Who Does AI Search Recommend in Singapore?
Only 37.9% of AI Overview citations come from Google's top 10 and 61.7% never name the brand. Inside the audit that shows who AI search recommends in Singapore.

Just ask ChatGPT, Gemini or Google AI Search. "What's the best 'anything' in Singapore?" Artificial intelligence convincingly lists several brands. But where did this recommendation come from? It's a different place than most people imagine. Only 37.9% of sources cited by Google's AI answers are in the top 10 of traditional searches. Additionally, over 60% of AI quotes never include the name of the brand being described. This guide only talks about three things. Where does AI Search really recommend in Singapore? Why does the list keep changing? And what can companies realistically do?
There is no better place than Singapore to ask this question. Even though Gemini is used more per person than anywhere else in the world, regular Google searches still make up about 92% of all searches. The two worlds go hand in hand, and both are necessary.
Why Singapore is The Place to Watch AI Search?
It can be said that Singaporeans are accepting AI answers faster than anywhere else. According to Google, Singapore has the "world's highest Gemini usage rate per capita", with users typing an average of nearly 10 prompts per day [1]. When research company Milieu asked 1,000 local residents, 80% use AI tools in their daily lives, with the most common use being simple research [2]. According to a national survey conducted by the Singapore Policy Institute, 64% of adults have used a chatbot [4].
But that doesn't mean traditional search is gone. Google still owns 92.46% [3] of Singapore's search share. Even if you ask Gemini ten questions a day, most of your searches will still come from Google. In other words, you need to optimize both at the same time.
Where AI Answers Took Over? "best" and "compare" Questions
The fastest growing AI responses occurred precisely when users were making purchase decisions. Semrush analyzed over 600,000 keywords and found that over a six-month period, AI overviews increased by an average of 71% on search results pages that include purchase intent. This is a steep 231% increase in the financial sector [5].
Buyers here have been using it for a long time. According to Adyen's Retail Report, 72% of Singaporean shoppers have used an AI assistant to shop in the past year, with this number rising to 85% among Gen Z [6]. In short, if you search for "Singapore's best ___" these days, the machine will give you the answer before you can even click.
Where AI Gets Its Picks? Mostly Not The Google Top 10
This is the most surprising discovery. Ahrefs analyzed 4 million links in its AI overview and found that only 37.9% were mentioned in the top 10 organic searches – compared to about 76% a year ago. About a third of the quotes were from pages 11-100. and another third were from pages outside the 100th place [7]. The same goes for chatbots, Semrush's research found that many of the pages cited by ChatGPT are ranked 21 or lower [8].
Local business media noticed the same blind spot. Columnist Nicholas Tung wrote in an The Business Times article [9]: "ranking Google and mentioning an AI answer machine are two different things." If you only share Google's rankings, you might completely miss what's happening on the AI side.
A Small Club of Websites Gets Cited Most
In the end, AI's answers always come from the same sources. An extensive MIT study of 24,000 search queries found that AI search “represents significantly fewer long-tail data sources” than traditional search [10]. Gradient Group's analysis of 1.2 million ChatGPT responses sums it up in one word: For one topic, "The top 10 domains account for 46% of all referrals and the top 30 domains account for 67%" [11].
However, there is one pitfall. Profound studied about 700,000 ChatGPT conversations and found that "sources come in groups. ChatGPT does not pick a single winner" [12]. In other words, most questions about Singapore are answered by a small group of review sites, comparison sites and major media outlets. If your company's name is not there, on-site Search Engine Optimization alone will not survive.
Being Named is Not The Same as Being Cited
Brands can be mentioned without a link, or they can be linked but not mentioned by name. Semrush, in collaboration with consultant Kevin Indig, found that 61.7% of referrals are “ghost referrals” – ones that refer to a fact but do not reveal the brand name behind them [13]. The pattern reverses engine by engine. Gemini names brands often but rarely links to them; ChatGPT links often but names brands less [13].
The further you go behind the tail, the higher the obstacle becomes. In 37,000 checks, 48-52% of special and local characters "never appeared in responses" [14]. RankSurf also confirms the same difference in B2B responses [15]. Measurement lesson: Calculate "named" and "referenced" separately. Otherwise, you're overestimating your appearance.
