AI Search Visibility for Dental Clinics in Dubai and Abu Dhabi: What Is Proven, What Is a Myth, and Who AI Actually Recommends

To be recommended by AI, a dental clinic in the UAE needs what Google already asks for. Keep the Google Business Profile complete and current, with fresh reviews. Give every dentist a page with credentials and publish original, expert content. Let the AI search crawlers OAI-SearchBot and PerplexityBot read the site. These are conditions, not guarantees: they remove obstacles, they do not promise a mention.

Here is what those three rest on. A current Business Profile with reviews is a documented lever for local visibility, including in AI answers: Google says so in its guide to AI search and its local ranking tips. Its direct effect on AI mentions is measured only by vendors and single cases: BrightLocal, which sells listings management, and an agency study of 118 practices. Dentist pages and original content are a platform requirement for medical topics, set out in Google's Search Quality Rater Guidelines. Their effect on AI citations is not measured. Crawler access follows the platforms' own documentation from OpenAI and Perplexity. Google-Extended is a separate switch for Gemini, not for Google Search, as Google's crawler documentation states; see section 5.

There is no special "AI file", markup or rewriting that moves the needle. Google says so in its AI search guide, and the same guide states that Google does not use llms.txt. There is no evidence that any other system does: that is the opinion of John Mueller, and vendor data from SE Ranking, which sells SEO tools, points the same way. We measured which clinics AI names for Dubai and Abu Dhabi: 16 prompts (12 patient, 4 B2B) in 6 systems over 3 runs, 22 to 29 October 2026. What separates the leaders from clinics with the same rating in the same search results is press, curated lists and awards, not llms.txt, markup or site speed. That is an observation on 3 + 3 clinics in section 10, not a proven cause. Single cases and opinions in the main text are marked as such and are not in the checklist.

Sourcing. Next to every claim we say what it rests on: an official document of a platform or regulator, an independent study with a disclosed method, vendor data, a single case or small sample, or expert opinion. When we say something was not found, none of our three research passes (three independent search passes by different tools) found it. Facts checked 22 October 2026, AI answer panel 1 November 2026. The checklist is linked in "What next" at the end of the article.

Contents. 1 What AI visibility means. 2 How AI picks a clinic. 3 Business Profile and reviews. 4 The website. 5 AI crawlers. 6 Mentions. 7 Regulation. 8 Measurement. 9 Myths. 10 Who AI recommends. 11 Practitioners. 12 Counterpoints. What next. FAQ.

AI visibility is how often an AI assistant names or cites the clinic across many prompts. It is not a "position". That is the framing of SparkToro, which sells audience research software, and an independent ACM Web Science audit documents the gap between organic results and AI Overview sources. GEO (generative engine optimization) and AEO (answer engine optimization) are ordinary SEO plus brand presence where AI pulls answers from. Google says as much in its AI search guide and on its page about AI features and your website. Google shows AI Overviews on most health queries. Vendor data from BrightEdge puts them on 89 percent of tracked healthcare keywords, country not stated. The ACM audit found them on 87.7 percent of 900 US health queries. A separate SE Ranking study, cited in SE Ranking's health report, found them on more than 82 percent of health searches in Germany. There is no UAE figure.

On local queries AI Overviews are rare, by vendor counts: 0 percent on "near me" queries in one BrightEdge sample, and 15 percent on "service + city" queries in another from Whitespark.

Terms, once. AI Overviews: the generated summary Google places above results with a source list. AI Mode: Google's chat-style search tab that runs sub-queries and cites pages. ChatGPT search, Gemini, Perplexity and Copilot: assistants that answer with citations, fed by their own crawlers or by Bing's index.

Do patients use it? About a third of US adults have turned to AI for health, according to two independent surveys: 32 percent of 1,343 adults in a KFF poll and 34 percent of 3,488 adults in a Pew survey. We found no UAE survey of patients using AI for health or for choosing a dentist. A general-population online survey is discussed in section 12.

AI referrals are under 0.15 percent of all visits, even after growing 66 percent in 2025. That comes from a global sample by Semrush, which sells SEO software. Clicks from Google shrink where an AI summary appears. In an independent Pew panel of 900 US adults (68,879 Google searches in March 2025, 12,593 of them with an AI summary), users clicked a result in 8 percent of visits with a summary versus 15 percent without. That figure covers all searches, not only informational ones. On informational queries, vendors measure minus 61 percent organic CTR (Seer Interactive) and minus 58 percent for the top page (Ahrefs). Seer then recorded a rebound from 1.3 to 2.4 percent by February 2026. Clicks compress on informational queries. They do not vanish.

What we do not know about the UAE. Every percentage in this article was measured in the United States or Germany. Treat it as a guide to behaviour, not as a fact about the UAE. Our only UAE measurements are our own Google SERP snapshot and the prompt panel in section 10 (3 runs in six systems). No research pass found a UAE patient survey on AI and health, a UAE clinic case with measured AI traffic, or anything in languages other than English (Arabic, Hindi, Urdu, Russian). Google prompts ran with hl=en only; the other systems got English prompts.

Good SEO is good GEO, or AEO, AIO, LLM SEO, or LMNOPO. Danny Sullivan, Director, Google Search, 2025-08-28, source

So far, AI isn't making the inroads into local in the same way it is in other SEO corners. Miriam Ellis, local search consultant, 2025-11-06, source

Table 1. AI surfaces a UAE patient meets

Audience figures are not system-specific. About a third of US adults have used AI for health, according to independent surveys by KFF and Pew. There is no UAE figure.

Refuted by: a UAE patient survey with AI absent from the channels used to pick a dentist, or stable clinic lists between panel runs.

Neither Google nor OpenAI publishes the selection logic (see Google's page on AI features and your website). Vendor audits put profiles and directories first. Google Business Profile is the largest single platform among sources: 28.5 percent of about 1.97 million local citations in a US, UK and Australia sample, according to BrightLocal, which sells listings management. Business websites (not clinics specifically) are 42 percent of the same citations, in the second part of the same study. Directories as a group are 52.6 percent of healthcare citations in a different, global sample of 6.8 million citations from Yext, another vendor. The two BrightLocal shares come from one dataset. The Yext share comes from another and cannot be added to them. AI Overview sources also differ from organic results. On 900 US health queries, an independent ACM Web Science audit found that the average domain overlap between AI citations and the organic top 10 was only 46.2 percent.

Cautions. The BrightLocal and Yext samples use different models and definitions, so shares cannot be added. Yext's websites category is over 40 percent of citations, in line with BrightLocal's 42, and its "local websites" at 8 to 20 percent are a subcategory, according to Yext's research. Both sell listings management. The largest dentistry-related measurement by number of answers comes from vendor Javvo. Of 3,234 AI answers about dental, podiatry and chiropractic practices in five US metros, AI cited only 16 percent of practices across the three specialties and 13 percent for dentistry alone. An agency's own study of 118 US practices found 82 percent of Google Maps top-10 practices named by at least one of three engines, against 47 percent of positions 11 to 20. It was a single run. In the same study, raw review counts were nearly flat within a tier, and star ratings, compressed at 4.7 to 5.0, did not predict mentions.

There isn't one source you can optimize for and call it done. Mike Hawkes, Senior Content Marketing Manager, BrightLocal (vendor), 2026-08-26, source

Table 2. Source types in local AI answers

Shares come from different corpora and cannot be added.

Refuted by: a gl=ae panel where clinics without a Business Profile are named consistently.

Google ranks local results on relevance, distance and prominence, and says plainly that more and better reviews can help. Google's guide to generative AI search and its page on AI features name an up-to-date Business Profile as a lever for local visibility in AI answers. The direct link "more reviews, more AI mentions" is not supported. In a single case documented by the agency Elevate DDS, one clinic with a 5.0 rating and 500+ reviews had 22.2 percent overall AI visibility and zero for implants, whitening and extractions. An agency's own study of 118 US practices by Identity Dental Marketing found that Maps position tracked AI mentions, while review counts barely mattered and ratings did not predict them.

