChatGPT search works.
In Brazil, the only measurement of Reclame Aqui in ChatGPT we found is from Promptado, a Brazilian platform that sells a brand monitoring tool for AI models. In March 2026, Reclame Aqui was among the sources of 6.5% of ChatGPT conversations about 204 Brazilian brands, on a sample of 28,938 conversations, 26,827 of them with sources. The questions were about the brands’ products, services and reputation. In murmur’s studies, with category questions, Reclame Aqui appeared in 11 of the 1,275 ChatGPT answers: in 9 among the sources and in 2 only by name.
In Claude, review sites appear more than in ChatGPT
On Claude, review sites appear more than on ChatGPT in the two measurements we read with the same questions asked to both AI models. In Orbit Media’s, Bill Widmer followed the sources of four AI models across 72 business-to-business purchase questions, through August 2026. Clutch was the third-party domain most cited by Claude, and ChatGPT preferred primary sources. The author states that one of the brands measured, Semrush, is his client. In murmur’s, Reclame Aqui appeared in 120 of Claude’s 1,273 answers (9.4%), often as a piece of advice: check the company’s reputation there before hiring.
Otterly, which sells brand monitoring in AI models, did not break the review site out among the 11 categories it used. Almost two thirds of Claude’s 379,321 citations in June 2026 went to company and product sites, in software and technology questions. The differences between Claude and the other AI models in the choice of sources are in the guide on how to get your company cited by AI models.
On Google AIO, the review site appears more when the search asks for a review
On Google AIO, the review site appears among the sources more often when the search asks for a review. SE Ranking, which sells an SEO tool, looked at 22,729 commercial searches with an AI summary in the United States, on 1 December 2025. In 34.5% of the summaries there was at least one review site among the sources: 49% in searches with “review” and 17.1% in those with “best” and “top”. Only 4.3% of the summaries wrote the site’s name in the text, and in murmur’s studies the name appeared in 1.6% of Google AIO answers.
Ahrefs’ September 2026 list, which also sells an SEO tool, covers all subjects in the United States and points the other way. No software review site is among the 50 domains most cited in AI Overview, and Trustpilot appears in 35th place.
On Perplexity, G2 and Trustpilot are among the most cited sources
On Perplexity, G2 and Trustpilot are among the most cited sources in the 2026 measurements made in the United States. Peec AI, which sells brand monitoring in AI models, counted 30 million sources and found G2 among Perplexity’s five most cited. In Ahrefs’ September 2026 list, Trustpilot rose 10 places on Perplexity, to 26th. In Orbit Media’s count, the most cited were LinkedIn, YouTube and Reddit. In murmur’s first studies, the only ones with Perplexity, it brought a review site in 41 of 372 answers, almost always Reclame Aqui. The step-by-step for this AI model is in the guide on how to appear in Perplexity.
When the AI brings data from Reclame Aqui, it repeats the reputation the site calculates
When it brings data from Reclame Aqui, the AI repeats the reputation the site calculates, the consumer score and the percentage of complaints resolved. In murmur’s studies, ChatGPT and Google AIO did this in some of the answers that name the site, sometimes with the RA1000 seal. Claude, in many of them, advised checking the reputation there. ChatGPT even wrote “nos últimos 6 meses” (“in the last 6 months”), which is the period Reclame Aqui shows first.
The reputation comes from the rating the consumer gives when the complaint is closed, in four numbers: complaints answered, solution index, who would do business again and consumer score. The solution index says how many consumers reported that the problem was solved. Companies that answer fewer than 50% of complaints are marked Not Recommended, whatever the score. The RA1000 seal requires, among other criteria, 90% of complaints answered and resolved. In insurance, the complaint starts with the claim, and what the insurer publishes about it is in claim and Reclame Aqui.
Reclame Aqui says it calculates reputation automatically, without manual interference, and denies agreements with AI developers:
“O Reclame AQUI não possui parcerias comerciais com desenvolvedores de IA (como OpenAI, Google, Anthropic ou Microsoft) para alterar, controlar ou favorecer respostas de ferramentas de terceiros.”
Original quote in Portuguese, preserved as published. In English: “Reclame AQUI has no commercial partnerships with AI developers (such as OpenAI, Google, Anthropic or Microsoft) to alter, control or favor responses from third-party tools.”
