Local Packs Appear in 93% of Local Searches, AI Overviews in 15%: What Singapore SMEs Should Do Next

Summary: AI Overviews appeared in 15% of local queries, while Google’s Local Pack appeared in 93%, according to Whitespark research reported by Entrepreneur on 23 October 2026. Those figures measure search-feature appearances, not leads, but they argue against abandoning local SEO. Separately, The Straits Times reported that AI adoption among Singapore enterprises increased from 14.7% in 2024 to 23.5% in 2025. Singapore SMEs should pursue two complementary priorities: maintain strong local-search foundations and make their businesses easier for AI assistants to understand and assess.

The latest local-search story challenges a fashionable marketing shortcut: that businesses should stop worrying about Google rankings because AI assistants will handle discovery instead. On 23 October 2026, Entrepreneur reported Whitespark research showing AI Overviews in just 15% of local queries, against 93% for Local Packs—the map-based business listings displayed within Google Search.

That is a strong reason to resist an AI-only strategy. A dental clinic in Bishan or an aircon servicing company covering Tampines should not neglect its existing local-search presence to chase an occasional chatbot mention. However, these figures concern Google search features. They do not measure how often ChatGPT, Gemini, Claude or Perplexity recommend businesses, and they should not be presented as Singapore-specific results.

The local backdrop matters too. On 5 October 2026, The Straits Times reported that AI adoption among Singapore enterprises rose from 14.7% in 2024 to 23.5% in 2025. That supports taking AI seriously—but does not establish that customers have abandoned Maps. Our recommendation is to protect current discovery channels while preparing for additional ones.

Image: Entrepreneur, accompanying the 23 October 2026 article on AI and local search.

1. The 93% Finding Is a Visibility Signal, Not a Revenue Forecast

The most useful interpretation is also the least sensational: Local Packs remain widely present in the local queries examined. Presence is not the same as attention, clicks, appointments or revenue. A Local Pack appearing on a results page does not prove that a particular business appears inside it—or receives an enquiry.

Likewise, 93% and 15% are not competing market shares. A results page can contain more than one feature. Do not add the percentages together, or assume the remaining 85% of searches have no exposure to AI elsewhere. The available findings do not answer those questions.

The supplied reporting also does not establish a Singapore-only sample, a neighbourhood breakdown or results for individual industries. An accountant in Tanjong Pagar should therefore treat this research as a strategic signal, not a prediction of their own enquiry mix. Their actual search results and lead records remain necessary checks.

Consider two illustrative searches: “emergency aircon servicing Tampines” and “what causes an aircon to leak?” The first asks for a provider; the second asks for an explanation. They represent different customer needs. An SME should examine both rather than assume that a finding about local queries describes every informational search related to its services.

For owners, the decision is straightforward: do not approve a wholesale budget shift based on feature-appearance percentages alone. Ask which channels currently produce qualified enquiries, which enquiries become customers, and whether your business information is accurate wherever those customers encounter it.

2. Fix Local SEO Before Buying an AI-Visibility Promise

Local SEO means improving your business’s visibility in location-based search, including Google Maps and local results. Given the reported 93% Local Pack appearance rate, our recommended starting point is a practical audit—not a promise that any checklist guarantees a ranking.

For a tuition centre in Ang Mo Kio, check whether the business name, category, address, telephone number and opening hours accurately describe the centre. Confirm that the linked website explains subjects, levels, lesson arrangements and how parents can enquire. These are basic customer-information checks, not speculative AI tactics.

For renovation contractors serving Woodlands and Jurong, distinguish the office location from the areas served. Do not imply that a staffed branch exists in every neighbourhood. A useful service-area page should explain genuine coverage, project suitability and relevant completed work, rather than simply replacing one district name with another.

• Check the business profile: Verify contact details, categories, hours and the destination of appointment or website links.
• Check service pages: Explain what is included, what is excluded and which locations the business genuinely serves.
• Check customer evidence: Request honest reviews without scripting claims, and respond professionally to feedback.
• Check measurement: Record calls, forms and bookings where possible, while avoiding double-counting the same enquiry.

A Queenstown dental clinic could use that audit to identify conflicting opening hours before commissioning another marketing campaign. A Bishan tuition centre could discover that parents cannot find its class levels. These are illustrative scenarios, not reported case studies or claims of measured uplift.

