Which Singapore industry has most to lose from not being named by assistants, and how to work out whether yours is one
Four conditions that make the exposure real, applied to Singapore sectors, with the arithmetic to run on your own numbers.
Published by AI Knows Us · Updated 22 October 2026
The exposure is not about industry size. It is about four conditions, and any Singapore business can check whether it meets them in ten minutes. High value per customer, a research step before contact, a small number of suppliers a buyer will consider, and no strong referral channel to fall back on. A business meeting all four loses real money from an answer that does not contain its name. A business meeting one or two does not.
This page gives the four conditions, applies them to Singapore sectors on the evidence we have, and sets out the arithmetic to run on your own figures. It does not rank industries by a visibility rate, because no per industry rate exists in any of our datasets and inventing one would be the error this page is warning about.
The four conditions
| Condition | Why it increases the exposure | Check |
|---|---|---|
| High value per customer | One lost enquiry is material rather than noise | Is a single customer worth more than a month of marketing spend |
| A research step before contact | The buyer asks somebody or something before calling you | Do new enquirers arrive already knowing what they want |
| A short consideration set | A buyer who considers three suppliers and you are not one of the three is out, not fourth | How many quotes does a typical buyer get |
| A weak referral channel | No warm route to replace the cold one | What share of last year came from referrals |
The arithmetic, on your own numbers
Take your average customer value over a year. Multiply by the number of new customers you won last year. Divide by the share of them who found you through a research step rather than a referral, to get the value of the research channel. Then ask what proportion of research now starts with a machine answer, and be honest that you do not know this number yet, which is why the enquiry form question matters more than any national statistic.
That calculation produces a range rather than a figure, and a range is enough. A business where the research channel is worth a quarter of revenue is exposed. A business where it is worth 5 per cent is not, however much the category is talked about.
Which Singapore sectors meet the conditions
On the four conditions rather than on a visibility rate, the sectors that meet three or four of them in this market are professional and financial services, business to business software and services, specialist healthcare and clinics, construction and specialist trades at the commercial end, and education and training at the paid professional end. What those share is a high value per customer and a buyer who researches.
Sectors that meet one or two, and are therefore less exposed than the noise suggests, include fast moving consumer goods, low value retail, and anything where the purchase is impulsive or the referral channel carries most of the business. Those businesses have other problems and this is not their biggest one.
What makes the Singapore exposure sharper than elsewhere
Two things, and the second is measured. A city state has no geographic spread, so the set of businesses an assistant can name for a Singapore category is small, which means absence from it is more visible here than in a large country. On 22 October 2026 we put 33 batches of buying questions to ChatGPT and Gemini with Singapore written into the text of each one, because the search interface we use accepts country, language and location parameters and ignores all three. The snapshot taken that evening holds 64 blind readings, 32 per engine.
When the engines named somebody, four firms carried almost all of it: First Page in 17 readings, MediaOne in 12, Impossible Marketing in 11, OOm in 10. A four name set is a short consideration set, and condition three above says that is where absence costs the most. For comparison, the Australian run taken the same day with the same method named a different six firms across its own 64 readings, so this concentration is a per market fact rather than a general one.
The other half of the finding, which is the good news
In 26 of those 64 Singapore readings the engines named no business at all. So the same measurement that shows a short consideration set also shows that the set is empty more often than it is full. A Singapore business that meets all four exposure conditions is at risk from the 38 occupied answers and has an opportunity in the 26 empty ones, and the second number is larger.
That asymmetry is the practical reason to measure rather than to worry. The exposure and the opportunity are measured by the same reading, and they are the same questions seen from two sides.
Questions about industry exposure to AI answers in Singapore
Why will you not rank the industries
Because there is no per industry rate in any of our datasets. The questions spanned sectors rather than sampling each one, so a sector table built from them would be a fabrication with a decimal point.
What is the single best indicator that my business is exposed
Customer value multiplied by a research step. If one customer is worth a lot and they research before contacting you, the exposure is real whatever your industry is called.
How do I find out what share of my buyers research this way
Two questions on the enquiry form. Where did you first hear about us, as free text, and did you check anything before contacting us. One quarter of answers beats any borrowed percentage.
Is a low value high volume business safe
Safer on this specific risk, and it has its own exposure through marketplaces and paid auctions. The conditions above are the test rather than the category name.
Does being cited protect me if the answer names a competitor
No, and this is the distinction worth holding on to. Being cited as a source and being named as the recommendation are two different outcomes that fail for different reasons, and a reading that merges them hides which one you have. Our page on being read and being chosen as two separate measurements works through why no single number reports both.
The exposure we measured on ourselves
We meet all four conditions and we are at the wrong end of our own measurement. Across 64 blind Singapore readings on 22 October 2026, our business was named zero times and our domain was cited zero times, while ten foreign tools and six local agencies were named around us. That is the exposure described on this page, happening to the company describing it. AI Knows Us runs the same fixed question list for any business across ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews, reports named and cited separately, and the first reading is usually the one that makes the arithmetic above feel concrete.