The single most useful thing to know about AI search in your industry is which sources the models read to build an answer. It is different everywhere, and the general advice, claim your Google Business Profile and get some reviews, is the part that is the same and therefore the part that differentiates nobody.
We track 520 citation sources across 27 verticals. Here is what the pattern looks like by industry.
Healthcare and dental
What the models lean on: licensing bodies, insurer provider directories, and established patient-matching services.
Why: health questions get stricter source selection than any other category. The models prefer sources they treat as verified over a practice's own claims, which narrows the citation surface and raises the cost of being absent from it.
Practical consequence: an incomplete insurer directory entry costs a practice more than a mediocre website does. We track 32 sources in healthcare and 29 in dental, and the high-weight ones are almost all in that verified category.
Legal
What the models lean on: a handful of long-established legal directories, disproportionately.
Why: legal is the most directory-dominated category we measure, and the models treat a firm's own claims about its expertise with visible caution.
Practical consequence: depth in one practice area beats presence in ten. A model answering a narrow question wants a source that answered that narrow question, and a firm with a thin page per practice area gets cited in none of them.
Real estate
What the models lean on: portals, review aggregates and local press. Very rarely agency websites.
Why: property is one of the few categories where the answer is assembled almost entirely from third parties.
Practical consequence: an agency with a beautiful website and an incomplete portal profile is invisible, and no amount of site work fixes it.
Home services
What the models lean on: review aggregates and booking marketplaces.
Why: the question is asked mid-problem and the asker needs two things, proximity and evidence that someone turned up. Review volume and recency are the strongest proxies available.
Practical consequence: in a category with little brand recognition, a thin review profile reads as a risk. This is the vertical where reviews matter most as a direct input.
Financial services and insurance
What the models lean on: regulator registers and established comparison services.
Why: money questions get the same guarded treatment as health questions, and the models are noticeably reluctant to name a single provider.
Practical consequence: being one of several named options is the realistic goal. Appearing consistently across many answers matters more than winning any single one.
B2B SaaS
What the models lean on: peer review platforms and comparison content, including your competitors'.
Why: the buying process was already research-led, and the alternatives question gets asked constantly.
Practical consequence: if the only pages answering "alternatives to X" are your competitors', they decide how you are described. This is the most contested category we measure, with 54 tracked sources.
Restaurants and hospitality
What the models lean on: review aggregates, booking platforms and local editorial.
Practical consequence: a photographed menu is invisible to a model. Dietary detail, group capacity and occasion suitability have to exist as text somewhere or they cannot be matched to a question.
Aged care
What the models lean on: government registers and quality ratings.
Practical consequence: the researcher is almost never the resident. It is an adult child under time pressure, and the content that reaches them explains the system, assessment, funding and waiting times, before it describes any provider.
The pattern underneath
Two things hold across every vertical.
Verified beats claimed. Wherever a category has a register, a licensing body or an insurer directory, that source outweighs anything a business publishes about itself. The models are looking for evidence, and a self-description is not evidence.
The high-weight sources are category-specific. The general directories everyone claims are the ones that differentiate nobody. The sources that matter are the ones a person in your industry would name and a person outside it would not.
How to find yours
Ask an assistant which sources it would use to check whether a business in your category is reputable, and which directories or registers matter in your field. It will tell you, in some detail, and the list is usually not the one you have claimed.
Then check whether you are in each of them, complete and consistent. That exercise is free, takes an afternoon, and is the highest-value thing most businesses can do about AI search.
How much this varies, quantified
The per-industry lists above are not a stylistic difference. The underlying data is genuinely different in every vertical.
We hold 520 scored citation sources across 27 industries, each rated by how much weight it carries when an engine builds an answer in that category. The spread is wide:
- B2B software has the most sources of any vertical we track, and the highest proportion of them are comparison and integration directories rather than anything editorial.
- Healthcare is second by volume, and the highest-weight sources are almost entirely institutional: licensing bodies, insurer directories, professional registers.
- Home services and financial services are close behind, and could not be less alike. One runs on review aggregates, the other on regulator registers and comparison sites.
- Fitness has plenty of sources and very few high-weight ones, which is why it is one of the harder categories to move quickly.
The practical consequence is that a generic directory checklist is close to useless. The sources that decide your category are findable, but you have to find them for your category rather than borrow someone else's list.
What working the right sources produces
“We've always grown through referrals - builders who know us pass our name on. That works, but it only reaches people who already know someone in the industry. Within a couple of weeks of the content going live, we had someone contact us directly through the website. That's a channel we didn't have before - and the enquiries have kept coming.
PROCERT is a building certification firm that had grown entirely on referrals. Within a couple of weeks of the work going live, someone who had never met them inquired directly through the website. Fur Magic went from absent to the most-cited brand in their category in three weeks, ahead of long-established competitors with far more online presence.
Both published with the starting number, the end number and the period: PROCERT and Fur Magic.
How to find yours in twenty minutes
The lists above are a starting point rather than an answer, because within any industry the high-weight sources vary by sub-category and market. Here is how to find your own.
Step one, five minutes. Write down five questions your buyers would genuinely ask, in their words.
Step two, ten minutes. Run each one in a logged-out or temporary chat, across at least three engines. Do not read the answer. Read the citations underneath it.
Step three, five minutes. Tally the domains. The ones that appear across multiple questions and multiple engines are your high-weight sources, and there are usually fewer than ten.
What this beats. Any generic directory checklist, including ours. The tally is derived from the actual answers to your actual questions, which is the only definition of a high-weight source that means anything.
What surprises people. Usually two things. A source they have never heard of turns up repeatedly. And Google Business Profile, which everyone claims, turns out to carry less weight than a category-specific register almost nobody has claimed.
Then check yourself on each. Present, complete, consistent. That list, in that order, is your first month of work.
See which questions name you
A free AI visibility audit on your own buying questions, across all five answer engines.
Book your audit
Ashur Homa
Built and scaled a digital brand to $100M+ in sales with zero ad spend. Has helped businesses generate millions through AI go-to-market strategy. Leads growth at Omni Eclipse.
Connect on LinkedIn
