Over the past 18 months, our audit desk has watched a strange pattern play out across Singapore client accounts. Sites ranking on Google’s page one for their money keywords have been losing 40 to 60 percent of their organic traffic year on year, while their rankings barely moved. The rankings didn’t collapse. The clicks did.
According to SparkToro’s 2026 analysis, only 32 percent of Google searches still send a click to any website; the other 68 percent end at Google itself. Ahrefs recorded that organic click-through rate on queries with AI Overviews dropped from 1.76 percent in June 2024 to 0.61 percent by September 2025, a 65 percent collapse. Meanwhile, Profound’s October 2025 entity-update analysis shows ChatGPT now surfaces just three to four brand mentions per response, down from six to seven before the update, with average brand visibility falling 31 percent across the board. Perplexity users cite Wikipedia eight times more often than any other domain.
Ranking is no longer the same thing as being found. You can hold position one and still lose to an AI Overview quoting a competitor. This is why the SEO, AEO, GEO conversation matters, and why most articles on the topic get it wrong. They treat the three as siblings when they are actually stacked layers with different economics, different signals, and different failure modes.
The stakes for this shift are why more Singapore business owners are quietly rethinking how they select a digital marketing agency in 2026. The evaluation criteria that worked in 2022 (rankings, traffic, backlinks) now miss two of the three layers where discovery actually happens.
Key Takeaways
- Traditional SEO still drives the majority of qualified traffic in Singapore, but it now converts a smaller share of impressions because Google increasingly answers within the results page rather than sending users to yours.
- AEO wins the featured snippet, People Also Ask boxes, voice assistants, and AI Overviews on informational queries. Getting cited here is often more valuable than ranking below.
- GEO wins the answer inside ChatGPT, Gemini, Perplexity, and Copilot. Different citation logic, different measurement, and different content patterns apply.
- The three disciplines share about 70 percent of the same technical foundations. The remaining 30 percent is where most brands are quietly losing visibility they don’t yet track.
- Singapore brands optimising for only one layer today are effectively invisible to a growing majority of buyers who never see a link.
The Three Layers of Modern Search Visibility
SEO: Still the Backbone, but the Job Has Changed
Search Engine Optimisation remains the practice of earning organic visibility on traditional search engines (Google, Bing, DuckDuckGo, and increasingly Baidu for Chinese-language queries in Singapore). Across our client base, SEO still drives 58 to 74 percent of total organic-attributed revenue depending on industry. The mechanics haven’t changed: technical health, on-page relevance, and off-page authority.
What has changed is what a good SEO position now delivers. The First Page Sage 2026 CTR study shows the #1 organic result receives 39.8 percent CTR when no AI Overview is present, but only around 15 percent when one appears above it. Position one is worth roughly a third of what it was in 2023.
How SEO still earns its keep in 2026:
- Transactional keywords (“buy” / “hire” / “book”) still convert through traditional blue links because users want to compare and click.
- Branded searches remain unaffected by AI Overviews; a Singapore brand’s own homepage still receives the click. This is why branded search optimisation has become a defensive priority for brands with rising AI citation rates.
- Commercial-investigation queries (“best CRM for property agents Singapore”) are increasingly answered inside AI Overviews, meaning ranking now competes with citation.
SEO is now a conversion channel more than a discovery channel. This is a genuine shift, and most Singapore brands still budget as if it were 2020. Working with an experienced SEO agency that understands this repositioning matters more than adding another headcount to execute the old playbook faster.
How much money is my brand losing to AI Overviews is the question we now start every audit with. For informational content sites, the answer has been in the 30 to 55 percent range.
AEO: The Answer Layer Where Zero-Click Actually Pays
Answer Engine Optimisation is the discipline of getting your content selected as the direct answer within Google’s featured snippets, People Also Ask boxes, voice assistant responses, and AI Overviews. The user rarely clicks, so the traditional CTR metric misleads teams into thinking AEO doesn’t work. It does. The value simply shows up differently.
Dataslayer’s 2026 study found that brands cited inside AI Overviews receive 35 percent more clicks on the branded searches that follow, compared to brands not cited. In practice, this looks like: user asks Google “best digital marketing agencies in Singapore,” AI Overview names three brands, user then searches “[brand name] reviews” separately, click happens on the branded query. AEO buys you brand-name presence in the moment of decision.
What actually earns an AEO placement:
- A direct 40 to 60-word answer positioned in the first 100 words of the page, before the fluff.
- Question-format H2s that mirror the user’s spoken query, not the keyword-optimised phrasing SEO writers instinctively use.
- FAQ, HowTo, and Article schema markup applied correctly. Our schema markup guide for structured data walks through the Article, Organization, FAQ, and HowTo implementations that most Singapore sites still get wrong. Structured data stops being optional at the AEO layer and becomes the entry ticket.