This also creates a sensible bridge to GEO—generative engine optimisation, or work intended to make a business easier for AI systems to understand and reference. Our guide to business identity consistency for AI search explains the principle: descriptions of the same business should not contradict one another. Consistency is a foundation, not a recommendation guarantee.

Image: The Straits Times, accompanying its report on Singapore’s digital economy and enterprise AI adoption.

3. Build AI-Ready Answers Without Abandoning Local Intent

The Singapore adoption figures require careful wording. The supporting Straits Times data identifies all enterprises: 14.7% in 2024 and 23.5% in 2025. That is an increase of 8.8 percentage points. It is not an SME-only statistic, and it does not measure customers asking chatbots for recommendations.

Nevertheless, business owners can prepare useful information without betting on a particular adoption forecast. Our recommendation is to publish answers that help a prospective customer decide whether to contact you. The same material can serve a human reader and give AI systems clearer source material to assess.

A Tanjong Pagar law firm might explain the matters it handles, what an initial consultation covers and which documents a client should prepare. An accounting practice serving Jurong SMEs might distinguish bookkeeping, tax filing and corporate secretarial services. Neither needs to claim expertise across every possible service to produce a useful page.

An aircon servicing company could explain whether a quoted package includes chemical cleaning, how additional work is authorised and what affects the final price. A tuition centre could describe class size, entry requirements and cancellation terms. Publish only details the business can stand behind; do not invent prices, qualifications or customer outcomes for search visibility.

For a repeatable format, see our guide to direct-answer content. Start with the question, answer it plainly, then add qualifications and supporting evidence. Avoid hiding the answer underneath several paragraphs of promotional copy.

Vercel’s article on adapting SEO for AI search cites research suggesting AI Overviews may reduce clicks by as much as 34.5% compared with similar searches without an overview. That is a reason to monitor discovery beyond website visits. It is not evidence that your clinic or law firm will lose 34.5% of its traffic.

Separate the outcomes you monitor: appearing in an answer, being cited with a link, receiving a visit and winning a customer are different events. A screenshot of a chatbot recommendation is useful evidence of that response at that moment—not proof of stable visibility or commercial impact.

Image: Vercel, author portrait supplied with its article on adapting SEO for LLMs and AI search.

4. A Practical 30-Day Plan for Singapore SME Owners

The following is a proposed operating plan, not a research-backed promise of results within 30 days. Its purpose is to establish accurate information and a baseline before increasing expenditure.

Week 1: Audit local discovery. Review your business profile and website. Check searches reflecting your actual services and locations, such as “dental clinic Bishan” or “accountant Tanjong Pagar”. Record the date and context: search results can differ, so one observation should not become a sweeping performance claim.

Week 2: Correct information gaps. Resolve conflicting contact details, clarify service areas and update unclear service descriptions. Prioritise issues that could prevent a customer from understanding the offer or contacting the business. Keep a change log so later comparisons have context.

Week 3: Publish decision-useful answers. Choose questions from real enquiries. Add clear answers about suitability, process, fees where appropriate and limitations. Link those answers to the relevant service and contact pages rather than creating disconnected articles solely to attract a chatbot citation.

Week 4: Establish an AI-search baseline. Test a documented set of relevant questions across ChatGPT, Gemini, Claude and Perplexity. Record the exact prompt, date, available search settings, named businesses and cited sources. Treat repeated checks as observations, not a comprehensive measure of every customer’s experience.

You can use our GEO Score as a starting point for reviewing AI visibility. Pair any assessment with business outcomes: qualified enquiries, appointments, conversion rates and acquisition costs. There is no defensible universal budget split in the supplied research; allocation should follow your own baseline and commercial priorities.

The verdict: the reported 93% versus 15% appearance rates support keeping local SEO central while testing GEO deliberately. Neither “AI replaces local search” nor “AI can be ignored” follows from the evidence.

Performance Marketing Lab helps Singapore SMEs assess local and AI-search visibility—start with the gaps you can verify, then invest against measurable business goals.

Sources

1. Entrepreneur: AI Hasn’t Cracked Local Search Yet: Why Local SEO Still Wins, 23 October 2026. Source of the reported Whitespark Local Pack and AI Overview findings.
2. The Straits Times: AI adoption rises as Singapore’s digital economy grows to $144b, 5 October 2026. Source of the all-enterprise AI adoption figures.
3. Vercel: How we’re adapting SEO for LLMs and AI search, publication date not supplied. Secondary reference for research suggesting a potential 34.5% click reduction.