Patient-side figures are US-only and come only from vendors. According to vendor data from rater8, 75 percent would not book a doctor rated below 4.0. Among patients who actively searched for a doctor in the past year, 36 percent cited AI tools such as ChatGPT as an influence, next to Google search at 34 percent (question wording as in the press release). Another vendor, BrightLocal, found in a survey of 1,002 US adults that 44 percent want a review under a month old.

The official rules are on record. Google's review snippet documentation says self-published reviews with Review markup earn no stars. Since 15 May 2026, Google's spam policy explicitly names attempts to manipulate generative AI responses in Google Search. Fake reviews, bought mentions and self-ranking lists are not named in its text. DHA's standard bans unverified testimonials and requires documented consent for any patient in social media advertising (ST-21 scope).

Other platforms work the same way. Copilot answers from Bing's index, and Microsoft recommends Bing Places for Business to keep address, hours and contact details current and eligible for inclusion in AI-generated responses. It reports no measured effect, and its post does not mention ChatGPT. On the role of Okadoc, Practo and similar directories in UAE AI answers we found no data, so the question goes to the panel. Instagram and LinkedIn sit in Google's UAE top 10 for "how to get dental clinic recommended by ChatGPT" and "dental clinic AI visibility", with no measured AI effect (our own snapshot of Google UAE results).

I do see ChatGPT referencing old/outdated citations but I do not anticipate that will last. Joy Hawkins, owner, Sterling Sky, 2025-11-06, source

Table 3. What is proven about reviews

Refuted by: a gl=ae panel with no relation between rating and mentions, or leaders without a profile.

The platform requirement is documented: original, expert content and dentist pages with credentials, because dentistry is YMYL for Google. That comes from Google's rater guidelines (YMYL section 2.3, E-E-A-T section 3.4, edition of 11.09.2025). The effect on AI citations is not measured. DHA's standard requires the advertising license number in official social accounts (clause 5.2). Showing the dentist's license on the website is a recommendation, not a DHA rule for web pages. Separate treatment pages and FAQ blocks are practitioner advice without measurement, an opinion from the Whitespark expert panel. That panel did not rate published prices, so they remain our untested recommendation. Dentist schema is valid and useful for rich results, according to Google's documentation and schema.org. But vendor data from Ahrefs show that adding JSON-LD to 1,885 pages barely moved AI citations: AI Overviews minus 4.6 percent, AI Mode plus 2.4 percent, ChatGPT plus 2.2 percent.

Google says you can skip "chunking", rewriting "for AI", special markup and llms.txt. Booking forms and chat collect health data. The UAE government portal lists the relevant acts; we did not open their texts in this run. At the federal level, Federal Law 2/2019 on ICT in health and the PDPL (Federal Decree Law 45/2021) apply. The DIFC free zone has its own data protection law (DIFC Law 5/2020, listed on the portal). ADGM is not listed on the portal (source to add). DHCC is still to be checked. This is a minimum check for counsel, not a compliance recipe.

Focus on your visitors and provide them with unique, satisfying content. John Mueller, Google Search Relations, 2025-05-21, source

Table 4. Website elements

Refuted by: a panel where dentist pages are cited no more than templated pages, or a controlled test showing a lift after adding FAQ or prices.

One opinion from a Russian-speaking practitioner (translated, not an expert): when he searched for implants, ChatGPT asked him about price and budget first. A site with "call us" instead of a price list is unlikely to be recommended, he argues (V-05, 12:51).

Allow OAI-SearchBot, which decides whether a site can appear in ChatGPT search, and PerplexityBot. GPTBot governs only training. ChatGPT-User and Perplexity-User are user-triggered agents that may ignore robots.txt anyway. That is how OpenAI and Perplexity document their own crawlers. Google-Extended controls whether content is used to train Gemini models and for grounding in Gemini Apps and Vertex AI. Google's documentation says it does not affect a site's inclusion in Google Search and is not a Search ranking signal. It does not mention AI Overviews. An independent observational study accepted to ACM SIGIR 2026 finds that sites blocking it are cited less often in both Gemini and AI Overviews. That is a correlation, not proof of cause, and for AI Overviews it sits uneasily with Google's statement. Keep it open unless there is a reason not to.

Do not spend on llms.txt. Google's documentation says it does not use it. In a 300,000-domain sample where 10.13 percent had the file, the vendor SE Ranking found no relation to citations. A second vendor, Search Atlas, attributes any visible effect to domain authority. As Google's John Mueller wrote, in his opinion, no AI system currently uses it. Access is a condition, not a factor: permission does not guarantee citation. For Copilot the condition is Bing's index, where the only official citation report outside Google lives, according to Microsoft.

For AI Search we are mostly focused on improving visibility in Google AIOs. We spend a bit of time on ChatGPT. Andrew Shotland, founder, Local SEO Guide, 2025-11-06, source

Table 5. AI bots

A Russian-speaking practitioner (translated, not an expert) argues that ChatGPT takes its results from Bing, so you should add the site to Bing Webmaster Tools (V-08, 09:06).

Brand mentions correlate with AI Overview visibility. Ahrefs, which sells Brand Radar, measured a Spearman correlation of 0.664 across 75,000 brands. YouTube is among the top cited domains in medical AI Overviews. It took 4.43 percent of citations in a German sample of 50,807 queries, according to vendor data from SE Ranking. In a 900-query US audit, an independent ACM WebSci study found it the second most cited domain. But since 15 May 2026 Google's spam policy explicitly names attempts to manipulate generative AI responses, and it treats paid links without rel="sponsored" or rel="nofollow" as link spam. The policy text does not name self-promotional "we are number one" lists, bought mentions or fake reviews. One consultant, Lily Ray, predicts they are the next target; that is her opinion. DHA's standard bans absolute wording ("the best", "unique", "safest", "the only": DHA 6.11) in the clinic's own social media advertising. ST-21 does not address third-party lists directly, but paid promotion must be disclosed (6.7).

Wikipedia, YouTube and Reddit together supply 15 percent of AI Overview sources, an independent Pew study found. Yet forums such as Reddit are about 2 percent of all citations in a global vendor sample from Yext. In a small sample of 100 queries, a brand that ranked itself first in its own listicle was excluded from the AI recommendation 83 percent of the time, Lily Ray found. On Wikipedia pages for clinics we found no data at all. In Google's UAE results for "dental clinic SEO Dubai AI Overviews", a consultant's listicle "Top 10 AI SEO Experts for Dentists in Dubai (2026)", which ranks its own author first, sits fifth. This comes from our own snapshot of the results page, and it is an example of the format, not a recommendation.

Many of the most popular GEO tactics right now are old SEO spam tactics with a new label: mass-produced AI content, self-promotional listicles where a brand ranks itself #1, fake reviews and “inauthentic mentions” bought to influence what AI tools recommend, and automated Reddit posting. Lily Ray, SEO consultant, 2026-10-04, source

All the factors we studied revealed moderate to very weak correlations on the Spearman scale, but that doesn’t necessarily mean they should be overlooked. Louise Linehan, Ahrefs (vendor), 2025-05-26, source

Table 6. Mention channels

Refuted by: a controlled test with no mention effect, or leaders with no third-party mention at all.

This section is not legal advice. We cite the rules with links to the primary sources; check how they apply to your clinic with a licensed lawyer in your emirate.

