Healthcare SEO Is Collapsing. GEO Is the New Gatekeeper.
If you run growth for an OTC brand, or a specialty clinic, here’s the uncomfortable reality: the organic patient funnel you’ve been investing in for a decade is being quietly rerouted into large language models.
Google’s AI Overviews — those big AI-generated summaries at the top of the results page — are already answering most mainstream health questions directly in the search results. For many queries, people now get the explanation, risk factors, and “when to see a doctor” guidance without ever needing to click a website. AI Overviews launched in May 2024; by mid-2025, they were already expected to appear on more than 80% of informational queries, up from 47% the year before. (Healthcare Success)
The consequences are measurable. Search impressions are up, but actual clicks out to the open web are down ~30% year-over-year, according to BrightEdge data cited by Search Engine Land on May 15, 2025. (Search Engine Land)
Healthcare is one of the most affected verticals. BrightEdge found that AI Overviews are showing for ~83%+ of “what is / symptoms / treatment” style healthcare queries — far higher coverage than in retail categories — and Google is preferentially citing high-authority medical systems, government guidelines, and major nonprofits. (brightedge.com)
In other words: Google is now answering the top of your funnel for you, using content it trusts, and your site might never get the click.
“But we’re a hospital / brand with authority. We’ll be fine.” Not necessarily.
Health systems are already taking hits. Healthcare Success, which tracks hospital marketing performance, reports that prominent U.S. hospital brands (Cleveland Clinic, Mayo Clinic, Cedars-Sinai) have seen organic traffic fall by more than 10% month-over-month since AI Overviews rolled out. (Healthcare Success)
Industry-wide, clickthrough rates on health-related search results dropped roughly 34% between May and December of last year as AI boxes took over above-the-fold real estate. (EMARKETER)
And remember: those are world-class brands with medical credibility and robust content teams. If it’s happening to them, local orthopedic groups, urgent care clinics, fertility practices, dermatology chains, and telehealth startups are even more exposed.
Here’s why healthcare is ground zero:
- High trust stakes. Google and other AI systems lean heavily on “safe” sources in medicine (NIH, major academic hospitals, guidelines bodies). That crowds out everyone else. (brightedge.com)
- Regulated language. If you’re in pharma / OTC, you know the dance: you can’t make unapproved claims. AI, meanwhile, summarizes aggressively. That makes you structurally slower to publish the kind of conversational, symptom-led, plain-language answers LLMs prefer to quote.
- Zero-click behavior. Similarweb and publisher data show that “zero-click” searches — where users never leave the results page — jumped from roughly 56% to 69% between mid-2024 and May 2025, and total organic visits to content sites fell from ~2.3B to <1.7B over that same period. Publishers say some articles lose up to 79% of their traffic when an AI summary appears first. Google disputes the methodology, but the pattern is obvious: users are satisfied without clicking. (New York Post)
Now apply that to healthcare discovery. Fewer clicks means fewer appointment requests, fewer email captures (“Download our ACL rehab guide”), and fewer first-party data opportunities for remarketing. For OTC and self-care brands, it means fewer chances to intercept intent (“sore throat at 2am”) before the consumer heads to Amazon, the pharmacy aisle, or a retailer chatbot.
Patients aren’t “Googling symptoms” anymore. They’re conversing.
The search behavior itself is mutating.
Patients are moving from keywords (“best orthopedic surgeon near me”) to natural language (“I tore my meniscus playing football, do I need surgery or physio and who should I see nearby?”). AI then returns a short curated answer — including which provider types to consider — right at the top of the page. The American Alliance of Orthopaedic Executives notes that patients increasingly get this AI-curated summary of “best options” without ever clicking through to a provider’s site. (aaoe.net)
At the same time, consumers are now openly using general-purpose chatbots — not just search engines — for health guidance:
- Pew Research Center found that by June 25, 2025, 34% of U.S. adults had used ChatGPT at least once, including 58% of adults under 30. (Pew Research Center)
- A recent survey cited by Rolling Stone and Yahoo News reported that 31% of Americans are already using chatbots to prepare questions for their doctor visit, and 23% have asked a chatbot directly for medical guidance, with nearly 40% saying they trust that advice. (Yahoo)
- Nursing and patient safety groups point out that 17% of adults use AI chatbots for health information today — and among 18–29-year-olds, roughly 1 in 4 are doing this monthly. Most people still say they’re not fully confident in chatbot accuracy (about 63% express concern), but they’re doing it anyway. (ons.org)
In parallel, specialty clinicians are seeing another signal: review credibility has become algorithmic fuel. In orthopedics, 84% of patients check online reviews before choosing a provider, and 61% say they trust those reviews more than personal recommendations — and those review snippets are being pulled straight into AI-generated provider summaries. (aaoe.net)
This is not “future behavior.” This is how patients are already shopping for care in August–October 2025.
