Free: a 48-hour Growth Audit and a 30-minute marketing masterclass. Learn what works before you spend. Get the audit →
The Local AI Findability Score · August 17, 2026

We scored 121 businesses on AI legibility. The local clinic beat the national brand.

Five industries, one identical rubric, every input a public file anyone can check. The venture-funded DTC supplement brands finished last, behind single-location clinics running off-the-shelf website builders. The reason is not talent. It is defaults.

121 businesses5 industries100-point rubricReproducible
58/100average AI Visibility Score across all 121 sites
79%tell an engine who they are
14%give it answers it can lift
4of 121 sites scored a perfect 100
The league table

Average AI Visibility Score by industry.

Five industries, 121 businesses, scored out of 100

Higher is more legible to an AI answer engine. Collected August 17, 2026.

1
Weight loss clinics14 sites measured
64
2
Med spas20 sites measured
63
3
Marketing agencies39 sites measured
61
4
Men's health / TRT clinics18 sites measured
58
5
DTC supplement brands30 sites measured
47
The result we did not expect

Weight loss clinics averaged 64. DTC supplement brands averaged 47, a gap of 16 points. The group with the largest budgets, the best engineering teams, and the most at stake commercially came last, beaten by single-location clinics in Orange County.

Why

Sophistication and legibility are not the same thing.

The obvious explanation would be that clinics care more or know more. Having looked at the markup on both sides, we do not think that is it.

Local clinics tend to run on website builders and local-SEO plugins that emit LocalBusiness and review markup automatically. The dentist did not decide to be legible to ChatGPT. Their platform decided for them, years ago, for local-SEO reasons that predate answer engines entirely.

Large DTC brands run bespoke headless storefronts. Nothing is automatic. Every piece of schema is a deliberate engineering ticket competing against a checkout experiment for sprint capacity, and the ticket that says "add Review markup" keeps losing. The brand out-engineered its way into being harder to read.

The competitive advantage here is not sophistication. It is a default someone else chose for you, and most large brands opted out of those defaults the moment they went custom.

The universal gap

Everyone solved identity. Almost nobody solved answers.

Share of all 121 businesses carrying each component

Measured on the homepage. Each component is worth 20 points of the score.

Machine-readable identitythe engine knows who you are
79%
Structured data that parsesthe markup actually works
83%
Trust markupratings and reviews it can quote
17%
Answer markupquestions it can lift answers from
14%

This is the finding that survives across every industry we measured. The first two components are largely solved: 79 percent declare a machine-readable identity and 83 percent ship structured data that parses cleanly. Those are the things traditional SEO rewarded, so the industry built them.

The last two are not solved anywhere. 14 percent carry answer markup and 17 percent carry trust markup. Those are the things answer engines actually consume when they assemble a recommendation, and essentially nobody has built them yet.

The detail

Component coverage by industry.

IndustrySitesIdentityValidAnswersTrustScore
Weight loss clinics147979362964
Med spas20859553063
Marketing agencies397987231361
Men's health / TRT clinics188383112258
DTC supplement brands3070700347
How to read this

Each cell is the share of that industry carrying that component, as a percentage. The pattern runs left to right in every single row: strong on identity and validity, collapsing on answers and trust. No industry in this sample breaks the pattern.

The uncomfortable one

We put the agencies in the sample on purpose.

If this gap were simply a knowledge problem, the firms that sell search marketing for a living should sit at the top of the table. We measured 39 agencies that market to these exact clinic verticals.

They averaged 61, placing them mid-table, and 5 of 39 scored 20 or below on their own websites. Several of the lowest scorers sell AI visibility or SEO as a headline service.

We are an agency. We published a study that invites exactly this scrutiny of us, because a benchmark you exclude yourself from is marketing, not research.

The charitable read, and we think the correct one, is that agency websites are the cobbler's children. The work goes to clients and the shop's own site runs on a theme nobody has audited since launch. It is still worth knowing before you hire someone to fix a problem their own site has.

The rubric

How the score is built.