The AI Tools Rarely Agree with Each Other
There is no uniform right answer for AI ranking. Out of 250 class tests, the three big engines topped the same brand in only 41.6% of cases. BrightEdge research shows the same thing. Everyone draws similar conclusions about brands, but opinions quickly diverge when it comes to the source of the information. Furthermore, even within Google there are differences. Brands appear in 90% of responses in the AI space, but only 43% of AI overviews.
Practical advice: try each tool separately. Just because you can win on ChatGPT is no guarantee that you can win on Gemini or Perplexity. This industry changes so quickly that retesting every few months is a minimum.
What We Actually Know About Singapore so Far?
To be honest: no matter where you look, the current local data for Singapore only comes from vendors selling similar services. According to one report, the three benchmark sites had a total of 17,411 AI referrals, while the big three banks only had a total of 5,349 AI referrals [19]. However, this number cannot be independently verified. Another study in Singapore rated 48/100 AI readiness of 46 small and medium-sized websites [20]. The average of the index calculated by VantrisAI for 750 B2B sites was also 49.6 points [21].
There is a clear trend here. AI responses often involve various intermediaries rather than the brand itself. That's why steady and continuous monitoring has value. For example, 24owlsGroup. We are a digital marketing and branding agency in Singapore that handles SEO, Google Ads, social media and event management for small and medium businesses in Singapore and Malaysia. The company incorporates this effort directly into its SEO activities. We come up with a fixed list of questions to ask about purchase intent in Singapore and record and rate each AI tool on a monthly basis. This way you can clearly see the real changes every month and take the guesswork out of it.
Read also: How Much a Lead Cost in Singapore?
What These Audits Cannot Promise?
AI responses are really unstable and tools that try to hide them are just hype. SparkToro ran the same prompt 2,961 times and found that the probability of the same brand placement appearing twice was less than 1%. Co-founder Rand Fishkin declares that any tool that gives an "investment in AI" is crap [22]. According to a review of 45 studies, there is still no "stable, long-lasting, cross-platform" effective technology [23]. Furthermore, even when the same prompt is repeated within a day, the overlap in referral sources is only 32–43% [29].
So don't just take the AI's answer with a grain of salt. At most it's like winning the lottery once. If you want to measure it properly, repeat the same prompt over and over again to see how often you achieve - not just a random ranking.
What Genuinely Helps? Good SEO Plus Being Quoted Elsewhere
The methods that actually work are actually quite simple. A large-scale study showed that if the page itself is of good quality and easily searchable, just adding influence, referrals and additional statistical information can improve the appearance of the page by up to 40% in artificial intelligence [25]. Practitioners in Singapore share the same view. Trades deemed to be "consistent, clear and reliable with multiple data sources" are recommended [26].
Google itself is pretty clear. "If you want to publish content on AI Overview, just do regular SEO. You don't need GEO, LLMO or anything else," Google's Gary Ilyes said at the 2025 event [27]. GEO's experts have also provided concrete figures. According to uSERP's Jeremy Moser, "80% of success is doing basic SEO well" [24].
This is what it boils down to. First, establish the basics of SEO. Go ahead to get featured on the review and comparison sites that AI Tool actually refers to. Then every month check how you look with each tool. It is this order that agencies like 24owlsGroup follow when incorporating AI visibility into SEO for small and medium businesses in Singapore.
Frequently asked questions
Is checking my Google ranking enough to know if AI recommends me?
No. Only 37.9% of AI Overview citations come from the top 10 results, and ChatGPT often cites pages ranking 21 or lower. Track whether AI tools name and link you directly, using classic rankings only as a reference point.
Which AI tool should a Singapore business focus on first?
Track all of them, and assume nothing carries over. Gemini names brands far more often than it links, and three tools agreed on one top pick just 41.6% of the time. ChatGPT tends to drive the most referrals, so weight it - but score each separately.
How many times should I test a question before trusting the result?