DHA standard DHA/HRS/HPSD/ST-21, Version 1.1, was issued 03.08.2022, took effect 03.10.2022 and has a revision date of 03.08.2027. We verified what follows against the PDF on 22.10.2026. Official clinic accounts show the MOHAP advertising license number (clause 5.2). Result claims are substantiated and carry risks (6.6). Any patient in advertising needs documented consent (6.10). Absolute statements and exaggerated claims are prohibited, and the standard gives its own examples: "unique, one of a kind, the best, exclusive, safest, the only, incomparable, unprecedented" (6.11). The medical director approves promotional content that names the facility (5.3). Patient testimonials without written consent, and comparisons with other treatments that are not scientifically proven, are listed as prohibited content (Appendix 2, items 8 and 9). Before/after photos show the same person, with the same lens and no retouching (9.6), and carry a disclaimer on outcomes (Appendix 2, item 12). The standard covers advertising content in social media and official accounts, not website pages.

DHA circular CIR-2026-00000144 of 17.08.2026 re-issues standard ST-21 v1.1 without changes. The editors verified this on 22.10.2026.

Abu Dhabi has a different regulator. The DoH announced training under its Health Media and Advertising guidelines, and the circular itself (26/2023) is an official document. We could not verify whether the training is mandatory or what the guideline requires: we did not find its text.

A guide by the law firm CMS on advertising medicines and medical devices says health advertising needs MOHAP approval and a license under Ministerial Resolution 430/2007 (reaffirmed by Cabinet Resolution 7/2007). It adds that online advertising of medicines also needs a UAE Media Council license. The guide is not about medical services, so it applies to dental clinics only by analogy. It is legal commentary from May 2024, and we have not checked the status of Resolution 430/2007 in 2026. We could not confirm an enforcement case in dentistry. UAE Media Council rules for paid posts and influencers apply to clinic promotions; this is to be verified for v2. Dubai Healthcare City is said to have its own DHCR advertisement policy. One pass found it mentioned, but we did not find the document itself. No lawyer or regulator is in our expert pool, so this section cites the standard only. Get legal review before publishing anything that counts as advertising.

Table 7. Tactic versus rule

Refuted by: a future DHA edition changing the clauses, or the DoH text widening or narrowing the rules.

Start with the free reports. Google Search Console's AI performance report shows impressions from AI Overviews and AI Mode by page, country and device, for all sites since 31.08.2026. Clicks are not listed in Google's announcement. Bing Webmaster Tools' AI Performance shows Total Citations and Average Cited Pages; Microsoft opened it as a public preview on 10.02.2026. If a report is empty, the site is not cited yet or the data is below the display threshold. Check weekly, and after opening the crawlers allow 2 to 4 weeks.

Then run a pre-registered panel: 10 to 12 prompts, 3 to 5 systems, 3 runs on different days, with share of mentions as the metric. There are two reasons. Across 2,961 runs of 12 prompts, the chance of two identical brand lists is under 1 in 100, according to SparkToro, whose co-author sells tracking. And an independent ACM WebSci study found that Google's AI Overview sources share on average only 46.2 percent of domains with its organic top 10. SparkToro advises at least 60 to 100 runs per prompt. Our 10 to 12 prompts over 3 runs is a minimum for a baseline, not a precise measure.

Paid trackers exist: Ahrefs Brand Radar, Gumshoe, SE Ranking, Semrush AI Visibility Toolkit, BrightLocal Local AI Visibility Tracker, Yext Scout, Seer SeerSignals, LocaliQ and Steady Demand AI Citation Ledger. We list them by name only, with no evidence behind the list. Google's documentation states that third-party trackers have no access to its systems. One vendor, Seer, reports that brands cited in an AI Overview get 35 percent more organic and 91 percent more paid clicks than uncited brands. That is a correlation on the agency's own clients; the cause is not established. If AI names your clinic with wrong hours, prices or dentists, fix the Business Profile, Bing Places and the directories the answer cited. Then recheck in 2 to 4 weeks. Visits that do come from assistants appear in GA4 as referrals from domains such as chatgpt.com and perplexity.ai. Our method: 4queries.com/method.

I’ve been swayed from my initial position and now believe visibility % across dozens to hundreds of prompts run multiple times is a reasonable metric . [sic] Rand Fishkin, co-founder, SparkToro, 2026-01-27, source

Table 8. Measurement tools

Refuted by: a study showing stable AI answers over repeated runs.

Four popular claims have no evidence behind them. The first is llms.txt, and it is a myth: Google's documentation rejects it, and so does vendor data from SE Ranking and Search Atlas. The second is chunking and rewriting pages for AI, which the same Google guide calls unnecessary. The third is self-promotional rankings. In a small sample by Lily Ray, the brand that ranked itself was left out of the AI recommendation in 83 percent of cases, and Google's spam policy does not name such rankings. The fourth is "position in AI" as a KPI, which vendor data from SparkToro and an independent ACM WebSci study do not support.

NAP is a different case. A consistent name, address and phone are needed so AI does not merge you with another clinic. That is entity matching, and it is our hypothesis. Google asks only for complete and accurate profile data. As a separate AI ranking factor, NAP is unmeasured: our hypothesis, not a myth. Google's official local factors remain relevance, distance and prominence. Pricing and treatment pages are sensible for patients, but as AI factors they are unmeasured, an opinion rather than a finding.

Where the advice shows up. Google's own AI Overview for "how to get dental clinic recommended by ChatGPT" (gl=ae; sources dentree.co.uk, reddit.com, revirto.com, serpner.com, youtube.com) recommends "keep directory data consistent", "publish clear pricing details" and "dedicated treatment pages". No measurement stands behind them, according to our own snapshot of the results. In the same snapshot, the top 10 for "dental clinic AI visibility" is agency and tool service pages plus LinkedIn, Instagram and Facebook posts. No domain holds a top-10 place for all four captured queries.

Prompt injection, fabricated reviews, manufactured community signals and citation gaming might end up producing short-term improvements, but they will get eventually addressed, and will trade durable trust for visibility that can backfire. Aleyda Solis, SEO consultant, 2026-09-20, source

Table 9. Myths and the primary source

Results as of 1 November 2026: three runs in six systems on the 16 UAE prompts (Dubai 8, Abu Dhabi 4, B2B 4). Unit of analysis: prompt × system × run. A clinic's share is counted per system; we do not average systems.

• Dubai has a leader, but no system agrees with the others on it. Dr. Joy Dental Clinic was named in all six systems, in 48 of 130 answers (37 percent), and in Google AI Mode in 14 of 24 (58 percent). Pearl Dental Clinic follows at 28 percent in five systems (GigaChat never named it), Dr. Michael's chain at 24 percent (Alice named it in 15 of 24 answers, GigaChat in none).
• Abu Dhabi has two clear leaders. International Center for Dental Excellence (ICDE) was named in 30 of 64 answers (47 percent), in Perplexity in 7 of 8; True Smile Dental Center in 25 of 64 (39 percent). Both appear in all six systems.
• Lists barely overlap. Answers for Dubai named 105 different clinics: four of them in all six systems, 95 in only one or two answers. Abu Dhabi answers named 25 clinics, three of them in all six systems.
• Sources differ by system. Google AI Mode and ChatGPT cite clinic websites most often (drjoydentalclinic.com, truesmile.ae, icdexcell.com); Perplexity leans on directories and medical-tourism sites (zavis.ai, placidway.com, bookimed); Alice and GigaChat mix clinic sites with local listings.
• 4queries.com was not named or cited in any answer.

Panel, runs 1 to 3, six systems, 22 to 29 October 2026. A cell shows the share of a system's answers that named the clinic; systems are not averaged. The 10 names with the most systems are shown, the rest were named in one or two answers. An observation, not a ranking: answers change between runs and systems.