On 6 October 2026, the site’s robots.txt, the file that tells crawlers what they may read, did not name any AI crawler. The rules for all bots closed off only a few paths, such as the site search and the complaint page.
In a survey by Reclame Aqui itself, released on 1 October 2026, 87% of consumers plan to use AI for their Black Friday purchases. Of those, 64% would disregard an AI recommendation for a store with a bad reputation. The article does not give the sample size, and the 2025 edition surveyed 3.3 thousand consumers on Reclame Aqui’s own website.
For those who sell to businesses, G2, Capterra and Clutch appear in some measurements and vanish from others
For those who sell software or services to other businesses, G2, Capterra and Clutch appear in some measurements and vanish from others. G2 announced in January 2026 the acquisition of Capterra, GetApp and Software Advice. On Google AIO in the United States, G2 took 23.1% of the links to review sites in SE Ranking’s count. In Ahrefs’ six September 2026 lists, G2, Capterra and Clutch are outside the 50 domains most cited. The lists cover the United States and all subjects. G2 keeps a page written also for AI assistants such as ChatGPT and Claude, which explains how it checks who is reviewing.
How to ask for reviews and take care of Reclame Aqui within each site’s rules
To take care of reviews and Reclame Aqui, ask every customer for a review, with nothing in return, and answer and resolve every complaint. Each site writes its own rule:
| Site | How to ask | What is not allowed |
|---|---|---|
| Google Business Profile | with Google’s link or QR code, to people you have served, answering both good and bad reviews | discount, free product or payment in return; asking only those who liked it; demanding or pressuring for the review on site; a review target for the team |
| Reclame Aqui | the review comes after the complaint is closed: resolve, answer and ask for the review at the end of the service, with the button in the company area | leaving a complaint unanswered, because if fewer than 50% are answered, the company is marked Not Recommended; suspicious activity puts the reputation Under Review |
| Trustpilot | invite all customers the same way, with or without a paid plan | discount, coupon, raffle, refund or gift; inviting only those who had a good experience |
| G2 | an incentive of up to US$ 100, given without looking at the opinion, with the review marked as incentivized | an incentive that depends on a positive review; collection that leaves out those who might review badly |
On Google, Trustpilot and G2, the common rule is to ask every customer, without picking only those who liked it and without tying anything to the score. G2 accepts an incentive of up to US$ 100, given without looking at the opinion and marked on the review, and Google and Trustpilot prohibit an incentive in exchange for a review. On Reclame Aqui, the review comes from the complaint, and what counts is resolving and answering. This is how Google describes what it allows merchants to do:
“Solicitem ou incentivem a postagem de conteúdo que represente uma experiência genuína, sem oferecer incentivos para isso ou tentar influenciar a classificação ou o conteúdo da avaliação.”
Original quote in Portuguese, preserved as published. In English: “Request or encourage the posting of content that represents a genuine experience, without offering incentives to do so or trying to influence the rating or the content of the review.”
The company’s reply is public, next to the review, and Google reminds you that the customer can change the review after reading the reply. Reclame Aqui advises replying with text written for the case, and not with an automated message. Showing the rating on your own site, with how many reviews, from when and where it came, helps people who check, as the guide on how to get your company recommended by AI advises. Marking that rating with structured data, the code that describes the page to machines, does not earn a star in Google search when the company controls the reviews about itself.
Buying or inventing reviews is another path, with risks for the brand and the site, covered in the guide on manipulating what ChatGPT recommends. Reviews are one of the areas of the step-by-step guide to make your company appear in ChatGPT. The other external sites each have their own guide: Reddit and Quora, Wikipedia and press relations. The calculation for all types of site is in the guide on which sites most influence ChatGPT in Brazil.
murmur writes the pages that answer the buyer, and the case studies have a name and a number
murmur writes the pages that answer the buyer’s questions and publishes them on a subdomain of the client’s brand, such as guia.suaempresa.com.br. The content is entirely the client’s, even after the contract ends.
