Which Australian industry has most to lose from not being named by assistants, and the four conditions to test

The exposure is not about industry size. It is about four conditions, and any Australian 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 Australian 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	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 them 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 new customers you won last year. Work out 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 that number yet, which is why the enquiry form question matters more than any national statistic.

That 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 Australian sectors meet the conditions

On the four conditions rather than on a visibility rate, the sectors meeting three or four of them here are professional and financial services, business to business software and services, specialist healthcare and allied health, commercial construction and specialist trades at the larger end, and paid professional education and training. What those share is a high value per customer and a buyer who researches.

Sectors meeting one or two, and 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 Australian exposure particular

Two things, and the second is measured. Geography means a buyer's consideration set is often shaped by who can actually service them, which narrows it further and makes absence more costly. And on 22 October 2026 we measured what the engines do with Australian buying questions: 33 batches put to ChatGPT and Gemini with Australia 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, six firms carried almost all of it: Online Marketing Gurus in 12 readings, Rocket Agency in 8, StudioHawk in 8, Prosperity Media in 8, Sparro in 3, King Kong in 3. A six name set is a short consideration set, and condition three above says that is where absence costs the most. For comparison, the Singapore 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 government framing, which changes who is exposed

One Australian feature adds a sector specific exposure that other markets show less of. Sixty six distinct Australian government hosts appeared 139 times across 25 of the 64 readings, including accc.gov.au in 13, acma.gov.au in 8, oaic.gov.au in 6 and the health practitioner regulators in single appearances. So in regulated sectors the buyer is handed the rules alongside the names, and a supplier absent from the names is absent from an answer that already has a compliance frame around it. That makes the regulated sectors doubly exposed: the consideration set is short and the answer carries authority the buyer trusts.

The other half of the finding, which is the good news

In 28 of those 64 Australian readings the engines named no business at all. So the same measurement showing a short consideration set also shows that the set is empty more often than it is full. A business meeting all four exposure conditions is at risk from the 36 occupied answers and has an opportunity in the 28 empty ones, and the second number is larger. That asymmetry is the practical reason to measure rather than to worry.

Questions about industry exposure to AI answers in Australia

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 lead platforms 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. Our page on being read and being chosen as two separate measurements works through why no single number reports both.

The conditions applied to the sector the regulators framed hardest

One Australian cluster meets all four conditions and carries an extra one. Health and allied health has a high value per patient, a research step before contact, a short consideration set, and in our reading it also had its regulators in the answer: the health practitioner regulator, the medical board and the therapeutic goods regulator each appeared across the 64 blind readings taken on 22 October 2026, inside a set where 66 distinct government hosts appeared 139 times across 25 of the 64. So a clinic absent from the names is absent from an answer that already carries official authority, which is a harder position than being absent from a search result. The same logic applies to financial services and credit, where the public registers are themselves part of how a buyer judges you. Against that, in 28 of the 64 readings nobody was named at all, so the exposure and the opportunity are the same questions read from two sides. AI Knows Us runs the fixed question list across ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews and reports named and cited separately, and our own entry is zero on both across all 64, which is the exposure on this page happening to the company describing it.