Dubai

Frequent sources: Google AI Mode: www.drjoydentalclinic.com (9), drjoydentalclinic.com (8), fresha.com (8), us-uk.bookimed.com (7), www.drmichaels.com (7); ChatGPT: drjoydentalclinic.com (4), drmichaels.com (4), services.dha.gov.ae (4), noadentalclinic.com (3), pearldentalclinics.com (3); Perplexity: placidway.com (9), zavis.ai (9), nicolasandasp.com (8), us-uk.bookimed.com (8), drjoydentalclinic.com (7); Alice: drmichaels.com (6), dentalclinicdirectory.ae (5), doctify.com (4), dubailondonhospital.com (4), harmonymc.ae (4); GigaChat: abudhabitiming.com (5), drjoydentalclinic.com (5), factorsmile.ae (5), aesthet.ae (4), bayavenuedentalcentre.com (4).

Abu Dhabi

Frequent sources: Google AI Mode: truesmile.ae (7), icdexcell.com (6), synergydental.ae (5), alkhajamedicalcenter.ae (4), bostondentaluae.com (3); ChatGPT: truesmile.ae (6), alsafwamedical.ae (3), centralclinic.ae (2), doctify.com (1), oaissbracescenter.com (1); Perplexity: zavis.ai (6), bestinhood.com (4), truesmile.ae (4), clevelandclinicabudhabi.ae (2), gcr.org (2); Alice: icdexcell.com (6), truesmile.ae (6), hsmcuae.ae (5), bestinhood.com (4), burjeel.com (3); GigaChat: truesmile.ae (11), abudhabitiming.com (5), davincidental.ae (5), drjoydentalclinic.com (5), smilefirstad.ae (5).

Agencies, B2B prompts

Frequent sources: Google AI Mode: boostmylocalbusiness.ai (7), www.linkedin.com (5), adgrow.ae (4), aliseoservices.com (4), www.instagram.com (4); ChatGPT: aidentalagency.com (4), lunasol.ae (4), aimenalabs.com (3), arabseo.ae (2), digitalarab.agency (2); Perplexity: lunasol.ae (6), curiousmindlabs.com (5), seotechexperts.ae (5), adjoltz.com (4), boostmylocalbusiness.ai (4); Alice: curiousmindlabs.com (8), newpatientsflow.com (3), yma.health (3), agenticlabs.ae (2), docrankr.ae (2); GigaChat: rush-agency.ru (6), curiousmindlabs.com (4), vc.ru (3), almond-medical.ru (2), digitalfloss.com (2).

What the leaders have that clinics from the same search results do not

To see what the leaders share, we compared the three panel leaders (ICDE, True Smile, Dr. Joy) with three "outsiders": dental clinics in the same cities whose own website ranks in Google's top 20 for patient queries (gl=ae), but which AI named at most once across all systems and runs (DentisTree in Dubai, Harley Street Dental Center and Gentle Dental Clinic in Abu Dhabi). The selection rule was written down before any feature was collected; features were captured the same way on 30 October and 1 November 2026, each with a file and a date.

Leaders vs outsiders, UAE

Observation: every leader has at least one third-party mention (a press article, a place in a curated list or an industry award); none of the outsiders has. Two of the leaders' press articles are old (2012 to 2016) and two leaders hold a 2026 award, so the signal is "named somewhere outside the clinic's own channels", not "fresh press". Ratings, review volume, review platforms and dentist pages are shared by both groups: they look like an entry ticket that does not by itself turn into an AI mention. llms.txt and Dentist schema do not separate the groups (outsiders have llms.txt more often), and ranking first in Google does not either. The same comparison for Moscow and St. Petersburg gave the same picture, with active YouTube channels as an extra trait of the Russian leaders.

Limits. This is an observation on three plus three clinics, not a proven cause: branded demand was not compared, and directories that block automated access (Google Maps, Practo) were read from search snippets. None of this argues for buying placements: Google treats manipulation of generative answers as spam, and DHA clause 6.11 bars "the best" claims in a clinic's own advertising.

Method, pre-registered before the first query. 16 prompts for the UAE (12 patient, 4 B2B), part of a 32-prompt panel across two countries. Patient prompts: 8 for Dubai (AE-01 to AE-08: implants, veneers cost, emergency near Dubai Marina, reviews near Business Bay, pediatric, Invisalign, comparison, top 5 by reviews) and 4 for Abu Dhabi (AE-09 to AE-12: implants, crown cost, Al Reem Island, top 5). B2B prompts (AE-B2B-01 to 04): whom AI names as an agency and what it advises. Systems captured: Google AI Mode, ChatGPT, Gemini, Perplexity, Alice (Yandex) and GigaChat. Dates: Google AI Mode 22, 25 and 29 October 2026; ChatGPT, Gemini, Alice and GigaChat 25, 27 and 29 October; Perplexity 22 October in the logged-out interface, then 27 and 29 October via its API (sonar-pro, country set as a parameter). Country: Google via gl=ae and hl=en (non-UAE answers discarded); other systems via the city in the prompt, actual country logged as an observation.

New chat for every run. Logged out: Google AI Mode and Perplexity; ChatGPT, Gemini, Alice and GigaChat after the owner's login into a clean account with memory off, model recorded as shown in the interface at capture time. Deviations from the plan: Copilot was not captured; ChatGPT does not answer logged out, so the panel uses logged-in answers; Gemini returned no answer for 9 Dubai and 4 Abu Dhabi pairs and Perplexity for 5 and 4 (column headers show what was captured).

Clinic share = mentions / captured prompt × run pairs per system, plus the number of systems naming the clinic; "named" = the clinic's name or domain appears in the answer. Spelling variants of one clinic (Dr Joy Dental Clinics and Dr. Joy, ICDE and its full name) are merged by a dictionary. Where brands were not extracted at capture (Perplexity via API, some browser answers), they were written out from the answer texts, and every name was checked against its text by a script. Log row: prompt_id, system, model_version, run, datetime, country_set, country_actual, brands_in_order, sources_cited, screenshot.

Done after the runs: leaders by share of mentions, stated neutrally, without "best" (DHA 6.11); a leader is a clinic named in every system of its city with the highest overall share; a reasons card for three leaders and three outsiders on 14 observable features (rating and review count, platforms, dentist pages, prices, FAQ, Dentist schema, media mentions, YouTube, Wikipedia, llms.txt, bot access, site speed, domain age, branches); overlap between systems. Not computed in this version: the share of answers citing directories versus clinic sites (a check on Table 2) and Maps top 10 versus positions 11 to 20 (a Dubai repeat of an agency's own measurement).

Expectations before the runs and what happened. Lists would overlap weakly between systems (SparkToro compared repeated runs within each tool; ACM compared Google's organic and AI Overview sources): confirmed, four Dubai clinics out of 105 were named by all six systems. Nobody has published who AI recommends in Dubai or Abu Dhabi before (we found no such publication); the largest dentistry measurement suggests AI cites a minority of practices: 13 percent of dental practices in the US sample, 16 percent across three specialties (vendor data from Javvo).

Refuted by: a clinic without any third-party mention that systems name consistently, or a leader without one, in a sample larger than three clinics per city.

The video corpus collected for this research is Russian-language (8 transcripts of 10). It illustrates the Russian market, not the UAE. Two observations about ChatGPT and Bing are kept as illustrations in sections 4 and 5 (V-05 at 12:51, V-08 at 09:06). The editor translated them and marked them as practitioner opinion. None of the speakers passed the expert rubric, and several sell the service they describe. We did not collect English-language practitioner videos for the UAE. That is an open item for v2.

We found no independent case of a UAE clinic with measured AI traffic. We found no UAE patient survey on choosing a dentist via AI. And we found no one who had published who AI recommends in Dubai or Abu Dhabi before this panel. One survey comes close. The communications agency Edelman ran an online survey (16,006 respondents in 16 countries, 1,000 per country, 28 February to 11 March 2026) and reports that "nearly six in ten" UAE respondents use AI to manage their health. The press relayed this as "over half" and 59 percent. The Khaleej Times report states only the overall sample of 16,000+ in 16 countries, not the UAE sub-sample. Edelman's live page was unreachable when we checked, so we saw only an archived copy. A general-population panel also says nothing about choosing a dentist, so we do not use it as a patient figure. The independent studies in the run are about the US: Pew's surveys on AI chatbots for health and clicks on AI summaries, KFF and ACM WebSci. The arXiv study, accepted to ACM SIGIR 2026, is observational: it shows a correlation, not a cause.

