Informational vs transactional AI queries in Singapore: what actually differs

Informational queries trigger a Google AI Overview 21.4% of the time. Transactional queries trigger one 2.1% of the time, a ten-fold gap across 146 million SERPs. That gap is closing fast: transactional queries grew from 1.98% to 13.94% of all AI Overview triggers in a year. For Singapore marketers, the practical split is not informational versus transactional. It is which of your pages an AI assistant is willing to cite.

If you sell in Singapore, you already know that 72% of local consumers have used an AI assistant to help with shopping. What you probably do not know is how differently AI search treats a question from a purchase. GeoVector tracks that split across six AI assistants: ChatGPT, Gemini, Claude, Perplexity, Google AI Overview and Google AI Mode. The pattern in the public data is consistent, and it changes where your content budget should go.

The two query types, defined

An informational query asks for an explanation. "What is generative engine optimization" or "how does GST work for imported goods" are informational. The user wants an answer, not a vendor.

A transactional query asks for a purchase path. "Buy noise cancelling headphones Singapore" or "GEO platform pricing" are transactional. The user has decided to act and is choosing where.

Between them sits commercial investigation: "best CRM for SMEs", "Shopify vs WooCommerce". Most tracking tools count that as its own class, and it behaves differently again. Generative engine optimization, or GEO, is the practice of getting your brand named and cited inside AI answers. Intent decides how hard that is.

The trigger gap is real, and bigger than most people assume

Ahrefs analysed 146 million SERPs in a study published in November 2025. AI Overviews appeared on 21% of all keywords. Broken out by intent, the spread is wide ^[1]:

Informational queries trigger an AI Overview ten times more often than transactional ones. Two related findings sharpen the picture. Question-phrased queries trigger AI Overviews 57.9% of the time, against 15.5% for non-questions. Queries of seven words or more trigger them 46.4% of the time, against 9.5% for single-word queries ^[1].

Read that together and the rule is simple. Long, question-shaped, explanatory searches are where AI answers live. Short, branded, buy-now searches mostly still return a normal results page.

One caveat before you plan around it. This is Google AI Overview data drawn from a global keyword set. No public study breaks trigger rates down for Singapore specifically, and assistant behaviour varies by engine. Treat the ratios as direction, not as a local forecast.

The gap is closing faster than most content plans assume

Planning for a permanent 10x gap would be a mistake. Semrush tracked the intent mix of AI Overview triggers through 2025 and found lower-funnel searches climbing hard ^[2]:

• Commercial queries: up from 8.15% to 18.57% of triggers
• Transactional queries: up from 1.98% to 13.94%
• Navigational queries: up from 0.84% to 10.33%

In January 2025, 91.3% of queries that triggered an AI Overview were informational. By October that share had fallen to 57.1% ^[2]. Google is pushing AI answers down the funnel, and it is doing so quickly.

So the honest framing is this. Informational content is where AI visibility is cheapest to earn today. Transactional coverage is where it is getting contested. If you wait until transactional trigger rates match informational ones, you will be building authority against competitors who started two years earlier.

Why your own product page rarely wins the transactional answer

Here is the part that surprises most marketing teams. Even when a transactional query does produce an AI answer, the citation usually does not point at a vendor.

In the Ahrefs data, 76.3% of keywords that trigger AI Overviews are non-branded, against 23.7% branded. AI Overviews are 1.9 times more common for non-branded keywords ^[1]. The engine is answering the category question, not the brand question, and it pulls from sources that read as independent.

That has a direct consequence for a Singapore brand. If a buyer asks an assistant to recommend a supplier, the answer is assembled from comparison articles, editorial reviews, directories and forum threads. Your pricing page is not the raw material. Earned coverage is.

So transactional AI visibility is bought differently from transactional SEO. You do not optimise a landing page and wait. You get named accurately in the sources the assistant already trusts, then make sure your own pages confirm the facts those sources state. Where the two disagree, the assistant tends to repeat the third party.

What the Singapore numbers say about urgency

Singapore is further into this shift than most markets. Adyen's 2026 index, run by YouGov across 1,041 Singapore consumers in March and April 2026, found that 72% have used AI assistants to help with shopping. Among Gen Z that rises to 85%, and among millennials to 80% ^[3].

That is your transactional traffic, already routing through an assistant before it reaches you. The behaviour is not evenly split by intent either. The same shoppers use AI to research a category and to shortlist a supplier, often inside one conversation.

The regulatory direction reinforces it. In July 2026, IMDA published voluntary Transparency Guidelines for consumer-facing generative AI chatbots, built around a standardised "chatbot info card" that discloses capabilities, limitations and data handling ^[4]. Google, Meta, DBS, OCBC, Singapore Airlines and Synapxe have indicated they will reference the guidelines within 6 to 12 months ^[5]. Singapore is normalising AI assistants as a mainstream consumer channel, not treating them as an experiment.

Check the zero-click assumption before you act on it

Most GEO advice leans on a scary zero-click number. The number is real, but it is often applied to the wrong thing.

SparkToro, using Similarweb clickstream data, found 68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% in 2024 ^[6]. That is a measure of all searches, not of AI Overviews specifically.

Semrush ran a narrower test. It took the same keywords before and after an AI Overview appeared on them. The zero-click rate went from 33.75% to 31.53%, a small decrease ^[2]. AI Overviews show up more often on searches that were already unlikely to be clicked. The AI Overview is not automatically what killed the click.

Why this matters for intent planning: if you assume AI Overviews destroy clicks uniformly, you will underinvest in informational content on the grounds that it no longer earns traffic. The data does not support that. Informational content is where you win a citation, and a citation is what carries your name into the transactional conversation later.

How to split content by intent, practically

1. Write the questions, not the categories. Question-phrased, long-form pages trigger AI answers at 57.9% versus 15.5% ^[1]. Title and structure your pages as the question a buyer actually types.
2. Answer in the first 60 words. Give the direct answer before the context. Assistants extract answers, not introductions.
3. Put the transactional facts where a third party can quote them. Price, coverage, availability, SGD figures, terms. If a comparison site cannot find your number, it will publish a competitor's.
4. Earn the non-branded coverage. Comparison articles, industry roundups and independent reviews are the citation supply for the decision stage. Budget for them like media, because that is what they are.
5. Measure by stage, not by keyword. Map prompts to Discovery, Research and Decision, then track citation share separately for each. A brand can dominate explanatory answers and be absent from every shortlist. GeoVector separates those three stages so the gap is visible rather than averaged away.
6. Re-check quarterly. Transactional trigger share moved from 1.98% to 13.94% in a year ^[2]. An intent strategy set in 2025 is already out of date.

Sources

1. ahrefs.com: "What Triggers AI Overviews? 86 Factors and 146 Million SERPs Analyzed", November 2025. AI Overview trigger rates by intent, question format, query length and branded share.
2. semrush.com: Semrush AI Overviews Study, data through November 2025. Intent mix of AI Overview triggers and before/after zero-click comparison.
3. adyen.com: Adyen Index Singapore 2026, YouGov survey of 1,041 Singapore consumers, March to April 2026. AI shopping assistant usage overall and by generation.
4. imda.gov.sg: IMDA, "New Transparency Guidelines to help consumers use Generative AI chatbots safely and responsibly", July 2026.
5. marketing-interactive.com: Coverage of the IMDA guidelines, 21 July 2026, including the organisations planning to adopt them.
6. searchengineland.com: SparkToro analysis of Similarweb clickstream data, US Google searches, January to April 2026.