If you sell software to one industry, dentists, contractors, property managers, HVAC companies, you know where your buyers trust information. The trade magazine. The association. The conference. The consultant everyone in the industry uses. Vertical software marketing is built around those places, and for good reason: that is where the buyers have always been.
So I wanted to know whether the AI engines read those places too. In November I asked Google AI Mode four questions, the way an owner or office manager in each industry might ask them, with the size of the business and what they needed. Best practice management software for a five-dentist clinic leaving an old server-based system. Best construction project management software for a 40-person commercial general contractor. Best property management software for a 200-unit residential portfolio, with accounting and a tenant portal. Best field service software for an HVAC company with 25 technicians. Then I opened every source behind every answer.
The trade press was not there
Across the four answers there were 60 linked sources. Not one was an industry association. Not one was a trade publication for dentists, contractors, property managers or HVAC companies. The only publication of any kind was a technology site covering Microsoft’s business software. The closest thing to a specialist industry source was a dental technology comparison site, which is a directory, not the trade press.
I do not know whether those publications have relevant pages. Some probably do, and some may be behind paywalls or logins. What the test shows is narrower and more useful: when a buyer in these industries asked an engine for a shortlist, the sources the industry itself trusts did not shape the answer.
Who did write it
The vendors’ own sites, mostly for the basics. About half the links went to the vendors being recommended, many of them homepages linked from the product names, plus pricing pages and some vendors’ own comparison pages: one product against another, written by one of the two.
Other software companies’ “best software” lists. This group wrote much of the actual judgement. In the construction answer, the description of the category leader, the one buyers would read first, was attributed to a list published by a construction software company based in India, answering a question about US commercial contractors. In the HVAC answer, the description of the leading platform came from a smaller field service software company’s list. A facilities management software company’s blog was cited in two of the four answers, construction and HVAC.
The businesses that serve the industry. In the property management answer, two of the sources were bookkeeping firms that work with property managers, comparing the software their clients use. In the dental answer, a dental marketing agency, the dental technology comparison site and a software development agency wrote comparisons. These are the advisers vertical buyers often trust, publishing in the open.
Directories. A software directory or review site appeared in three of the four answers, and in the dental answer there were four directory pages, side-by-side comparisons of named products.
Industry Reddit and YouTube. Reddit threads from the industry’s own subreddit appeared in three of four answers. One was from 2021 and was still being used to describe how a product handles invoices. YouTube videos appeared in three of four, including one from a vendor’s own channel.
Why this matters more in vertical software
In most software categories, buyers are used to reading roundups written by other software companies, and they discount them. Vertical buyers are often not software people. A dental office manager or a contractor’s operations lead reading an AI answer does not see that the line describing your product came from a company in a different industry, or a different country, that sells something else. They see a confident answer with a table.
And the categories are small. A vertical market might have five or ten serious products. If two or three outside lists describe them, those lists can define the whole category for the engine. In the construction answer, a list that was not written for US buyers set the opening description of the market leader.
There is also a trust gap you can use. Vertical software companies usually have relationships the horizontal ones do not: the industry’s accountants, consultants, IT providers and associations. In these answers, those relationships were only visible where the advisers had published something public and specific, like the bookkeepers’ comparisons.
What to do
Run your buyers’ questions in their words. Use the size and situation details your buyers actually give: chairs, units, technicians, annual volume, the system they are leaving. Read who the engine quotes. Map the sources before you decide anything.
Find the outside lists that describe your category, and correct the facts. They are mostly other software companies’ pages. Some will fix a wrong price or a missing feature if you ask plainly with evidence. For the rest, the answer is your own honest page on the same question, so the engine has a second source.
Help the industry’s advisers publish. The accountants, consultants and IT providers your customers already trust are the most credible outside voices you have, and in these answers the few who had published were cited. Give them accurate, current material for the comparisons they already write for clients. Do not pay for placement, and do not write it for them. Earning these sources is slower than buying a listing and worth more.
Publish your own comparisons, specific to the industry. Several vendors in these answers had pages comparing themselves to a named rival, and those pages were cited. Make yours fair, and specific to the size and situation your buyers describe.
Do not ignore Reddit and YouTube because your buyers are not “online”. They are. The industry subreddits and practical walkthrough videos were in three of four answers. Show up there honestly, as a named person, and keep in mind that old threads keep getting quoted. Why AI search cites Reddit covers how.
Keep the trade press, but do not count on it for AI answers. The magazine feature and the association listing still matter for the buyers who read them. In this test they did not reach the engine. If a trade publication will publish something public and specific about your category, that is worth more than an ad in it.
What I got wrong
I expected the opposite. My working assumption was that vertical software would be the one place where industry authorities still decided the answer, because the engine would recognise them as the specialist sources. I planned to tell vertical vendors to lean harder on their associations and trade press.
None of those sources appeared. The specialist knowledge that did make it in came from advisers who had published in public, and much of the judgement came from software companies writing lists, some from other industries or other countries. The industry’s own institutions were absent.
These were four questions, run once each on one engine, from outside the US, and Google’s answers vary between runs. The Indian construction list may have appeared partly because of where I ran the test from. The absence of trade sources in four answers is a pattern worth checking in your own industry, not a law. Run your own questions before you move budget.
Where this leaves you
Vertical software buyers trust their industry’s own sources. The engine they now ask first, in this test, did not read them. It read other software companies’ lists, directories, the vendors’ own pages, a few advisers who had published in public, and the industry’s Reddit threads. If you sell to one industry, the work is to make those sources accurate about you, and to help the advisers your customers trust put their knowledge somewhere the engine can read it.
If you want to see who writes the AI answers in your industry, that is part of the engagement, and it runs on a documented method.
1 response