Where the evidence contradicts itself. Clicks first. An independent Pew study measures a drop, and so does vendor data from Seer and Ahrefs. Other vendor figures point the other way. Semrush sees zero-click falling from 33.75 to 31.53 percent on the same keywords, and Seer's 2026 update records a rebound.

Citation shares, on a closer look, do not conflict. Both figures are vendor data. BrightLocal's 42 percent for business websites matches Yext's websites category at over 40 percent. Yext's 8 to 20 percent is the narrower "local websites" subcategory. Patient choice is a similar case. The same rater8 release gives 36 percent (AI as an influence among patients who searched for a doctor) and 39 percent (AI as the top digital influence among patients who switched doctors). These are two questions from one vendor, so quote the question wording.

Reviews as a direct AI factor are not supported. In a single agency run, Maps position tracked AI mentions (82 versus 47 percent), while review counts barely mattered and ratings did not predict them. In another single case, a 5.0-rated clinic had zero visibility for its main services.

Risk, not tactics. After a Guardian investigation into AI Overviews on liver blood tests, Google removed some of its medical summaries, the Guardian reported on 11 January 2026. Errors in AI health answers are a trust problem for patients. Disclosure: 4queries.com sells an AI visibility audit, so we have an interest in clinics measuring this. We say what each claim rests on so that this interest can be checked. "One tactic works the same in Russia and the UAE" is also unproven: regulators, directories and reports differ.

AI Overviews are designed for speed, not accuracy, and that leads to mistakes in health information, which can be dangerous. Gina Neff, professor of responsible AI, Queen Mary University of London, 2026-02-16, source

Table 10. Claims and status

The full evidence table, claims-matrix.csv with 79 claims, grades and counter-evidence, accompanies this article. The same method applied to Moscow and St. Petersburg is in the companion article on klientlab.ru: the same source base, a different market.

This week, about an hour: checklist items 1 to 3 (crawlers in robots.txt, the two free reports, the Business Profile). This month: dentist pages, the DHA items and the directory listings. The 5-prompt self-check takes about 2 hours and gives a baseline to revisit in a month. Checklist: below.

Our service. 4queries.com runs the prompt panel for your clinic: 16 prompts for your city in 5 systems, 3 runs, with the log and screenshots (4queries.com/audit) for USD 190. We do not quote other providers' prices.

Twenty items. Next to each one we say what it rests on: a platform's or regulator's own document, an independent study with a disclosed method, or vendor data to act on with a caveat. Claims backed only by a single case or by opinion did not make the list. They sit under "Hypotheses: test them yourself". One item (16) is listed but not scored until verified. Each item gives what to do, how to check it in 5 minutes, and what it is based on. This is a companion to the article "AI Visibility for Dental Clinics in Dubai and Abu Dhabi" (4queries.com, 22.10.2026). DHA clause numbers follow ST-21 v1.1, which DHA re-issued unchanged by circular CIR-2026-00000144 of 17.08.2026. We verified them against the PDF on 22.10.2026.

Do this first: platform and regulator rules

1. robots.txt allows OAI-SearchBot and PerplexityBot. GPTBot is a separate decision, since it is for training only. Google-Extended is not blocked: it controls Gemini training and grounding, and Google's documentation says it does not affect inclusion in Google Search. An independent study accepted to ACM SIGIR 2026 observes fewer citations in Gemini and AI Overviews for sites that block it. That is a correlation. Check: open your-domain/robots.txt and search for these names next to Disallow. Based on: the crawler documentation of OpenAI and Perplexity, plus Google's page and the SIGIR study above.
2. Google Search Console AI performance report and Bing Webmaster AI Performance are connected and checked weekly. Check: in Search Console open Performance and filter by AI features. In Bing Webmaster Tools open AI Performance and confirm the site is verified. If a report is empty, the site is not cited yet or the data is below the display threshold. Check weekly, and after opening the crawlers allow 2 to 4 weeks. Based on: the launch announcements from Google and Microsoft Bing.
3. Google Business Profile is complete and current: services, dentists, hours, photos, booking link. The clinic replies to reviews neutrally and never confirms treatment, diagnosis or the patient's identity (DHA 6.10 covers consent). Check: open the profile in Maps logged out. Count empty fields and unanswered reviews from the last 30 days. Based on: Google's guide to generative AI search, Google Business Profile Help and DHA's ST-21 standard.
4. Every dentist has a page: name, specialty, experience, education, photo, services. A license number is recommended (ST-21 covers social accounts, not web pages). Articles carry an author or "reviewed by" with a date. Check: open three dentist pages and one article. A missing page or templated text is a fail. A missing license number is a recommendation gap. Based on: the platform requirement is in Google's rater guidelines. The AI effect rests only on opinion. It is cheap, so do it.
5. Content is original and expert, not templated or unedited AI text (YMYL, E-E-A-T). Check: paste two sentences from a treatment page into Google in quotes. If other sites return the same text, it is templated. Based on: Google's guide to generative AI search and Google's rater guidelines.
6. Dentist or LocalBusiness schema matches the visible text. Expect no AI lift from it. Check: run the page through Google's Rich Results Test. Compare name, address and phone in the markup with what the page shows. Based on: Google's local business markup documentation. For the AI effect, vendor data from Ahrefs.
7. No Review markup on the clinic's own testimonials to get stars. Check: in the Rich Results Test look for Review or AggregateRating on your own pages. If it is present on self-published testimonials, remove it. Based on: Google's review snippet documentation.
8. The MOHAP advertising license number is shown in official social accounts (DHA 5.2, ST-21 v1.1, verified 22.10.2026). The medical director approves promotional content (DHA 5.3). Check: open the Instagram bio and the latest promotional post. The license number must be visible. Check the website footer too: that is recommended, not a ST-21 rule. Based on: DHA's ST-21 standard.
9. No absolute claims about results or rank: "the best", "unique", "safest", "the only", "guaranteed". This is the DHA 6.11 list for social media advertising. Apply the same wording rule on the website as a precaution. Result claims come with risks (DHA 6.6). Check: search the home page, treatment pages and the last 10 social posts for these words. Ignore "best practice" and similar. Based on: DHA's ST-21 standard.
10. Testimonials only with documented consent (DHA 6.10). Before/after photos show the same person, the same lens, no retouching, with the outcome disclaimer (DHA 9.6; Appendix 2, item 12). Check: pick the latest before/after post, then find the disclaimer and the signed consent on file. Based on: DHA's ST-21 standard.
11. Booking forms and chat have a privacy notice and consent, and counsel has confirmed which law applies. Federal: Federal Law 2/2019 and the PDPL 45/2021. DIFC: its own law, DIFC Law 5/2020. ADGM: not listed on the UAE portal, source to add. A notice and a checkbox are a minimum, not compliance. Check: open the booking form. A visible privacy notice and a consent checkbox must be there, and counsel's note on the applicable law must exist. Based on: the UAE government portal, which shows the acts exist. Their scope is not verified, so confirm it with counsel.
12. No attempts to manipulate AI answers: no paid links without rel="sponsored" or rel="nofollow", no fake reviews, no undisclosed paid mentions or self-promotional rankings. Check: ask your agency where reviews, links and mentions come from, and to show the source of each. Based on: documents, but only for what their texts name. Google's spam policy names manipulation of generative AI responses (since 15.05.2026) and paid links without rel="sponsored". DHA's standard requires consent for testimonials (6.10) and disclosure of paid promotion (6.7). Google's spam policy text does not name fake reviews, bought mentions or self-ranking lists. Avoiding them is our risk call.
13. No budget for llms.txt, "chunking" or rewriting pages "for AI". Check: look at the last SEO invoice or proposal for these words. Based on: Google's guide to generative AI search, plus vendor data from SE Ranking on llms.txt.
14. Bots actually get the page. A request with the bot User-Agent (curl) or the server logs show status 200. No CDN or WAF blocks bots. No nosnippet or noindex on dentist and treatment pages. Check: run curl with the OAI-SearchBot and PerplexityBot User-Agent against a dentist page, or ask the host for the log lines. Then open the page source and search for nosnippet and noindex. Based on: the crawler documentation of OpenAI, Perplexity and Google.
15. Treatment and dentist pages show a last-updated or "reviewed by" date within 12 months. Check: open three treatment pages. Based on: the platform requirement is in Google's rater guidelines. The AI effect rests only on opinion. It is cheap, so do it.

