And what to do before AI search leaves your brand behind entirely
There's a conversation I have with clients more and more often these days. It usually starts with a question like: "Why isn't our brand showing up when people ask ChatGPT about [category]?"
Sometimes the answer is competitive, a better-resourced rival is doing the right things and winning the citation game. But sometimes the answer is more fundamental, and harder to hear: the website itself is structurally incapable of earning AI visibility. Not because of anything wrong with the brand, but because of decisions made years ago about what the website was supposed to be.
The four-page microsite is the clearest example of this problem. And if your brand lives on one, this post is for you.
What Is a Microsite, and Why Did It Seem Like a Good Idea?
A microsite is a stripped-down web property built for a narrow purpose, typically product awareness, a single campaign, a regulatory-compliant HCP portal, or a DTC brand entry point. Done right, they're lean, fast, and focused. A homepage. A "get access" form. Maybe a product information page or a price list. Four pages, maybe six. Clean.
For years, this approach made sense. SEO for branded queries was manageable with even minimal on-page optimization. Paid media drove traffic directly. Legal and regulatory teams appreciated the reduced surface area. Pharma brands in particular embraced the microsite model because it was easier to get approved, easier to update, and easier to control.
Then AI search happened. And the microsite's strengths became its fatal weaknesses. In the age of generative engine optimization, less is definitely not more.
How AI Engines Decide What to Cite , And What to Ignore
To understand the problem, you need to understand how generative AI engines work when they construct an answer. When someone asks ChatGPT, Perplexity, or Google's AI Overviews a question, the model does not just return a blue link. It synthesizes an answer, drawing from training data, real-time web retrieval, and indexed content, and then (sometimes) cites the sources it used.
For your brand to appear in that answer, the AI has to:
- Find your content during its crawl or retrieval pass.
- Understand what your brand does and for whom.
- Judge your content as authoritative and trustworthy relative to other sources.
- Match your content to the specific question being asked.
- Extract a citable, quotable insight that adds value to the synthesized answer.
A four-page microsite fails at nearly every one of these steps.
The Five Structural Failures of the Microsite in an AI Search World
1. There Is Nothing to Cite
AI engines cite content because it answers a question well. Price lists answer the question "how much does this cost?", barely. Access forms answer nothing. An animated homepage built to delight a consumer or impress a prescriber contains almost no prose that an AI can extract and attribute.
For an AI to cite your brand in response to a question like "What are the treatment options for [condition]?", there has to be actual, substantive written content on your site that addresses those questions. Not animation. Not a PDF download gated behind a form.
Prose. Claims. Context. Evidence. That's what gets cited. A microsite has almost none of it. You might have spent $50k on content last year, but if it isn't in a format AI can digest, it might as well not exist.
2. Animated, JavaScript-Rendered Content Is Invisible to AI Crawlers
This one catches a lot of brands off guard. Your homepage may be beautifully designed, perhaps with a clever split experience where the page dynamically changes based on whether the visitor self-identifies as a patient or a healthcare professional. Users love it.
AI crawlers don't see any of it.
Most AI crawlers, including the indexing infrastructure behind Google's AI Overviews and Perplexity, are not full JavaScript rendering engines. They fetch the raw HTML. If your homepage's meaningful content lives inside React components, JavaScript-triggered DOM updates, or audience-toggle scripts, the crawler sees a skeleton. Maybe a loading spinner. Not the brand narrative you spent three months getting approved.
3. You Have No Topical Authority
AI engines favor sources that demonstrate deep, consistent expertise. This is the concept behind E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). Topical authority is built through coverage breadth and depth on a subject. A brand that publishes clinical explainers, mechanism-of-action content, and patient support resources builds an associative signal: this domain knows this topic.
A four-page microsite covers nothing in depth. It makes a claim or two, shows a price, and asks for an action. There's no topical footprint for the AI to anchor to. When the model is deciding whether to cite your brand or a competitor with a rich content library, the choice is straightforward, even if your product is clinically superior.
4. You Are Generating No Structured Data the AI Can Process
One of the most powerful tools in the AI visibility toolkit is schema markup: structured data that explicitly tells crawlers what your content is. Organization schema. MedicalCondition schema. FAQPage schema.
A microsite almost never implements comprehensive schema. Not because the team doesn't want to: but because there's nothing to mark up. You can't implement FAQPage schema if you have no FAQ content. The sparse content creates sparse structured data, which leads to sparse AI retrieval. This is a key reason why traditional SEO isn't enough for the AI era.
5. You Have No Answer to the Long Tail of Questions
Think about how your target audience actually uses AI search. It's rarely a branded query out of the gate. As we discuss in our guide on what is a prompt, the new search query is conversational. It looks like this:
- "What's the difference between [drug class A] and [drug class B]?"
- "What should I ask my doctor about [treatment]?"
- "Is [condition] treatable with [mechanism]?"
These are the moments of intent that precede a decision. Your microsite doesn't answer any of these questions. So when these moments happen, your brand is simply absent.
"But We're in a Regulated Industry. We Can't Just Publish Content."
This is the objection I hear most often. Regulatory review cycles are real. MLR (Medical, Legal, Regulatory) processes exist for good reasons. Pharma brands can't publish unvetted claims.
All of that is true. And none of it means you're stuck with four pages.
There is a substantial body of content that can be developed, reviewed, approved, and published without triggering the highest-risk promotional review categories:
- Disease state education: Mechanism of disease, diagnosis, and burden of illness.
- Treatment landscape overviews: Without making comparative claims that require head-to-head data.
- Patient support resources: Adherence tips and "how to talk to your doctor" guides.
- FAQ content: Built around commonly asked questions: reviewed and approved like any other promotional asset.
- llms.txt files: A simple text file that helps AI models find your most important information quickly.
The question isn't whether content is possible. It's whether the organization has prioritized it. Many haven't: because they built their digital strategy in an era when microsites were sufficient. That era is ending.
What This Means for Your AI Visibility Audit
When I conduct an AI visibility audit for a brand living on a microsite, the report tells a consistent story:
- Citation rate near zero across major AI platforms.
- No brand presence in response to category or treatment queries.
- Competitor brands with richer content ecosystems capturing share of voice by default.
That last point is the hard truth. AI visibility isn't a technical fix. It's not a metadata tweak or a sitemap submission. It's a content problem, and content problems require content solutions.
The good news is that this is fixable. The brands that act now: that treat AI visibility as the new frontier of organic brand presence: will establish authority that compounds over time. You need to start by tracking your AI citations to see exactly where you stand.
The Bottom Line
Your microsite was built for a different era of digital discovery. In that era, you could drive awareness through paid channels and capture branded search with minimal effort.
AI search has changed the calculus. When a patient asks an AI what treatment options exist, the brands that show up are the brands with content worth citing. A four-page microsite, built around a form and an animated homepage that AI crawlers can't read, cannot participate in that conversation.
The question isn't whether your brand needs more content. It's how quickly you can build it.
Ready to see how invisible your site really is? Get your ModelScore → citemetrix.com