Many. The same brand list reappeared in fewer than 1 in 100 repeats, and same-day source overlap was only 32 to 43%. Run each question repeatedly and report how often you appear, never a single "rank".
Is GEO just SEO with a new name?
Mostly, on the evidence. Google says to "use normal SEO practices" for AI Overviews, and one expert puts the overlap near 80%. The genuinely new parts are consistent brand information, third-party mentions, and regular per-tool measurement.
Do Singapore SMEs get recommended by AI today?
Often not yet. Business Times reports SMEs "winning at Google, losing at AI", a 46-site audit scored them 48 out of 100, and comparison sites out-cite the big banks by more than 3 to 1 in vendor data. The fix starts with your own tracked question list.
Will AI mentions actually bring traffic or sales?
Not much yet. Pew found users clicked a result on only 8% of AI-summary visits, and just 1% clicked a citation. Amsive found AI referral conversions similar to organic while AI stayed under 1% of sessions. Treat AI visibility as a brand signal for now, not a direct sales channel.
References
- [1] Grow with Google APAC (Google), "The Gemini Report | Southeast Asia 2026 | Singapore - Grow with Google APAC", published 2026-06-30, retrieved 2026-08-17, https://grow.google/intl/en_sg/gemini-report/
- [2] Milieu Insight, "AI in daily life—use, trust, and limits", published 2026-06-29, retrieved 2026-08-17, https://mili.eu/ai-in-everyday-life-how-we-use-it-trust-it-and-where-we-draw-the-line/
- [3] Statcounter Global Stats, "Search Engine Market Share Singapore | Statcounter Global Stats", published 2026-07-31, retrieved 2026-08-17, https://gs.statcounter.com/search-engine-market-share/all/singapore/2026
- [4] Institute of Policy Studies, NUS, "Fraternity and the Social Fabric in the Digital Age", published 2026-01-20, retrieved 2026-08-17, https://lkyspp.nus.edu.sg/docs/default-source/ips/sp2026_findings_survey-on-measuring-social-capital-in-the-digital-age.pdf?sfvrsn=8395040a_1
- [5] Semrush, "AI Overviews are expanding across commercial intent search [Study]", published 2026-07-02, retrieved 2026-08-17, https://www.semrush.com/blog/ai-overviews-commercial-search-study/
- [6] MARKETECH APAC, "Singaporean shoppers turn to AI as rising costs drive generational divide in spending: report", published 2026-07-30, retrieved 2026-08-17, https://marketech-apac.com/singaporean-shoppers-turn-to-ai-as-rising-costs-drive-generational-divide-in-spending/
- [7] Ahrefs, "Update: 38% of AI Overview Citations Pull From The Top 10", published 2026-03-02, retrieved 2026-08-17, https://ahrefs.com/blog/ai-overview-citations-top-10/
- [8] Semrush, "We Studied the Impact of AI Search on SEO Traffic | Our Findings", published 2025-06-09, retrieved 2026-08-17, https://semrush.com/blog/ai-search-seo-traffic-study
- [9] The Business Times (Singapore), "Singapore's SMEs are winning at Google, losing at AI", published 2026-08-07, retrieved 2026-08-17, https://www.businesstimes.com.sg/opinion-features/singapores-smes-are-winning-google-losing-ai
- [10] arXiv (Sinan Aral, Haiwen Li, Rui Zuo - MIT), "The Rise of AI Search: Implications for Information Markets and Human Judgement at Scale", published 2026-02-13, retrieved 2026-08-17, https://arxiv.org/abs/2602.13415
- [11] The Gradient Group, "The Science Of How AI Picks Its Sources", published 2026-03-24, retrieved 2026-08-17, https://gradientgroup.com/the-science-of-how-ai-picks-its-sources/
- [12] Profound, "How ChatGPT sources the web", published n.d. (data window October to December 2025), retrieved 2026-08-17, https://www.tryprofound.com/blog/chatgpt-citation-sources
- [13] Semrush, "Why 62% of AI citations don't lead to brand mentions [Study]", published 2026-06-09, retrieved 2026-08-17, https://www.semrush.com/blog/the-ghost-citations-study/