To verify, not scored until verified:

16. Abu Dhabi locations: check with DoH whether Health Media and Advertising training is mandatory (circular 26/2023). We did not find the DoH guideline text. Check: ask the medical director for the training certificate date and for DoH's answer on whether it is mandatory. Based on: still to verify. The DoH standards page confirms only that the circular exists.

Independently measured

No item of this checklist is independently measured at clinic level for an AI effect. The independent studies in the article measure domains and query types, not what a single clinic can change. The candidates sit in "Hypotheses: test them yourself".

Do it with a caveat: vendor data

17. Visibility is measured as share of mentions per prompt, system and run, over 10 to 12 prompts, 3 to 5 systems and 3 days. Not as a single "AI rank". Check: your last visibility report must show runs per prompt and a share, not one screenshot. Based on: vendor data from SparkToro and the ACM audit of health AI Overviews. This is a method, not a factor, and SparkToro is a vendor. Our minimum for a baseline is 10 to 12 prompts over 3 days. SparkToro advises at least 60 to 100 runs per prompt.
18. Mentions in local and industry media and third-party lists, never self-published "top" lists. Check: search "your clinic name" -site:your-domain and count pages with a real description from the last 12 months. Based on: vendor data from Ahrefs, which measured brand mentions in general, not lists specifically.
19. Reviews are collected lawfully: ask after the visit (QR, link), offer no incentives or discounts, keep documented consent for any testimonial reused in advertising (DHA 6.10), reply neutrally. Check: open your review request template and the last five replies. No reward offered, no treatment details confirmed in a reply. Based on: DHA's ST-21 standard for the rules. The effect rests on vendor data from BrightLocal.
20. Listings on the directories AI cites for healthcare. Vendor data: listings make up 52.6 percent of healthcare citations. Run 1 of our panel saw fresha.com, dentaldepartures.com and bookimed among Google AI Mode sources for Dubai. The full list comes after run 3. Check: ask ChatGPT and Perplexity "dental clinic in [your area]" and note which directories appear in the citations. Be listed there. Based on: vendor data from BrightLocal, Yext and Javvo.

Hypotheses: test them yourself

• A dedicated page per treatment (second place in the practitioner panel; a single agency case with zero visibility for services).
• Published "from" prices (practitioner opinion: one practitioner says ChatGPT asks about budget first).
• An FAQ block from real patient questions (we found no measurement in any research pass).
• Area pages (Marina, Business Bay, Al Reem) with real differences between them (classic local SEO, not tied to AI by data).
• Reviews that mention the treatment by name (practitioner opinion).
• Informational pages such as "how much do veneers cost in Dubai" aimed at AI Overviews. An ACM audit found AI Overviews on 87.7 percent of 900 US health queries. A separate SE Ranking study, cited in SE Ranking's health report, puts them on more than 82 percent of health searches in Germany. They are rare on local queries, and there is no UAE figure. The evidence is indirect, at the level of vendor data.
• Dentist videos. The youtube.com domain is cited in medical AI Overviews: vendor data from SE Ranking for Germany, and the ACM audit for the US. The clinic-level effect is unmeasured (vendor data or a single case). Every video is medical advertising under DHA's ST-21: license number, medical director approval, no result claims. Background on AI summaries: Pew Research.
• Review thresholds: at least one new review a month per platform and a rating above 4.0 are US vendor thresholds from BrightLocal and rater8. They are unmeasured for AI (vendor data or a single case).

How to test: pick one hypothesis and change it on the site. Then compare share of mentions in your prompt panel before and 30 days after, with the other items unchanged. One prompt, one day, one system is not a test. A 30-day window is monitoring, not causal proof. Log model versions and country.

5 prompts to test yourself

Run each prompt in a new chat three times on different days. Use Google AI Mode and Perplexity logged out, plus one more system of your choice (ChatGPT, Gemini or Copilot), logged into your own account with memory off. "Named" means your clinic name or domain appears. Five systems are for the full audit; three are enough for a self-check. Count in how many of the three answers your clinic is named, and which sources the answer cites. Wordings are taken from our pre-registered panel, with the city replaced.

1. Which dental clinic in [your city] should I choose for dental implants? Name specific clinics and why.
2. How much do veneers cost in [your city] and which clinics do them well? Name clinics.
3. I have severe tooth pain, which dental clinic near [your area] can see me today? Name specific clinics.
4. Dental clinic with good reviews near [your area], [your city]. Recommend specific clinics.
5. Top 5 dental clinics in [your city] by patient reviews in 2026, with names.

How to read the result:

• You are missing from the list and competitors are there: check items 3 and 4 in "Do this first", then mentions and directories in the vendor-data group (items 18 and 20).
• You are named, but the source is a directory or someone else's site: your own pages give AI nothing to cite. Work on items 4 and 5, dentist pages and original content.
• AI quotes old hours, prices or dentists: your profile and pages are stale. See item 3 and the dates in item 15.
• Answers differ across the three runs: that is normal. Count the share out of three per system. That is your baseline to revisit in a month.
• AI calls you the best or quotes a testimonial: check that nothing on your own channels makes that claim, per DHA 6.10 and 6.11 (items 9 and 10).
• AI names you with wrong hours, prices or dentists, or merges you with another clinic: fix the Business Profile, Bing Places and the directories it cited, then recheck in 2 to 4 weeks.

How can a dental clinic in Dubai get recommended by ChatGPT and Google AI Mode?

Open robots.txt to the AI crawlers that OpenAI, Perplexity and Google document. Complete the Business Profile and keep reviews fresh, as Google advises in its AI search guide and Business Profile help. Publish dentist pages and original content, in line with Google's rater guidelines. Measure your share of mentions across 10 to 12 prompts and 3 runs. The method comes from SparkToro, which sells audience research software and advises 60 to 100 runs. The free reports are official: Google's Search Console AI report and Microsoft's Bing Webmaster Tools AI report. The same Google guide says nothing "AI-specific" is needed.

What must a dental clinic website in the UAE have so that AI assistants recommend it?

Dentist pages with credentials, in line with Google's rater guidelines. Showing the license number on the site is recommended, and clause 5.2 of DHA's standard requires it in official social accounts. Forms need at least a privacy notice and consent, with counsel confirming which law applies (the UAE government portal lists the acts). Valid Dentist schema is for rich results, as Google's documentation describes, with no expected AI lift (vendor data from Ahrefs, which sells SEO tools). Our opinion: prices and FAQ help patients, but they are unproven as AI factors.

Which agencies in the UAE do AI search optimization (GEO, AEO) for dental clinics?

Across three runs in six systems, the answers to panel prompt AE-B2B-02 named 32 agencies. SEO Tech Experts appeared in four systems, Lunasol and AI Dental Agency in three each, and 4queries.com was not named. Any such list is unstable between runs, according to SparkToro, which sells audience research software. In our own snapshot, Google's UAE top 10 for related queries is agency service pages. That is a position, not an assessment.