- Table and list formats. Paragraph snippets still dominate at 70 percent of featured snippet formats, but list and table snippets have higher conversion when the query implies comparison.
How do I get my page into a Google featured snippet is a question we hear weekly. The honest answer: the page must already rank in positions 1 to 5 for the target query. Featured snippets are almost never awarded to results ranked below page one. AEO isn’t a shortcut around SEO; it’s a layer on top of it.
From our AEO audit finding across 60+ Singapore B2B sites (2024–2026), we found that 82 percent had FAQ content but only 11 percent had FAQ schema markup implemented correctly, and only 4 percent had answer paragraphs positioned in the first 100 words.
The gap between “we have FAQ content” and “our FAQ content is optimised for answer engines” is where most of the AEO opportunity in Singapore currently sits. For the specific citation-pattern side of this, our breakdown of how AI overview citations work explains why some sites get quoted while similarly ranked competitors don’t.
GEO: The AI Citation Layer Where the Rules Are Still Being Written
Generative Engine Optimisation is the practice of getting your brand and content cited inside AI-generated responses on ChatGPT, Perplexity, Google Gemini, Microsoft Copilot, and increasingly niche vertical AI tools. This is the newest and most misunderstood layer, and it’s also where the most competitive advantage currently exists because most Singapore brands are not yet optimising for it at all. A specialist GEO agency approach is required because the signals, tooling, and content patterns are meaningfully different from traditional SEO.
GEO does not care about your Google ranking. It cares about whether your content trains well, retrieves well, and looks trustworthy to a language model deciding what to include in its answer. The signals are related to SEO but not identical.
What actually earns GEO citations:
-
Named entity clarity
If your brand, product, and category aren’t mentioned in the same sentence structure repeatedly across your site and third-party sources, AI models struggle to associate you with your category.
-
Third-party mentions
ChatGPT and Perplexity rely heavily on Wikipedia, Reddit, YouTube transcripts, industry publications, and news sites. Your own website is rarely the primary source; it’s the confirming source.
-
Statistics with attribution
AI models preferentially cite numbers that include a visible source. A page claiming “68 percent of searches end without a click” without attribution is less likely to be quoted than the same page attributing it to SparkToro.
-
Recency signals
For fast-moving topics, models check publication date and last-updated timestamps. Older pages get filtered even when accurate.
-
Semantic completeness
Content that answers a topic from multiple angles (definitions, mechanics, examples, comparisons, exceptions) trains models to see the page as authoritative on that topic.
How do I get ChatGPT to recommend my brand is the question every founder in Singapore now asks. The unsatisfying but accurate answer: you don’t optimise ChatGPT directly. You optimise the sources ChatGPT trusts. That means Wikipedia entity presence, review site profiles, curated directory listings, Reddit discussion, YouTube coverage, and industry publications mentioning your brand alongside your category descriptors. Our companion analysis on how to still get mentioned in Google AI Overviews walks through the specific citation triggers on Google’s side of this problem.
Our internal tracking across 12 Singapore B2B clients over the past six months shows the strongest correlations with AI citation weren’t domain authority or backlink count.
They were:
- (1) Number of third-party publications mentioning the brand within the past 12 months
- (2) Presence of first-hand data or original research on the brand’s own site
- (3) Consistent use of the brand’s category-defining phrase across owned and earned media
This connects to our broader work on off-page GEO, which walks through how to build the third-party citation network that AI models actually pull from.
The Overlap Most Guides Miss
Standard articles on this topic list “user-centric design” and “content quality” as shared principles, which is true and useless. The overlap that matters commercially is narrower and more specific.
The 70/30 Rule
From our audit data, roughly 70 percent of the technical foundation and content practices that improve SEO also improve AEO and GEO. The remaining 30 percent is where each layer diverges, and where most brands are quietly underperforming on one or two layers without knowing it.
Shared 70 percent (do these once, benefit three times):
- Site speed and Core Web Vitals
- Clean site architecture and internal linking
- Original research, statistics, and firsthand data
- Author bylines with verifiable credentials
- Schema markup (Article, Organization, FAQ, Product)
- HTTPS, mobile-first responsive design
- Content freshness and update timestamps
- Consistent brand-plus-category phrasing
Diverging 30 percent (this is where budgets should now shift):
| Layer | Where the Extra Work Goes |
| SEO-specific | Long-tail keyword clusters, backlink acquisition, conversion optimisation on landing pages |
| AEO-specific | 40 to 60-word answer paragraphs, question-format H2s, HowTo and FAQ schema, comparison tables |
| GEO-specific | Wikipedia entity work, Reddit and YouTube presence, industry publication mentions, structured statistics with attribution |
A brand budgeting the same way as 2023 is overinvesting in the diverging SEO-specific work and underinvesting in the diverging AEO and GEO work. What percentage of my SEO budget should go to AEO and GEO in 2026 is the correct budgeting question, and the honest answer varies by industry. For informational-heavy sites, we now recommend 40/35/25 SEO/AEO/GEO. For transactional e-commerce, 60/25/15.