- [14] arXiv (Will Jack, Noah Lehman, Keller Maloney, Sarah Xu - Unusual.ai), "Prominence-Stratified Failure Modes in Retrieval-Augmented Commercial Recommendation: A 37,000-Run Audit", published 2026-05-22, retrieved 2026-08-17, https://arxiv.org/abs/2605.27439
- [15] RankSurf, "What ChatGPT & Perplexity Actually Cite (6,500 Answers)", published 2026-07-15, retrieved 2026-08-17, https://ranksurf.com/blog/what-chatgpt-and-perplexity-cite
- [16] arXiv (Dmitrij Żatuchin), "Who Owns the AI Recommendation? A Multi-Industry Empirical Map of Brand Category Ownership Across Large Language Models", published 2026-06-22, retrieved 2026-08-17, https://arxiv.org/abs/2606.23057
- [17] BrightEdge, "Why AI Engines Cite Different Sources but Recommend the Same Brands", published n.d., retrieved 2026-08-17, https://www.brightedge.com/resources/weekly-ai-search-insights/ai-search-same-brands-different-sources
- [18] Search Engine Land, "Brands dominate Google AI Mode, struggle in AI Overviews: Study", published 2025-07-30, retrieved 2026-08-17, https://searchengineland.com/google-ai-mode-overviews-brands-study-459754
- [19] Hashmeta AI, "2026 Singapore GEO Industry Report — Generative Engine Optimisation & AI Visibility in Asia's AI Command Centre", published 2026-07-01, retrieved 2026-08-17, https://files.hashmeta.ai/geo-report/2026_Singapore_GEO_Industry_Report_EN.pdf
- [20] Nick Tung, "Singapore AI Search Readiness Report 2026. SME Benchmark", published 2026-07-04, retrieved 2026-08-17, https://drnicktung.com/singapore-ai-search-report
- [21] VantrisAI, "Singapore AI Search Visibility Index 2026 | VantrisAI", published 2026-04-30, retrieved 2026-08-17, https://vantrisai.com/singapore-ai-search-visibility-index-2026/
- [22] SparkToro, "NEW Research: AIs are highly inconsistent when recommending brands or products; marketers should take care when tracking AI visibility", published 2026-01-27, retrieved 2026-08-17, https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/
- [23] arXiv (Olivier Martinez), "Optimizing Visibility in Generative Engines: A Critical Survey of Generative Engine Optimization (2023–2026)", published 2026-07-15, retrieved 2026-08-17, https://arxiv.org/html/2607.14035v1
- [24] Digiday, "GEO hype busted: How it differs (and how it doesn't) from SEO", published 2026-03-10, retrieved 2026-08-17, https://digiday.com/media/geo-hype-busted-experts-call-it-more-seo-than-new-discipline/
- [25] arXiv / ACM SIGKDD, "GEO: Generative Engine Optimization", published 2024-08-24, retrieved 2026-08-17, https://arxiv.org/abs/2311.09735
- [26] Marketing-Interactive, "Generative engine optimisation: The next battleground for brands in Asia", published 2025-11-26, retrieved 2026-08-17, https://www.marketing-interactive.com/generative-engine-optimisation-geo-the-next-battleground-for-brands-in-asia
- [27] Search Engine Roundtable, "Google: Normal SEO Works To Get Into AI Overviews", published 2025-07-24, retrieved 2026-08-17, https://www.seroundtable.com/google-ai-overviews-normal-seo-39817.html
- [28] Pew Research Center, "Google users are less likely to click on links when an AI summary appears in the results", published 2025-07-22, retrieved 2026-08-17, 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/
- [29] arXiv (Schulte, Bleeker, Kaufmann), "Don't Measure Once: Measuring Visibility in AI Search (GEO)", published 2026-04-08, retrieved 2026-08-17, https://doi.org/10.48550/arxiv.2604.07585
- [30] Search Engine Land, "LLM traffic converts about the same as organic search: Research", published 2025-09-03, retrieved 2026-08-17, https://searchengineland.com/llm-organic-search-traffic-convert-same-research-461567
