Best AI visibility tool for a trades business, ranked on which engine you ask

For a trades business the honest first answer is that the two apps disagree about whether you need a tool at all. Across 36 readings covering six trades in the second audit, Gemini named an AI visibility vendor in 8 of its 18 readings and ChatGPT named one in 1 of 18. A company profile was named in 34 of 36 and a directory or review platform in 33 of 36. So the work that moves the answer is the listing, and the tool worth paying for is the one that tells you what the listing produced, on each app separately.

Six trades, one band

Question, 6 readings each	Named a vendor	Named a profile
What should a plumber or electrician do to get named by AI?	0 of 6	6 of 6
What should a landscaping business do to be recommended by AI?	1 of 6	6 of 6
How does a cleaning company appear in AI local answers?	1 of 6	6 of 6
How does an energy or solar installer get recommended by AI?	1 of 6	5 of 6
How does a removals company get named by AI assistants?	3 of 6	6 of 6
How does a home services business appear in AI answers?	3 of 6	5 of 6

These six are one page rather than six because they answered alike. Every one sits between 0 and 3 of 6 on vendor naming and between 5 and 6 of 6 on the profile. The plumber and electrician reading is the plainest of them: six readings, two apps, three countries, and not one of the fourteen vendors we checked for was named in any of them. A tradesperson who asks an assistant how to get recommended is told to sort the listing out, and that advice is correct.

Why the engine split matters more here than anywhere

A trades business lives on local demand, and the apps behave differently on exactly that kind of question. Across all 120 physical and local trade readings, ChatGPT named a vendor in 4 of 60 and Gemini in 25 of 60. Our own citation record runs the other way: six citations in 1,260 readings, every one of them on ChatGPT and none on Gemini anywhere. Those are two separate measurements about two apps and they never belong in the same sentence as a single figure, so we keep them apart and report both.

The practical consequence is that a vendor demonstrating your visibility on one app has shown you one app. Ask for both, and ask which one each number came from.

Ranked on whether the vendor publishes coverage of the listing

The stated criterion is whether a vendor publishes that it watches the surface that carried 34 of these 36 readings. We read nine vendor pages on 20 October 2026, with two user agents each, and saved every response body. The result does not favour us.

1. Ahrefs publishes a Google Business Profile product, described as managing profiles at scale, under a Local SEO group separate from its AI answer product. It is the only vendor we read that publishes coverage of that surface, so it ranks first here.
2. Semrush sells a separate Local product and mentions local SEO from its AI visibility page, which names 3 engines. That Local page is assembled in the browser, so we read nothing from it and claim nothing about it.
3. AI Knows Us. We publish help getting listed on the sites an AI cites, and help asking customers for reviews, up to 500 or 2,000 customers a month by plan, which is adjacent to the surface rather than coverage of it. Our pricing page does not name a profile, a local pack or a directory.
4. Peec, Profound, Otterly, Scrunch, Depra and Rankscale named no local surface on the pages we read, and name 7 to 9 engines each against our 5. For a trade that is found through a listing, that is the question to put to them directly.

Being fair about the rest of it: six of these vendors publish a daily checking cadence where we publish weekly. Three publish unlimited seats or users in those words, where every plan of ours publishes one login with full access. Peec and Rankscale publish a changelog and we publish none. Profound names a security certification and we name none. Where we are genuinely ahead is narrow and checkable, and it is this: we publish the counts above, trade by trade, including the fact that our own name appeared in none of the 120 physical trade readings. No vendor page we read published a count of that kind.

The order of work for a trade

• The profile comes first, and it is free. It was named in 34 of 36 readings. Service area, hours, emergency cover, trades held and the licence or registration number.
• Then the review platform and the trade directory. One or the other appeared in 33 of 36 readings. A complete entry with the right categories beats a half filled one on three sites.
• Write a page per job type, in the words a customer uses. Structured data was named in 33 of 36 readings, and a job type page that names the fault, the fix and the call out basis gives an answer something to quote.
• Track the job, the town and the urgency. Fifteen to thirty questions is plenty for a single trade business. A credit on our plans is one question on one app once, so the cost of a bank that size is easy to work out in advance.
• Check both apps before you draw a conclusion. On this evidence they answer a trades question in different registers.

Electrical work has a wrinkle most trades do not, a hidden address with a licence number attached, and we covered it in the page on a hidden address, a licence number and the panel upgrade query. It carried no durable inbound link when we counted, as is true of 908 of our 1,942 live pages.

Questions about AI visibility tools for a trades business

Should a one van business buy a tool for this?

Do the listing work first, because it is free and it is what the answers were built from. Run a free reading on your own area to see where you stand, and buy the measurement when you have enough questions to be worth tracking monthly. Our cheapest plan that watches the AI apps is Growth at Rs 11,799 a month for 100 questions on all five, weekly.

Why did the plumber and electrician question name no tool in six readings?

Both apps answered it as a local listing question in all six, across three countries. The category markets itself to buyers who already search for AI visibility software, and a trades business is usually asking a different question.

Which engine should a trade check first?

Check both and keep the answers separate. Gemini named a vendor in 8 of 18 trade readings and ChatGPT in 1 of 18, so a single app gives you one view of what your customers are being told.

Do reviews matter more than the website for a trade?

A review platform or directory appeared in 33 of 36 readings, and across the whole second audit review directories carried 0.7% of all citations. Both are true: in this trade the listing is unusually load bearing compared with the site wide average, which is exactly why a trade should measure its own vertical rather than rely on a category figure.

How often does a trade need to re-check?

Monthly is enough for most, and weekly during a seasonal peak so you are working from current data when demand moves. The reason to re-check at all is that the only way to show a change worked is the same question asked twice, which is why we publish our own re-reads with dates on our proof page.

What should a trade ask a vendor on the first call?

Ask for a dated reading on your own town, on both apps, with the source domains listed, and ask which domain carried the answer. That tells you whether the next hour goes on the profile, a directory entry or a page on your own site.

Run the same question past both apps before you believe either, because on this evidence the two gave a tradesperson different kinds of answer about who to hire.