Five components, 20 points each, every one objectively checkable from a public file.

20 pointsIdentity

Organization, LocalBusiness, ProfessionalService or an equivalent type. Tells the engine what kind of entity you are.

20 pointsValidity

Every JSON-LD block on the page parses. Broken markup scores zero, because a parser that fails gets nothing.

20 pointsAnswers

FAQPage or QAPage markup. Gives an engine clean question-and-answer pairs instead of prose it has to infer from.

20 pointsTrust

AggregateRating or Review markup. The evidence an engine leans on when it has to justify recommending you over someone else.

20 pointsAccess

robots.txt does not root-block the named AI crawlers. Half credit where no robots.txt could be read at all.

Scoring noteUnreachable sites

A site that would not return a homepage scores zero on the four content components. We report how many, per industry, rather than hiding them.

Method

Sample, limits, and what we refused to do.

Perfect scores

4 of 121 businesses scored 100 out of 100: mgmtdigital.com, plumpmedicalspa.com, prestigemedigroup.com, pulsedigital.health. Each one carries identity, valid markup, answer markup and trust markup on its homepage, and blocks nothing.

Questions

About the index.

What is the AI Visibility Score?

A 0 to 100 score built from five equally weighted, machine-checkable components worth 20 points each: whether the site declares a machine-readable identity, whether its structured data actually parses, whether it carries answer markup, whether it carries trust markup, and whether its robots.txt lets AI crawlers in. It measures legibility to an answer engine. It does not measure quality, revenue, or whether the business is good.

Why did national DTC brands score lowest?

DTC supplement brands averaged 47 against Weight loss clinics at 64. The likeliest explanation is platform defaults. Local clinics tend to run on website builders and local-SEO plugins that emit LocalBusiness and review markup automatically, while large DTC brands run bespoke headless storefronts where every piece of schema is a deliberate engineering ticket that nobody filed. Sophistication and legibility are not the same thing.

Does a high score mean AI engines will recommend me?

No. This measures whether an engine can cleanly parse and use your site once it has found it. It does not measure whether the engine finds you, which depends on authority and how often credible third parties mention you. A perfect score removes a barrier. It does not create demand, and we would rather say that plainly than sell you something else.

Why include marketing agencies in the sample?

Because it is the fairest test of the thesis. Agencies sell this exact capability, so if the gap were merely a knowledge problem they should be at the top. They averaged 61, effectively the middle of the pack, and 5 of 39 scored 20 or below. We included ourselves in that reasoning: we are an agency, and we published this knowing it invites the same scrutiny.

How were the businesses selected?

Every business came from lists we had already researched and published for other purposes, which prevents us from cherry-picking a sample to fit a conclusion. Clinic groups are Orange County and Southern California businesses. Supplement brands are national DTC brands named across independent 2026 listings. Agencies are firms that market to the same clinic verticals. This is a convenience sample of real, findable businesses, not a random sample of any industry.

Why is the answer and trust markup so low everywhere?

79 percent of these sites declare who they are and 83 percent ship structured data that parses cleanly, so the basic technical layer is largely solved. But only 14 percent carry FAQ or Q&A markup and only 17 percent carry rating or review markup. The industry built the layer that traditional SEO rewarded and stopped before the layer that answer engines actually need.

Google deprecated FAQ rich results. Why does FAQ markup still count?

Google ended FAQ rich results in May 2026, so the markup no longer earns a SERP dropdown. Google also confirmed it still reads the markup to understand pages, and answer engines outside Google use it for clean question-and-answer extraction. We score it because it still does the job we are measuring, not because it still wins a rich result.

Can I reproduce this?

Yes. Every input is a public robots.txt file or the public HTML of a homepage. The methodology below states the sample, the date, the crawler list, the scoring rules, and the exclusions. Sites change, so results will drift from our collection date.

Want your own score?

We will run this exact rubric on your site and send back the score, the missing components, and what each one takes to fix. Free, and we will tell you when the honest answer is that you do not need an agency for it.