A Singapore Case Composite: The Three Layers Working Together
A Singapore B2B software client (SaaS, mid-market, roughly $180K MRR) came to us in Q2 2025 with a familiar complaint: rankings stable, traffic dropping, pipeline softening. Their previous agency had been doing textbook SEO. Nothing wrong with it, and nothing wrong with the rankings, which held positions 2 to 4 for their core commercial-investigation keywords.
The problem showed up when we ran their queries through our AI citation tracker. For 18 of their top 20 keywords, ChatGPT and Perplexity were returning answers that named three competitors and never named them. Google AI Overviews were appearing on 71 percent of their target queries; the client was cited in 8 percent.
We restructured their programme across the three layers:
SEO layer (35 percent of monthly hours): Kept the technical foundation strong. Redirected content spend from broad keyword coverage to depth on the 24 keywords where organic clicks were still converting.
AEO layer (30 percent of monthly hours): Rewrote the top 40 landing pages to include a direct 45-word answer in the first 100 words, restructured H2s into the question format their buyers actually type, and implemented FAQ and HowTo schema across the resource library.
GEO layer (35 percent of monthly hours): Built out Wikipedia entity presence for the founder and product category, sponsored two independent third-party reviews on industry sites, published three pieces of original research with methodology sections designed to be quoted, and structured their homepage messaging around a consistent “category-defining phrase” model.
Nine-month results: Google traffic dropped 12 percent (expected, as informational queries lost to AI Overviews). Featured snippet capture rose from 3 to 27 keywords. ChatGPT citation rate on target queries rose from 8 to 43 percent. Sales-qualified leads from “found us through ChatGPT” or “saw us in AI Overview” doubled. Overall pipeline grew 34 percent.
Measuring only traffic in 2026 hides half the story. Citation share, snippet share, and AI mention frequency are now leading indicators of the pipeline. This shift also explains why AI agent search visibility is now a category of its own, particularly for B2B brands whose buyers now research vendors through AI assistants before ever visiting a website.
The Diverging Measurement Problem
The three layers cannot be measured with the same dashboard, and pretending they can is what most agencies are still doing.
| Layer | What to Measure | Where to Measure It |
| SEO | Organic clicks, keyword rank, conversion rate from organic, revenue per session | Google Search Console, Ahrefs, GA4 |
| AEO | Featured snippet share, People Also Ask presence, AI Overview citation, voice answer selection | GSC “impressions vs clicks” gap analysis, Ahrefs SERP features, manual query tracking |
| GEO | Citation frequency in ChatGPT/Perplexity/Gemini, share of voice against competitor brands, referral traffic from LLM products | Otterly.AI, Profound, Peec.ai, or manual tracking; GA4 referral segmentation for chat.openai.com and perplexity.ai |
Most Singapore SME dashboards still show only column one. Founders then wonder why traffic is flat while pipeline is growing (they’re winning AEO and GEO without measuring it) or why traffic is flat while pipeline is shrinking (they’re losing on AEO and GEO without measuring it either). Both are common. Both are diagnosable.
How do I track whether my brand is being cited in ChatGPT now belongs in every marketing team’s monthly review. The tools are new, imperfect, and improving fast; Otterly.AI, Profound, and Peec.ai currently lead in citation tracking, and Ahrefs Brand Radar has moved into the category.
The Practical Sequence: How to Actually Start
Optimising for three layers sounds expensive. In practice, the shared 70 percent means the incremental cost of extending an existing SEO programme to include AEO and GEO is smaller than founders expect, if the sequencing is right.
Month 1 to 2: Foundation audit Confirm the SEO fundamentals are solid, technical health, Core Web Vitals, schema markup, structured content hierarchy. Broken foundations sink all three layers simultaneously. Our technical SEO audit report tool is a useful starting benchmark.
Month 2 to 3: AEO retrofit on existing top pages Take the 30 to 50 pages already ranking in positions 1 to 10. Add a 45-word answer paragraph near the top of each. Restructure at least one H2 per page into a question format. Add FAQ or HowTo schema where relevant. This is usually the fastest ROI move because the pages already have authority; you’re just repackaging for a different reader (the answer engine).
Month 4 to 6: GEO entity and third-party work Audit third-party mentions. Where is the brand cited outside its own site? Where is it not? Build Wikipedia entity presence where warranted, secure independent review coverage, publish original research designed to be quoted. This is the slowest layer to move because it depends on other publishers, which is exactly why it becomes a competitive moat once established.
