The 20-Minute Test That Tells You If AI Has Ever Heard of You

Across 1,700 companies in 32 industries, 88% never appear when someone asks ChatGPT for a recommendation (Omni Eclipse's 2026 AI Search Visibility Report). Healthcare is not the exception to that — it's one of the industries in it. Not because these are bad practices. Because nobody told them AI search doesn't work like Google, and nobody's going to tell you either unless you go check yourself. So go check yourself. Twenty minutes, twelve questions, five platforms, and you'll know exactly where you stand — which, statistically, is probably not where you think.

• 88% of businesses never appear in ChatGPT recommendations (Omni Eclipse's Jan–Mar 2026 audit of 1,700 businesses) — invisibility here is the default, not the exception.
• The test: twelve questions across ChatGPT, Gemini, Claude, Perplexity, and Google AI Overviews — twenty minutes, no tools, no cost.
• Each platform reads different sources: ChatGPT draws on Bing (tests suggest it sometimes uses Google as a fallback), Gemini and Google AI read Google plus your Business Profile, Claude leans on training data, Perplexity crawls live. Fixing one doesn't fix the others.
• Fix in order: Bing Webmaster Tools (30 min), Google Business Profile (1 hour), schema markup (2–4 hours). Check monthly, not weekly.

Why This Is Worth Twenty Minutes of Your Day

Nearly half of patients — 47%, up from 31% just nine months earlier — now use AI to research providers, according to rater8's 2026 Patient Choice Report. Among adults aged 45 to 60, the group booking implants, LASIK and fertility care, it's 64%. They ask "who's the best dermatologist in Austin," and ChatGPT gives back one name. Not ten links to weigh against each other. One name. If it isn't yours, you weren't in the running — you were never in the room.

The one genuinely good part of those numbers: almost nobody has fixed this yet. Which means the practices that do this test today and act on it are, for a little while, competing against almost no one.

What You Need

Your phone or laptop, 20 minutes, and somewhere to jot down what each tool says. That's the entire supply list.

Step 1: Ask ChatGPT

Type these three in, swapping in your real details.

Question 1

Who is [your name]? Are they a [specialty] doctor in [city]?

Correct name, specialty, and city back means you're on the map. "I don't have information" or wrong details means you're not.

Question 2 — the one that actually matters

Who should I see for [specific service] in [city]?

Recommended by name: visible. "I can't recommend specific providers," or your competitor's name instead of yours: invisible. There's no in-between reading of this one.

Question 3

What's the difference between [your practice] and [competitor] in [city]?

If ChatGPT can only describe one of you, guess which one it's going to send the patient to.

Write down or screenshot every answer. You'll want the receipts later.

Step 2: Ask Perplexity

Perplexity crawls live instead of running off a fixed index, so it sometimes finds practices ChatGPT misses entirely.

Question 4

[Your practice name] [city]

Your site should show up in the cited sources. If it doesn't, your site isn't built in a way AI can find.

Question 5

Best [specialty] in [city] for [specific service]

Same question as ChatGPT's. Compare the two answers side by side — they will not necessarily agree, and that disagreement is itself useful information.

Question 6

What do patients say about [your practice name]?

"No information available" means your review data isn't reaching AI at all, regardless of how many five-star reviews you actually have sitting on Google.

Step 3: Check Google AI Overviews

Healthcare leads every industry for AI Overview coverage — roughly 88% of health queries return one, per BrightEdge. Other datasets put it lower; WebFX measured 51% across 130,000 health queries. The methodologies differ, the direction doesn't — this is the AI-generated summary sitting above the actual results, and it is showing up more each year.

Question 7

[Specialty] near me

Search it on your phone. If an Overview appears, scan it for your name. Seeing everyone but yourself in that box is exactly as bad as it sounds.

Question 8

How do I choose a [specialty] in [city]?

Same test, different angle. In the answer or not?

Step 4: Ask Gemini

Gemini reads Google's own record of you — the index, Business Profile, reviews. It's the platform where your Google SEO work and your AI visibility actually overlap, so it's usually the most forgiving of the five. Check it anyway; the disagreements between engines are the useful part.

Question 9

Who is [your practice name] in [city]?

Name, specialty, city correct back means Google's record of you resolves cleanly. "Here's what I found" with three competing businesses means it doesn't.

Question 10

Best [specialty] in [city] for [specific service]

Same question again on purpose. If Gemini names you and ChatGPT doesn't, that's not a pass — it's a Bing indexing problem showing itself.

Step 5: Ask Claude

Smaller footprint than ChatGPT, growing fast, and it rewards actual substance over marketing copy more than any of the others.

Question 11

Tell me about [your practice name] in [city]

Real, detailed service pages give Claude something to work with. A site that's mostly "compassionate, patient-centered care" language gives it nothing — which, fair, that phrase doesn't tell a patient anything either.

