Nick Tung, published inThe Business Times
Nick Tung is a Singapore AI consultant published in The Business Times on 7 August 2026, arguing that Singapore SMEs which rank on Google are invisible to AI answer engines. His audit of 46 SME sites averaged 48 out of 100.
Below: the piece as it ran, the argument in my own words, the data behind it, and the way to check where your own site stands.
Opens on businesstimes.com.sg. Headline, standfirst and byline reproduced here for reference. The article itself stays with the publisher.
The piece
Singapore's SMEs are winning at Google, losing at AI
Artificial intelligence-powered answers tend to skip over the Republic's small businesses
“The writer, an AI consultant, is CEO of Freemansland Holdings and founder of drnicktung.com. AI tools were used for research and editing before submission. The writer remains fully accountable for the commentary's accuracy, originality and final form.”
Why it matters
Singapore's national business broadsheet ran the argument, and the data behind it is my own audit of 46 Singapore SME websites.
The argument, in my own words
Findable on Google. Invisible to the answer.
A summary of my own findings, written fresh for this page. In July 2026 I audited 46 Singapore SME websites that already rank on Google, and scored each on the signals AI engines actually use to decide what to quote.
48/100
Across 46 Singapore SME sites that already rank on Google, average AI Search Readiness came out at 48 out of 100. Being findable by search is not the same as being quotable by an AI engine.
~75%
Three in four had written a real FAQ and never marked it up with FAQPage schema. The content was already done. The machine-readable layer that AI engines actually extract was simply missing.
~85%
Almost none had an llms.txt telling AI crawlers what the site is authoritative about. The handful that passed did so because their platform auto-generated a generic file, not because anyone curated one.
~35
Education and training scored the weakest of the five sectors audited. Tuition and enrichment sites lead with emotive hooks instead of answers, and every site tested failed the answer-first check.
~63
Construction and trades came out on top, which surprised me. Plain WordPress sites running Yoast quietly inherit LocalBusiness and FAQ markup, so they read to an AI engine far better than they look to a human.
The record
Press & features
Headlines, standfirsts and bylines are reproduced for reference only. Copyright in each article remains with its publisher. Every link goes to the publisher's own page.
Questions
Where this ran, and what it argued.
Where was Nick Tung published?
Nick Tung wrote an op-ed for The Business Times, Singapore's national business broadsheet, published in Opinion & Features on 7 August 2026 under the byline Nicholas Tung. The headline was "Singapore's SMEs are winning at Google, losing at AI". The Business Times ran the argument; the data behind it is Nick's own audit, published in full at drnicktung.com/singapore-ai-search-report.
What did the Business Times op-ed argue?
That being findable on Google is no longer the same as being quotable by an AI engine. Across 46 Singapore SME websites that already rank on Google, average AI Search Readiness came out at 48 out of 100. Roughly three in four had written a real FAQ and never marked it up with FAQPage schema, and almost none published an llms.txt. The content is largely already written. The machine-readable layer AI engines actually extract is what is missing.
What does an AI Search Readiness score measure?
It scores a page from 0 to 100 on the signals AI engines use to decide what to quote: whether the page answers the question first rather than after a slogan, whether FAQ content carries FAQPage schema, content depth, entity and structured-data signals, and whether the site publishes an llms.txt. The rubric is deterministic and evidence-graded, so the same page scores the same twice and every deduction names a specific weakness.
How do I check my own site?
Paste any URL into the free scorer at drnicktung.com/tools/ai-readiness. It returns a score out of 100 in about 30 seconds, with the specific gaps named. There is no sign-up, and it runs the same rubric used to score the 46 sites in the report.
The 30-second version
46 sites averaged 48 / 100.
What does yours score?
The same rubric behind the audit, free and open. Paste any URL and see exactly why AI engines do or do not quote you. No sign-up.