Pharma / OTC: welcome to the AI health funnel
For consumer health and OTC, the funnel is getting rewritten in a way that should make every brand manager sit up.
LLM-style assistants (think ChatGPT, Gemini, Perplexity, Amazon’s Rufus) aren’t just answering “why does my throat hurt?” — they’re starting to suggest what to buy and in some cases where to buy it. Industry strategists now describe these models as gatekeepers of demand: if your product is the one the model names, you win the shelf before the shopper ever sees a shelf. (FRANKI T)
Consumer health marketers are already warning: if your brand doesn’t show up in those AI answers, you’ve effectively lost salience at the exact point of need. And that lost mention doesn’t just cost you impressions — it hands your competitor the recommendation moment. (Varn Health)
Pharma marketers are feeling the same squeeze in disease education. Agencies advising top pharma clients are saying (a) zero-click AI answers are eroding traditional branded and unbranded landing-page traffic, and (b) this forces teams to build content that matches what humans actually ask, in human language, while also surviving legal review and safety constraints. They’re also blunt that AI hallucinations are now a brand risk: if the model gets dosing, contraindications, or safety wrong, you’re dealing with a trust event you didn’t cause — but your name might still be in the paragraph. (fiercepharma.com)
And yes, hallucination risk in health is real. Researchers at Mount Sinai (August 6, 2025) showed that mainstream AI chatbots can be easily “led” into repeating and even elaborating on false medical claims with total confidence. Reuters reported similar findings in July 2025: several leading LLMs could be prompted to give authoritative-sounding but flatly incorrect health advice, complete with fake citations. (Mount Sinai Health System)
So pharma/OTC is now dealing with two fronts at once:
- Visibility risk. Are we even mentioned in the answer when a consumer asks about a symptom?
- Liability / trust risk. If we are mentioned, is the answer safe, accurate, and on-label?
Let’s talk CAC: when organic dries up, you pay — a lot
Healthcare marketers used to love SEO because it funneled motivated, self-triaging patients to owned assets (“hip pain quiz,” “Is this strep throat?”). That discovery path fed your nurture programs, physician finders, Rx copay cards, or “book consult” CTAs.
If AI search now captures that intent without sending traffic, you’re left with two choices:
- get named inside the AI answer (we’ll come back to this), or
- buy attention.
The “buy” part is brutal.
Healthcare and adjacent high-stakes service categories are already among the most expensive in paid search. In elective care and aesthetics, clinics are paying well above $20 per click for high-value procedures, and it’s not unusual for competitive sub-specialties to see CPCs push $50. Dentistry is commonly $6–$11 per click, but cosmetic surgery, medspas, and other cash-pay verticals climb fast because lifetime value is high. (Anzolo Medical)
Cost per lead is spiking too. Benchmarking across healthcare verticals shows cosmetic surgery averaging a cost per lead north of $130, versus ~$32 for general hospitals and clinics, with an industry-wide average >$50. (promodo.com)
If you drift into “medical + legal risk” areas (malpractice, surgical injury, etc.), the auction gets downright insane: some “medical malpractice lawyer” keywords are trading at $70–$250 per click, and in certain metros even higher. That’s because one converted lead can be worth tens of thousands. (PPC Agency)
Translation for provider groups and OTC brands: if AI Overviews and chatbots intercept symptom intent before it ever hits your site, your blended cost to acquire a patient (or a consumer for your product) is going to rise — sometimes sharply — because you’re forced to compete in the most expensive ad auctions in digital media.
And on top of that, attribution is breaking. Marketers point out that for most of 2024–2025, Google Search Console didn’t even tell you when your content was surfaced inside an AI Overview, so you literally couldn’t see whether you influenced the user even if you didn’t get the click. Healthcare marketers are calling this “invisible impact”: your expertise powers the AI answer, the user is satisfied, but your dashboard shows nothing. (Healthcare Success)
“GEO”: the operating system for this new world
This is why everyone is suddenly talking about GEO — Generative Engine Optimization.
GEO is the discipline of getting your brand (or product, or clinician, or clinic) cited inside AI-generated answers from engines like Google’s AI Overviews, ChatGPT, Gemini, Perplexity, and retailer assistants like Amazon’s Rufus. Wired recently called GEO “the world brands are sprinting toward,” and highlighted that the overlap between classic Google top links and what chatbots cite has collapsed from ~70% to below 20%. In other words, ranking #1 on Google no longer guarantees you get mentioned by an AI assistant. (wired.com)
Brands like Estée Lauder, LG, and Aetna are already restructuring product and service information into highly specific, machine-readable chunks — FAQs, bullet answers, structured attributes — because that’s what LLMs prefer to ingest and reuse. The goal is simple: when someone asks the AI a specific, high-intent question (“best lotion for sunburn blister?” / “how to manage heartburn at night?”), your brand is in the answer. (wired.com)
Google is openly saying this is the direction: more conversational queries, more complex questions, and “higher quality clicks” when people do leave the AI box. Google claims AI Overviews drive more queries overall, expose more links, and deliver clicks that stay longer and engage deeper. (blog.google)
Publishers, hospitals, and pharma marketers counter that “quality of clicks” doesn’t pay for lost volume, lost brand exposure, or lost first-party data — especially in healthcare, where trust and conversion tend to happen after multiple touchpoints, not in one instant answer. (Healthcare Success)
Both things can be true:
- AI search can deliver warmer, later-stage clicks.