Each murmur case study gives the client’s name, the measured number and the date of the measurement. ChatGPT recommended Mindo, a motion design studio from São Paulo, in 12% of 134 questions in June 2026, and in 23% in July. In the case of Visio, operations software for retail, the same 135 questions were asked in June and in July 2026. Google AIO answers with the brand went from 0% to 45%, and ChatGPT answers from 2% to 22%. At FotoRegistro, a photo printing e-commerce, ChatGPT cited the guide for the first time on 24 July 2026, 30 days after the guide launched. The fourth case study is Fly Vet, a veterinary marketing agency.
At murmur, the same person handles the brand from start to finish. They study the company and the sector and see who AI puts in its place. They track the presence week after week, on a dashboard included in the service. When AI gets something wrong about the brand, murmur helps the company correct the information and position the brand the right way.
Frequently asked questions
Does the company’s reply to a review count for AI?
The company’s reply is public text, next to the review, and none of the research we read measured whether AI uses it. Reclame Aqui says that all the public content on the site is indexed by Google and by other search engines. On Reclame Aqui, replying also feeds into the reputation that AI repeats.
Can I offer a discount or a gift in exchange for a review?
On Google and Trustpilot, no. Google prohibits payment, discounts and free products or services in exchange for reviews, and Trustpilot prohibits discounts, coupons, sweepstakes, refunds, gifts and any other benefit. The path that works on any site is to ask every customer you served, with nothing in exchange and without cherry-picking the ones who liked it.
Do I need to be on G2 and Capterra?
It depends on what the company sells and on the buyer’s question. For software sold to companies, G2 and Capterra appear among Google AIO sources in the United States and disappear from DerivateX’s ChatGPT recommendations. In your sector, the answer is in the sources AI models show for the purchase question, and the guide on how to measure your brand’s presence in ChatGPT teaches you how to check.
Does Reclame Aqui carry more weight than Google reviews?
No study we read compares the two. In Promptado’s calculation, Reclame Aqui was among the sources in 6.5% of ChatGPT conversations about 204 Brazilian brands. In murmur’s September 2026 studies, it appeared in 120 of the 1,273 Claude answers, often in advice to check the reputation there. Google Business Profile reviews, which also appear on Maps, are the subject of the guide on appearing in ChatGPT when someone looks for a service in your city.
What if the complaint or the review is fake, made by a competitor?
Each site has its own path. On Reclame Aqui, the company requests moderation through its page, and the categories include third-party complaint, inappropriate content and fraud case. Google’s policy prohibits posting content in a competitor’s space to harm their reputation, and the review can be reported. If AI already repeats the false information or an old complaint, the fix starts at the source, in the guide on correcting wrong information that ChatGPT says about your company.
Where to start, and where murmur comes in
To find out whether reviews and Reclame Aqui influence AI in your market, start with two questions. Ask ChatGPT, Claude and Google AIO the purchase question for your sector and the reputation question about your company, and note which review sites appear in the sources. Most of murmur’s sector studies from September 2026 list the sites that ChatGPT and Claude consulted in that market. For a first read of your market, murmur’s free diagnostic asks for the site URL and delivers the AI answers, with no sign-up.
Read also
- Manipulating what ChatGPT recommends: it worked in a test and puts your site at risk
- How to get your company recommended by AI, not just cited
- How to appear in ChatGPT when someone searches for a service in your city: the six steps
- How to correct wrong information ChatGPT says about your company: the step by step
- Sites that influence ChatGPT in Brazil: company sites lead, and Reddit fell
References
- Seer Interactive, “Study of 800K AI Responses: How Review Profiles Shape Brand Presence in AI Search”, by Nick Haigler, 14 May 2026. 804,491 answers from ChatGPT, Google AI Mode, Gemini and Perplexity, from 1,926 brands in 8 sectors; study commissioned by Trustpilot, which sells plans to the reviewed companies; Seer writes that the relationship is one of correlation. https://www.seerinteractive.com/insights/study-of-800k-ai-responses-how-reviews-shape-brand-presence-in-ai-search (accessed on 6 October 2026).
- G2, “Do More G2 Reviews Mean More AI Visibility? Insights from 30k Citations”, by Kevin Indig, 23 October 2025. 30,000 citations in AI models measured by Profound, across 500 software categories; reviews from the last 12 months and citations from the last 4 weeks; R2 of 0.009. G2 sells services to the reviewed vendors. https://learn.g2.com/do-more-g2-reviews-mean-more-ai-visibility (accessed on 6 October 2026).