AI search visibility for dental clinics in the UAE: where to start and what does it cost?

Start with the two free official reports, Google's Search Console AI report and Microsoft's Bing Webmaster Tools AI report, and the 20-item checklist (below). The self-check on 5 prompts takes about 2 hours. Our own service is described in "What next".

What does a clinic need to be named for an implant prompt in Dubai?

That is panel prompt AE-01. Across three runs in six systems, 16 answers named 16 clinics, and no clinic appeared in every system. NOA Dental Clinic appeared in four systems. Dr. Joy Dental Clinic, the Dr. Michael's chain, Drs. Nicolas & Asp and SameDay Dental Implants appeared in three each. The most cited sources were drmichaels.com, drjoydentalclinic.com, bookimed and the DHA services site. The need behind the prompt is an implant page with a named, licensed dentist, in line with Google's rater guidelines, and a price range, which in our opinion is an untested AI factor. Any promise about results must be substantiated and carry risks under clause 6.6 of DHA's standard.

What does a clinic need to be named for a "good reviews near Business Bay" prompt?

Panel prompt AE-04. Across three runs in six systems, Pearl Dental Clinic was named in five systems (13 of 16 answers). Trio Dental Center, Bay Avenue Dental Centre, Dr Gad Dental Clinic and Legacy Care appeared in four each. We expected practices from the Google Maps top 10. The Business Profile as a lever rests on Google's own guide. The link between Maps position and AI mentions rests only on an agency's own measurement by Identity Dental Marketing, where rating was not a predictor. Testimonials on your own channels need documented consent under clause 6.10 of DHA's standard.

What does a clinic need to be named for an emergency prompt?

Panel prompt AE-03. A patient in pain should call a clinic or emergency services, not wait for an AI answer. Urgency prompts lean on opening hours and maps. AI Overviews are rare on local queries: 0 percent on "near me" and 15 percent on "service + city" queries in two vendor samples, from BrightEdge and Whitespark. For a clinic, that means accurate hours and emergency slots in the Business Profile, as Google's help page advises.

Why can a clinic not publish its own "top 5"?

Panel prompts AE-08 and AE-12 ask AI for such lists. Across three runs, AE-08 (Dubai) named Dr. Joy Dental Clinic in all six systems, and Go Dental Clinic and Pearl Dental Clinic in five. AE-12 (Abu Dhabi) named ICDE in five systems (13 of 15 answers) and Dental Experts Center in four. The most cited sources were directories and lists: zavis.ai, bestdentists.ae, bestinhood.com. Lists differ across runs, according to SparkToro, which sells audience research software. An independent ACM Web Science study found that Google's AI Overview sources differ from its organic results. In its social media advertising a clinic may not call itself "the best" or use other absolute claims: that is clause 6.11 of DHA's standard. Google's spam policy, read as written, does not name self-promotional rankings. In a small sample by Lily Ray, the self-ranking brand was left out of the AI recommendation 83 percent of the time.

Do I need llms.txt or special markup for AI?

No. Google says plainly that llms.txt, chunking and AI-specific markup are not needed. A 300,000-domain sample found no relation between llms.txt and citations, according to SE Ranking, which sells SEO software. Use Dentist schema for rich results only, as Google's documentation describes.

Can I publish patient testimonials and before/after photos to show results to AI?

Only within the rules of DHA's standard: documented consent (6.10); before/after of the same person, with the same lens and no retouching (9.6); the outcome disclaimer (Appendix 2, item 12); and substantiated result claims with risks (6.6). In Abu Dhabi, DoH guidelines apply, as the DoH circular confirms, but we did not find their text in the run. These rules apply to advertising and social media. Website content still needs legal review.