Month 7 onwards: Measurement and iteration Add AI citation tracking to the monthly reporting cadence. Measure share of voice against competitors across ChatGPT, Perplexity, and Gemini. Adjust budgets based on which layer is producing pipeline growth in your specific vertical. For the Google side of this, our Google AI Mode SEO strategy walks through the newer AI Mode surfaces (Preferred Sources, Search Profiles) that most agencies are still ignoring.
How long does it take to see results from GEO optimisation is a question we now include in first-call scripts. Featured snippet gains show within 4 to 8 weeks; AI citation gains show within 3 to 6 months; sustained citation share takes 6 to 12 months. Anyone promising ChatGPT visibility in 30 days is selling you something Google’s own ranking teams cannot deliver.
For teams wanting a structured entry point into this work, our GEO optimisation services walk through the same sequence with a monthly cadence and citation-tracking dashboard.
What Competing Advice Currently Gets Wrong
Three claims currently circulating in Singapore digital marketing content deserve pushback.
1. “GEO will replace SEO.”
No serious data supports this. AI Overviews increasingly source their citations from top-ranking pages. If you’re not in the SEO layer, you’re rarely in the AEO or GEO layer either. SEO is the entry ticket, not the outgoing champion.
2. “AEO is just SEO with schema markup.”
Schema is one signal. The bigger AEO shift is content architecture: writing for the reader who will never scroll past your first paragraph and never click your link. The mental model is different from writing for a reader who wants a long-form guide.
3. “GEO requires expensive AI tools.”
The current best-in-class GEO tactics are Wikipedia entity work, third-party mentions, and structured original research. All are labour-intensive rather than tool-intensive. The tools track results; they don’t produce them.
The Compounding Advantage Nobody Is Talking About
Here’s the point most competitors miss. The three layers reinforce each other in ways that create a compounding moat over time.
- Strong SEO earns rankings. Rankings earn featured snippets (AEO). Featured snippets increase branded search volume, which reinforces GEO signals because AI models see rising brand mention frequency.
- Strong AEO earns citations in AI Overviews. Citations drive branded search. Branded search reinforces SEO signals.
- Strong GEO earns citations in ChatGPT. Users then search Google for the brand name. This reinforces both SEO (branded search) and AEO (branded featured snippet capture).
Brands doing all three well pull further ahead of brands doing only one, at an accelerating rate. This is the compounding structure we build into every retainer engagement.
At MediaOne, our SEO services now integrate AEO and GEO work into the same monthly cadence, because separating them into three vendors creates gaps at exactly the overlap points that drive compounding results.
If you want a diagnostic of where your brand currently stands across the three layers, book a complimentary three-layer visibility audit with our team. You’ll receive a citation share benchmark against your top three competitors, a featured snippet gap report, and a prioritised 90-day roadmap. No sales pitch. Just the diagnostic.
Frequently Asked Questions
What is the difference between SEO, AEO, and GEO in plain terms?
SEO earns you a ranking on a search engine results page. AEO earns you the position of being the direct answer, in a Google featured snippet, People Also Ask box, or voice assistant response. GEO earns you a citation or mention inside an AI-generated answer from ChatGPT, Perplexity, Gemini, or Copilot. The three overlap heavily on technical foundations but diverge on content structure and measurement.
Do I need to hire three separate agencies for SEO, AEO, and GEO?
No, and doing so usually hurts results. The 70 percent shared foundation means work done for one layer benefits the other two. Three vendors create coordination gaps at exactly the overlap points that produce compounding gains. One integrated team, or a single agency with capability across the three layers, tends to outperform three specialists working in isolation.
How do I know if my brand is being cited by AI models like ChatGPT?
The manual approach is to run your top 30 buyer questions through ChatGPT, Perplexity, and Gemini and note which brands appear in the answers. The scalable approach uses citation-tracking tools such as Otterly.AI, Profound, Peec.ai, or Ahrefs Brand Radar. How do I track whether ChatGPT recommends my brand is now a standard question in most agency audits, and the tools are improving fast.
Does Wikipedia really matter for AI citations?
Yes, more than most Singapore marketers realise. Perplexity cites Wikipedia roughly eight times more often than any other source. ChatGPT and Gemini also draw heavily on Wikipedia for entity information. Building a well-cited, notability-compliant Wikipedia entry for a brand or founder is one of the highest-leverage GEO moves currently available, though it requires meeting Wikipedia’s notability standards, which is not trivial.
Can small businesses in Singapore compete on GEO, or is it only for large brands?
Small businesses can compete, sometimes more effectively than large brands, because niche category clarity favours GEO citation. A specialist Singapore accounting firm serving F&B businesses is more likely to be cited for “accountant for restaurants Singapore” than a large generalist firm, because the AI model recognises category specificity. The moves that matter (third-party mentions, structured original research, category-defining phrase consistency) don’t require enterprise budgets. They require discipline and time.