- Xi Chu and Yupeng Hou, “Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems”, arXiv, 16 June 2026, revised on 21 August 2026. Laboratory test with invented brands, three models used via the API, without search. https://arxiv.org/abs/2606.17443 (accessed on 6 October 2026).
- Surfer, “The 50 Most-Cited Domains in ChatGPT (September 2026)”, by Paulina Kaleta, updated on 6 October 2026. 47,213 questions and 1,229,618 ChatGPT answers, from 1 to 28 September 2026; the questions come from Surfer users, and Surfer sells an SEO and AI monitoring tool. https://surferseo.com/blog/most-cited-domains-chatgpt/ (accessed on 6 October 2026).
- DerivateX, “B2B SaaS AI Citation Study: How ChatGPT Recommends Software”, 31 May 2026. 40 software categories, one purchase question each, ten rounds, 233 recommendations of 219 tools; DerivateX is an SEO and GEO agency for software companies. https://derivatex.agency/report/b2b-saas-ai-citation-study/ (accessed on 6 October 2026).
- Promptado, “De onde o ChatGPT tira informações sobre empresas no Brasil? [Março/2026]”, 4 April 2026. 85,213 conversations in six AI models about 204 Brazilian brands, 28,938 of them in ChatGPT; Promptado is a Brazilian platform that sells brand monitoring in AI models. https://blog.promptado.com/de-onde-o-chatgpt-tira-informacoes-sobre-empresas (accessed on 6 October 2026).
- Bill Widmer, “I Tracked 13,184 AI Citations Across 4 LLMs: Here’s What I Learned”, Orbit Media, 2 September 2026. 1,765 answers from four AI models to 72 purchase questions about three brands that sell to companies, with data up to 23 August 2026. Orbit Media sells website creation with an SEO focus, and the author states that Semrush, one of the brands, is his client. https://www.orbitmedia.com/blog/ai-citation-sources/ (accessed on 7 October 2026).
- Otterly, “Claude AI Citations Study: How to Get Cited in 2026”, 30 June 2026. 379,321 Claude citations across 16,406 domains, in June 2026, on software and technology questions; Otterly sells brand monitoring in AI models. https://otterly.ai/blog/claude-ai-citation-study/ (accessed on 6 October 2026).
- SE Ranking, “Despite 90% Traffic Loss, Review Platforms Top AI Overview Citations”, by Yulia Deda, 29 January 2026, updated on 20 April 2026. 30,000 commercial keywords, 22,729 with an AI summary, in the United States, on 1 December 2025; SE Ranking sells an SEO tool. https://seranking.com/blog/review-platforms-in-ai-overviews/ (accessed on 6 October 2026).
- Ahrefs, “The 50 Most-Cited Websites in Google AI Overviews (September 2026)” and “The 50 Most-Cited Websites in Perplexity (September 2026)”, by Si Quan Ong and Ryan Law, updated on 2 September 2026. United States, all subjects; the share is the portion within the sum of the 50 domains. https://ahrefs.com/blog/most-cited-domains-ai-overviews/ and https://ahrefs.com/blog/most-cited-domains-perplexity/ (accessed on 6 October 2026). The other four lists, from the same date: https://ahrefs.com/blog/most-cited-domains-in-chatgpt/, https://ahrefs.com/blog/most-cited-domains-ai-mode/, https://ahrefs.com/blog/most-cited-domains-gemini/ and https://ahrefs.com/blog/most-cited-domains-copilot/ (accessed on 7 October 2026).
- Peec AI, “Top domains cited by AI search: Analysis based on 30M sources”, by Tomek Rudzki, 31 March 2026, updated on 21 September 2026. 30 million sources from five AI models, in the United States; Peec AI sells brand monitoring in AI models. https://peec.ai/blog/top-domains-cited-by-ai-search-analysis-based-on-30m-sources (accessed on 6 October 2026).
- Reclame Aqui, “Conheça os indicadores de reputação do Reclame AQUI”, published on 21 December 2022 and updated on 23 July 2026. https://blog.reclameaqui.com.br/conheca-os-indicadores-de-reputacao-do-reclame-aqui/ (accessed on 6 October 2026).
- Reclame Aqui, “Manual de Reputação”, undated page. https://manual.reclameaqui.com.br/reputacao-no-reclame-aqui (accessed on 6 October 2026).