Question 12

Who are the best [specialty] doctors in [city]?

Last recommendation test. In or out.

Reading Your Results

Showed up everywhere, correctly — genuinely rare. Recheck monthly and don't get comfortable; competitors catch up.

Showed up on some platforms, not others — useful, because each gap points at a different fix:

• Invisible on ChatGPT → you're probably not indexed on Bing. Thirty minutes fixes this.
• Invisible on Gemini → your Google Business Profile or your practice's own details disagree with Google's record. An hour of cleanup fixes this.
• Invisible on Perplexity → your content isn't structured for AI to extract from. Schema and clearer content fix this.
• Invisible on Google AI → your Google Business Profile is incomplete. An hour fixes this.
• Invisible on Claude → your content is too thin or too promotional. Only more substance fixes this — there's no shortcut version.

Invisible everywhere — that's the default, not the exception. Not a crisis, a starting line. Everyone in this bucket has the same three fixes below.

The Fixes, in Order

Bing — fixes ChatGPT — 30 minutes

Go to bing.com/webmasters, sign in, add your site, submit your sitemap, request indexing for the homepage. That's the entire list. ChatGPT draws on Bing's index, and tests suggest it sometimes falls back to Google — if you're missing from Bing, you're missing from ChatGPT's main source, no matter how good your content is.

Google Business Profile — fixes Google AI — 1 hour

Fill out every field. Name, address, phone, hours, services, specialties, actual named procedures — not "various treatments." Add photos. Respond to reviews. An incomplete GBP is the single most common reason practices get skipped in local AI Overviews, and it's also the fix requiring the least skill of anything on this page.

Schema markup — fixes everything — 2 to 4 hours

Code that tells AI exactly what your practice is, instead of making it guess: MedicalOrganization or Physician schema, FAQPage schema for your FAQ section, Service schema for treatments. If you're not technical, hand this list to whoever built your site. This is the difference between AI extracting your information and AI ignoring it.

How Often to Actually Check

Once a month. AI engines take 4 to 12 weeks to fully process changes — checking weekly just shows you the same unchanged result and wastes twenty minutes you could spend elsewhere. Check, fix what's broken, wait, repeat.

Tools, If You Want to Track This Over Time

You can run this entire test by hand, for free, forever. If you'd rather have something watching it continuously: Otterly.ai (~$29/mo) tracks presence across ChatGPT, Perplexity, and Google AI; Ahrefs' AI Visibility Checker (free) spot-checks it; ZipTie (free tier) watches Google AI Overview mentions specifically. None of these replace a manual audit — see why a single score isn't the same as an answer — but they're a reasonable way to keep an eye on things between checks.

Frequently Asked Questions

How do I check if my practice shows up in AI search?

Run the twelve questions above across ChatGPT, Gemini, Claude, Perplexity, and Google AI, and write down what each one says. Correct information across the board means you're visible. Anything else means you have a specific, fixable gap.

Why doesn't my practice show up in ChatGPT?

ChatGPT draws on Bing's index — tests suggest it sometimes uses Google as a fallback — and not Google's directly. If your sitemap was never submitted to Bing Webmaster Tools, ChatGPT's main source has never been told you exist. Submit it. No amount of Google SEO changes that.

Why doesn't my practice show up in Google AI Overviews?

Overviews pull from your Google Business Profile and your site's structured data. Incomplete GBP or missing schema means nothing to extract.

Why doesn't my practice show up in Perplexity?

Perplexity searches live. No clear, structured, factual content means nothing worth citing when it crawls you.

Can I pay to appear in AI recommendations?

No. Nobody's selling that slot. The practices that show up are the ones with clean, consistent, well-structured data — that's the entire ranking factor.

What if my practice appears but with wrong information?

Worse than not appearing at all — wrong information actively erodes trust instead of just failing to build it. And it is common: two-thirds of patients who used AI to research a provider hit incorrect provider information, while 60% trusted the AI's summary without verifying it (rater8's 2026 Patient Choice Report). Trace it back to the source (usually a stale directory listing), fix it there, and recheck in a month.

Key Takeaways

• 88% of businesses never surface in ChatGPT recommendations. Fixing it now means you're ahead of almost everyone.
• Twelve questions, five platforms, twenty minutes.
• Each platform reads different sources — ChatGPT draws on Bing, Google AI reads Google, Perplexity reads live.
• Check monthly, not weekly. AI is slow to update; weekly checks just waste your time.
• Wrong information beats no information for being a problem, not the other way around. Fix the source first.

If the test comes back bad — or comes back with the wrong information about you — that's exactly what a real AEO audit maps and fixes. Claim Free Snapshot or see how the process works.