- AI search can also starve the top of the funnel.
So the new KPI can’t just be “sessions from organic search.” Healthcare marketers are starting to track brand-led intent instead: how often do people search for you by name (vs. generic symptom terms), how often do you get cited in AI answers, how often do reviews and ratings for your clinicians show up in AI summaries. Some call this “branded search visibility,” and they’re treating it as the new north star for awareness and trust. (Healthcare Success)
What health & OTC marketers should actually do right now
Here’s the short list we’re advising:
- Structure your expertise for machines (not just humans).
Turn your clinical/medical guidance into tightly scoped Q&A, decision trees, dosage tables, contraindication callouts, and “when to see a doctor” guidance. LLMs prefer concise, structured answers they can safely quote — not glossy hero copy. This is literally what big brands are doing to win GEO right now. (wired.com) - Own the symptom moment — within compliance.
People don’t ask “Is Brand™ good?” in ChatGPT. They ask “What can I take for nighttime acid reflux if I’m also on blood pressure medication?” If you’re pharma / OTC, you need approved, high-signal content that maps to real symptom language and real usage context. Otherwise the AI will answer that question using somebody else’s content — and possibly recommend somebody else’s product. (FRANKI T) - Engineer trust assets that AI will lift.
Reviews, clinician bios, credentials, operating hours, locations, “accepted insurance,” safety guidance, black box warnings, pregnancy precautions, etc. All of that should be accurate, current, marked up with schema, and consistent everywhere (site, profiles, directories). AI Overviews and LLMs are already scraping those structured facts and review snippets to decide who’s “reputable.” In orthopedics, for example, AI surfaces star ratings and review language right in the zero-click summary. (aaoe.net) - Monitor (and correct) what AI is already saying about you.
Do regular pulls of:
- “What does ChatGPT say about [brand/product]?”
- “Who are the top [specialty] clinics in [city]?”
- “Is [active ingredient] safe during pregnancy?”
If you see unsafe, off-label, or outdated claims, you’ve identified both a reputational risk and a content gap. Mount Sinai’s August 2025 study and other analyses show how easily LLMs can repeat confident but wrong medical statements. You need escalation paths internally (medical, legal, pharmacovigilance, PR) when that happens. (Mount Sinai Health System)
- Rethink spend allocation through a CAC lens, not a channel lens.
Don’t just “shift budget from SEO to paid search.” Paid search in health is getting brutally expensive — $20–$50 CPCs for high-value procedures are now normal in aesthetics and medspa; some medically-adjacent lead gen terms go $70–$250 per click. (Anzolo Medical)
Instead, ask: “What’s the cheapest way to get named — credibly — at the moment of intent?” Sometimes that means content (GEO). Sometimes it means review hygiene. Sometimes it’s still paid, but maybe not just Google Ads (local service ads, retailer assistants, pharmacy marketplace placement, etc.). (promodo.com) - Start reporting on ‘AI answer share,’ not just rankings.
Classic rank tracking (“We’re #2 for ‘knee pain at night’”) is becoming less meaningful if an AI box sits above every result. The practical question is: “Are we in the answer block patients see?” Healthcare marketers are already building internal dashboards that log whether their hospital, clinician, product, or guidance appears in AI Overviews / ChatGPT answers for critical queries, because that is now top-of-funnel visibility. (Healthcare Success)
The punchline
Healthcare spent years mastering SEO. That world is disappearing in real time.
AI systems — Google’s AI Overviews, ChatGPT-style assistants, Amazon’s retail LLMs — are becoming the first touchpoint for symptom research, provider selection, and OTC product choice. They’re absorbing intent, rewriting comparison shopping, and compressing entire patient journeys into a single synthesized answer. (Search Engine Land)
If you’re not present in that synthesized answer, you’re invisible at the exact moment of need. If you are present but the answer is unsafe or off-label, you have a regulatory and trust crisis. Both scenarios are existential for growth teams in pharma, OTC, and provider marketing.
At BrandHack.AI, we call the discipline to solve this GEO-as-a-Service: making sure your brand is (1) machine-readable, (2) medically accurate, (3) discoverable in AI answers, and (4) defensible from a compliance and trust standpoint.
This isn’t “SEO, but with some AI keywords.” This is the new distribution layer for health decisions. Owning it is how you keep patient acquisition cost sane — and how you keep your brand from vanishing.
—