1. Google Search Central, Optimizing for generative AI search (guide), 2026-05-15, updated 2026-07-10. https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
2. Google Business Profile Help, Tips to improve your local ranking on Google. https://support.google.com/business/answer/7091
3. BrightLocal, Top sources and directories for local AI search: What 1.9 million citations across 116,670 domains tells us, 2026-08-26 (vendor; US, UK and Australia; AI Overviews, AI Mode, ChatGPT). https://www.brightlocal.com/resources/ai-directory-sources/
4. Identity Dental Marketing, When AI Picks a Dentist: 2026 AI Visibility Study, 2026-06-09 (agency, 118 practices, single run). https://identitydental.com/dental_marketing_visibility_report_2026
5. Google, Search Quality Rater Guidelines, 2025-09-11. https://static.googleusercontent.com/media/guidelines.raterhub.com/en//searchqualityevaluatorguidelines.pdf
6. Dubai Health Authority, Standards for Medical Advertisement Content on Social Media, DHA/HRS/HPSD/ST-21 Version 1.1, issued 2022-08-03, effective 2022-10-03, revision date 2027-08-03; verified against the PDF 22.10.2026, file in evidence/ (DHA ST-21 v1.1, PDF linked from circular CIR-2026-00000144). https://www.dha.gov.ae/uploads/042022/Standards%20for%20Medical%20Advertisement%20Content%20in%20Social%20Media2022433965.pdf
7. OpenAI, Overview of OpenAI Crawlers. https://platform.openai.com/docs/bots
8. Perplexity, Perplexity Crawlers. https://docs.perplexity.ai/guides/bots
9. Google Search Central, Google's common crawlers (Google-Extended), updated 2026-10-15. https://developers.google.com/search/docs/crawling-indexing/google-common-crawlers
10. arXiv 2604.27790, How Generative AI Disrupts Search: An Empirical Study of Google Search, Gemini, and AI Overviews, 2026 (accepted to ACM SIGIR 2026; observational, correlation, grade B). https://arxiv.org/abs/2604.27790
11. SE Ranking, Does LLMs.txt impact your AI visibility and citations?, updated 2026-10-15 (vendor). https://seranking.com/blog/llms-txt/
12. SparkToro, New Research: AIs are highly inconsistent when recommending brands or products, 2026-01-27. https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/
13. ACM Web Science Companion 2026, Information Diversity and Authority Shifts in Google's AI Overviews: An Audit of Health Queries, 2026-05-25. https://doi.org/10.1145/3795513.3807429
14. Google Search Central, AI features and your website, updated 2025-12-10. https://developers.google.com/search/docs/appearance/ai-features
15. BrightEdge, Healthcare and AI Overviews: How Google Sharpened Its Approach Over Three Years, 2025-12-24 (vendor; country not stated). https://www.brightedge.com/resources/weekly-ai-search-insights/healthcare-ai-evolution-google-2023-2025
16. SE Ranking, Health-related AI Overviews turn to YouTube 2-3x more than trusted medical sites, 2026-01-14 (vendor). https://seranking.com/blog/health-ai-overviews-youtube-vs-medical-sites/
17. Whitespark, New Research: The Prevalence of AI Overviews in Local Search, 2025-05-12 (vendor). https://whitespark.ca/blog/case-study-the-prevalence-of-ai-overviews-in-local-search/
18. KFF, Tracking Poll on Health Information and Trust: Use of AI for Health Information and Advice, 2026-03-25. https://www.kff.org/public-opinion/poll-finding/kff-tracking-poll-on-health-information-and-trust-use-of-ai-for-health-information-and-advice/
19. Pew Research Center, From Diagnoses to Treatments: Why Americans Use AI Chatbots for Health, 2026-08-25. https://www.pewresearch.org/science/2026/08/25/from-diagnoses-to-treatments-why-americans-use-ai-chatbots-for-health/
20. Edelman, 2026 Trust Barometer Special Report: Trust and Health, UAE page, 2026-06 (vendor; online survey, 16,006 respondents in 16 countries, 1,000 per country, 28.02-11.03.2026; live page unreachable in check, Wayback copy of 2026-06-21; cited only in section 12). https://www.edelman.com/ae/trust/2026/trust-barometer/special-report-health
21. Khaleej Times, ChatGPT before doctor? Over half of UAE residents use AI for health decisions, 2026-06-10 (press relay of item 20; 16,000+ respondents in 16 countries, UAE sub-sample not stated; cited only in section 12). https://www.khaleejtimes.com/lifestyle/health/ai-use-uae-residents-healthcare-advice
22. Semrush, We analyzed billions of web visits: How AI is reshaping traffic channels, 2026-04-27 (vendor). https://www.semrush.com/blog/traffic-channel-mix-study/
23. Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results, 2025-07-22. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
24. Seer Interactive, AIO Impact on Google CTR: September 2025 Update, 2025-11-04 (vendor). https://www.seerinteractive.com/insights/aio-impact-on-google-ctr-september-2025-update
25. Ahrefs, Update: AI Overviews Reduce Clicks by 58%, 2026-02-04 (vendor). https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/
26. Seer Interactive, AIO Impact on Google CTR: 2026 Update, 2026-04-24 (vendor). https://www.seerinteractive.com/insights/aio-impact-on-google-ctr-2026-update
27. Google Search Central Blog, Introducing Search Generative AI performance reports in Search Console, 2026-06-03, all sites from 2026-08-31. https://developers.google.com/search/blog/2026/06/gen-ai-performance-reports
28. Microsoft Bing Webmaster Blog, Introducing AI Performance in Bing Webmaster Tools Public Preview, 2026-02-10. https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview
29. Yext, AI Doesn't Rank, It Cites. And 86% of Its Sources Are Brand-Managed, 2025-10-09 (vendor). https://www.yext.com/blog/ai-citations-86-percent-of-sources-are-brand-managed
30. BrightLocal, What 200k local AI searches tell us about what businesses get recommended, and why your website matters (Local AI Visibility Study, part 2 of the study in item 3), 2026-09-16 (vendor). https://www.brightlocal.com/research/local-ai-visibility-study/
31. Yext Research, AI Citations, User Locations, and Query Context, 2025-10-09 (vendor). https://www.yext.com/research/articles/ai-citations-user-locations-query-context/
32. Javvo, AI Visibility Report for Local Medical Practices, 2026-08 (vendor). https://javvo.ai/research/local-medical-practices/
33. Elevate DDS, AI SEO for Dentists: Do Dentists Really Need to Rank in ChatGPT in 2026?, case study, 2026-04-28 (agency). https://www.elevatedds.com/ai-seo-for-dentists/
34. rater8, 2026 Patient Choice Report, press release, 2026-06-03 (vendor). https://rater8.com/blog/2026-patient-choice-report/
35. BrightLocal, Local Consumer Review Survey 2026, 2026-02-11 (vendor). https://www.brightlocal.com/research/local-consumer-review-survey/
36. Google Search Central, Review snippet structured data, updated 2026-09-08. https://developers.google.com/search/docs/appearance/structured-data/review-snippet
37. Google Search Central, Spam policies for Google web search, generative clause added 2026-05-15. https://developers.google.com/search/docs/essentials/spam-policies
38. 4queries research run dental-ai-visibility-2026-10, Google UAE SERP snapshot (gl=ae, hl=en, pws=0), 2026-10-22, 4 of 6 queries captured; a block headed "AI Overview" on the standard results page for 4 of 4 (its hidden heading in the page markup reads "AI Mode reply for" the query, so the label is not certain); files 03-serp/google-ae/summary.md and serp.json, published with the article.
39. Whitespark, 2026 Local Search Ranking Factors (expert panel, opinion), 2025-11-06. https://whitespark.ca/local-search-ranking-factors/
40. Google Search Central, Local business (LocalBusiness) structured data, updated 2026-09-08. https://developers.google.com/search/docs/appearance/structured-data/local-business
41. schema.org, Dentist. https://schema.org/Dentist
42. Ahrefs, We Tracked 1,885 Pages Adding Schema. AI Citations Barely Moved, 2026-05-11 (vendor). https://ahrefs.com/blog/schema-ai-citations/
43. UAE Government portal (u.ae), Data protection laws. https://u.ae/en/about-the-uae/digital-uae/data/data-protection-laws
44. Google Search Central Blog, Top ways to ensure your content performs well in Google's AI experiences on Search, 2025-05-21. https://developers.google.com/search/blog/2025/05/succeeding-in-ai-search
45. YouTube, Viktor Chistilin / Medical Marketing, "SEO для стоматологии: 20 крупнейших клиник и их фатальные ошибки в продвижении", 2025-10-06 (Russian, practitioner). https://www.youtube.com/watch?v=LvOdD8k_aak
46. YouTube, Dmitry Varskoy, "Продвижение SEO-статьями в нейросетях, Яндексе и Google 2027: стоматология, клиника, стройка", 2026-10-13 (Russian, practitioner). https://www.youtube.com/watch?v=4ywFX6AVBV4
47. Search Atlas, The Limits of /llms.txt for AI Search: Citation Analysis Across Major LLMs, updated 2026-06-19 (vendor). https://searchatlas.com/blog/limits-of-llms-txt-for-ai-search-citation-study/
48. John Mueller, Bluesky post on llms.txt, 2025-06-17 (opinion). https://bsky.app/profile/johnmu.com/post/3lrshm4gggs2v
49. YouTube, Sinergium, "Продвижение стоматологии в нейросетях: как попасть в выдачу ИИ", 2025-11-28 (Russian, practitioner). https://www.youtube.com/watch?v=bgmhWqK-7tI
50. Ahrefs, An Analysis of AI Overview Brand Visibility Factors (75K brands), 2025-05-26 (vendor). https://ahrefs.com/blog/ai-overview-brand-correlation/
51. Lily Ray, Prediction: the next massive Google update is around the corner, Substack, 2026-10-04 (small sample: 100 "best B2B software" queries in Google AI Overviews). https://lilyraynyc.substack.com/p/prediction-the-next-massive-google
52. Department of Health Abu Dhabi, Standards page with training notice, circular 26/2023. https://www.doh.gov.ae/en/resources/standards
53. Dubai Health Authority, circular CIR-2026-00000144, Standards for Medical Advertisement Content on Social Media, 2026-08-17; re-issues ST-21 v1.1 unchanged, verified by the editors 22.10.2026. URL checked 22.10.2026, copy in evidence/. https://services.dha.gov.ae/sheryan/wps/portal/home/circular-details?circularRefNo=CIR-2026-00000144&fromHome=true&isPublicCircular=true
54. CMS law firm, Expert Guide to advertising of medicines and medical devices, UAE, updated 2024-05-08 (legal commentary, not a vendor). https://cms.law/en/int/expert-guides/cms-expert-guide-to-advertising-of-medicines-and-medical-devices/united-arab-emirates
55. Semrush, AI Overviews Study, 2025-12-15 (vendor). https://www.semrush.com/blog/semrush-ai-overviews-study/
56. The Guardian, "'Dangerous and alarming': Google removes some of its AI summaries after users' health put at risk", 2026-01-11; file 06-foreign-research/opencode/evidence/N01.html. https://www.theguardian.com/technology/2026/jan/11/google-ai-overviews-health-guardian-investigation

World, all quotes verified against the source page, interest disclosed where the employer sells a related service: Danny Sullivan (Google), Miriam Ellis (local search consultant), Mike Hawkes (BrightLocal, vendor), Joy Hawkins (Sterling Sky), John Mueller (Google), Andrew Shotland (Local SEO Guide), Lily Ray (independent consultant), Louise Linehan (Ahrefs, vendor), Rand Fishkin (SparkToro), Aleyda Solis (independent consultant), Gina Neff (Queen Mary University of London). No UAE-based expert, lawyer or regulator is in the pool; this is an open gap.

Russian-speaking practitioners (YouTube, translated by the editor, not part of the expert pool): Dmitry Varskoy, Sinergium, Viktor Chistilin.