What each of five UK sectors should buy first, and the one thing each should not
Professional services, B2B SaaS, retail ecommerce, hospitality and property have different first purchases, and in four of the five the usual recommendation is wrong.
Published by AI Knows Us · Updated 22 October 2026
The five sectors below are sold similar marketing programmes and should buy different first items, because the thing that decides a sale differs in each. For each one here is the first purchase, the thing to refuse, and the measured reason from our own British reading where there is one.
Professional services
Buy first. The page that states what you charge or how you charge, with the stages and the people. Professional buyers self select on price and scope before they call, and a firm that publishes neither gets fewer and worse enquiries. In some regulated professions a version of this page is already required, which makes it the cheapest high value work available.
Refuse. A brand identity project. You need consistency and a readable mark, not a system, until you know which service you want to be known for.
B2B SaaS
Buy first. The comparison page for the competitor you lose to most often. Neutral, specific about which buyer each product suits, and accurate about theirs. Fairness is persuasive and it is the page buyers look for during an evaluation.
Refuse. A content calendar produced in a workshop. Your objections list, written by your own prospects, is the real plan, and it costs nothing to collect.
Retail ecommerce
Buy first. Delivery, returns and stock information stated plainly in text on the product and category pages. These are the top pre purchase questions in this market and the commonest reason a basket is abandoned.
Refuse. Paid social at a budget too small to read. If you cannot afford enough weekly volume to make a decision from the data, you are buying noise rather than a small campaign.
Hospitality
Buy first. Your own booking path, and the information a person needs before they commit: opening times, the actual menu or room types, prices, and how to get there. Then the review habit, asked for at the moment of relief rather than at the moment of payment.
Refuse. Paying a third party for visibility you could own, before your own booking page works. Platforms are a channel, not a substitute for being bookable.
Property
Buy first. The local page nobody has written, about the specific area and the specific transaction type, with the numbers and the process. Property is the sector where generic national content is most abundant and local specificity is scarcest.
Refuse. A volume content programme. It competes in the most crowded part of this market and our reading below is the argument against it.
The measured part, and what it says to all five
On 22 October 2026 we put a 198 question set of British buying questions to ChatGPT and Gemini, with the market fixed inside the prompt because the gl, hl, location and country parameters on the search API are accepted, return 200 and answer from somewhere else. That produced 64 blind readings, 32 per engine. Three findings apply across every sector here.
- 30 of the 64 readings named no supplier at all, and 34 named at least one. In half of these answers there is no incumbent to displace.
- The source list is wide and shallow. 1,022 distinct hosts across the set, 805 of them cited exactly once. There is no gatekeeper list to buy onto, and adding generic volume competes in the busiest part of the pool.
- Official British hosts are read heavily. gov.uk appeared in 16 readings, ico.org.uk in 7, asa.org.uk in 5 and legislation.gov.uk in 4, while trustpilot.com appeared in 1 and checkatrade.com and yell.com in none.
And one finding that is specifically about how to read your own results. A single British agency site, whito.co.uk, was cited in 18 of ChatGPT's 32 readings and named in the answer text of only 4. Being read and being recommended are different outcomes, and whichever sector you are in, a report with one number will describe only one of them.
We are not publishing a per sector naming figure, because 64 readings across a 33 batch question set does not support splitting the result five ways. Anybody offering you a British sector by sector AI visibility benchmark should be asked for the denominator.
The item all five should buy second
A standing reading of the questions your own buyers ask. It is the same price in every sector because the question set is the same length, and it is the only thing that tells you whether the first purchase worked. Take it before the first purchase ships, because the baseline cannot be reconstructed afterwards.
What British companies ask about sector specific marketing
Is our sector harder than the others? The channels differ and the discipline does not. The sector specific part is which question decides the sale, and the answer to that is in your own enquiry calls rather than in a benchmark.
Should we hire an agency that specialises in our sector? It shortens the learning curve and it narrows the field. Ask to see a page they wrote for a client in your sector and judge whether it contains anything that could only be true of that sector.
How much should we spend? Enough to do the first purchase properly and to keep a small measurement running. Spreading a small budget across channels is the failure mode in all five sectors.
What if we are in none of these five? Apply the same test. Which single question decides the sale in your category, and is it answered in text on your own site. That is the first purchase in any sector.
Does the VAT line change the first purchase in any of the five? It changes retail and hospitality most, because a price shown to a consumer has to be the price they pay, so the decision about how prices appear on your pages is a commercial one rather than a formatting one. The standard rate of VAT is 20%, from gov.uk's VAT rates page read on 22 October 2026, and how it applies to your own supplies is a question for your accountant.
When do we review? At ninety days against the one number you chose before starting. Without that number, the review is two opinions.
The reading that tells you which sector advice to ignore
Most sector advice, including ours, is a generalisation that your own category may contradict. The way to find out is to ask the questions your buyers ask, with the United Kingdom in the question text, and record who is named and which sources are cited, monthly, with the two kept apart. AI Knows Us does that across ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews, and for a British buyer the plan that reads all five engines is 299 dollars a month, the price for the United States and the United Kingdom, read from aiknowsus.com on 22 October 2026.
For the full host by host version of the directory finding above, see our page on do AI assistants read British directories.