- Reclame Aqui, “Como as LLMs decidem se uma empresa é confiável?”, 7 July 2026. The numbers of mentions in AI models come without the method behind the count. https://blog.reclameaqui.com.br/como-llms-decidem-confianca-marca-reclame-aqui/ (accessed on 6 October 2026).
- Reclame Aqui, “Reclame AQUI para Empresas”, frequently asked questions, undated page. https://conteudo.reclameaqui.com.br/faq-para-empresas (accessed on 6 October 2026).
- Reclame Aqui, “Veja como conseguir avaliações dos consumidores no Reclame AQUI”, 10 April 2018, updated on 9 March 2026. https://blog.reclameaqui.com.br/como-conseguir-avaliacoes-dos-consumidores-no-reclame-aqui/ (accessed on 6 October 2026).
- Reclame Aqui, robots.txt. https://www.reclameaqui.com.br/robots.txt (accessed on 6 October 2026).
- Exame, “Black Friday: 64% dos brasileiros descartam sugestões de IA se a reputação da loja for ruim”, 1 October 2026, about a Reclame Aqui study; the article does not give the sample base. The 2025 edition surveyed 3.3 thousand consumers on the Reclame Aqui website: https://blog.reclameaqui.com.br/black-friday-2025-63-dos-consumidores-nao-identificam-golpes-feitos-com-ia/. https://exame.com/bussola/black-friday-64-dos-brasileiros-descartam-sugestoes-de-ia-se-a-reputacao-da-loja-for-ruim/ (accessed on 6 October 2026).
- G2, “G2 Acquires Capterra, Software Advice, and GetApp”, 29 January 2026. https://company.g2.com/news/g2-acquires-capterra-software-advice-getapp (accessed on 6 October 2026).
- G2, “Community Guidelines”, undated page. In the original: “G2 limits the value of any incentive offered in exchange for a review to $100 USD.” https://legal.g2.com/community-guidelines (accessed on 6 October 2026).
- G2, “About G2’s Reviews, Data, and Rankings”, updated on September 2026. https://www.g2.com/ai-instructions (accessed on 7 October 2026).
- Trustpilot, “Guidelines for businesses”, September 2026 version. https://corporate.trustpilot.com/legal/for-businesses/guidelines-for-businesses/sep-2026 (accessed on 7 October 2026).
- Trustpilot, “Brands that build trust through reviews increase AI citations from 1% to 75%, earn competitive advantage over ‘invisible’ brands”, press release of 12 May 2026, which says the research was commissioned by Trustpilot and carried out by Seer Interactive in March 2026. https://www.prnewswire.com/news-releases/brands-that-build-trust-through-reviews-increase-ai-citations-from-1-to-75-earn-competitive-advantage-over-invisible-brands-302768554.html (accessed on 7 October 2026).
- Google, Business Profile, “Dicas para receber mais avaliações”, undated page. https://support.google.com/business/answer/3474122?hl=pt-BR (accessed on 6 October 2026).
- Google, Business Profile, “Gerenciar avaliações de clientes”, undated page. https://support.google.com/business/answer/3474050?hl=pt-BR (accessed on 6 October 2026).
- Google, Google Maps, “Conteúdo proibido e restrito”, undated page. https://support.google.com/contributionpolicy/answer/7400114?hl=pt-BR (accessed on 6 October 2026).
- Google Search Central, “Dados estruturados de snippet de avaliação”, updated on 17 September 2026. https://developers.google.com/search/docs/appearance/structured-data/review-snippet?hl=pt-br (accessed on 6 October 2026).
- murmur, 51 sector studies from 8 to 29 September 2026: 25 category questions per sector, without brand names, with 1,275 ChatGPT answers, 1,273 Claude answers and 1,258 Google AIO answers; Perplexity only in 16 of them, with 372 answers. The count uses the same base as the guide about the websites that influence ChatGPT in Brazil, with the Google AIO sources read from the HTML of the answer. https://murmur.marketing/estudos/.
- murmur case studies: https://murmur.marketing/en/cases/mindo/, https://murmur.marketing/en/cases/visio/, https://murmur.marketing/en/cases/fotoregistro/ and https://murmur.marketing/en/cases/fly-vet/.