We Tested 99 Health Websites Across ChatGPT, Claude, and Google AI Overviews. Here's What We Found About WordPress vs Next.js.

AI is eating search. Google AI Overviews now sit above organic results for most health queries. ChatGPT and Claude are becoming the first place people go to find a doctor, choose a fitness app, or compare telehealth platforms. The question every health brand should be asking isn't "how do I rank on Google?" anymore — it's "how do I get cited by AI?"

We wanted to know: does your tech stack actually matter? Specifically, do Next.js sites have an edge over WordPress when it comes to Generative Engine Optimization (GEO)?

So we ran a study. 99 health industry websites. 49 built on Next.js, 50 on WordPress. 25 real health queries. Three AI platforms. Here's what the data says.

The Headline Numbers

Next.js sites captured 70% of all AI citations across ChatGPT, Claude, and Google AI Overviews. Out of 163 total study site mentions, 114 went to Next.js sites and 49 went to WordPress.

But the gap goes deeper than raw counts:

81.6% of Next.js sites were cited by at least one AI platform. 48.0% of WordPress sites were cited by at least one AI platform. Next.js sites averaged 2.29 citations per site vs WordPress at 1.02.

That's not a marginal difference. Next.js sites were cited at nearly double the rate.

What the AI Platforms Actually Recommended

We asked all three platforms questions that real people ask every day — things like "what are the best telehealth platforms," "what apps help with anxiety and meditation," and "what are the best health wearables."

ChatGPT was the most selective — and the most brutal for WordPress. When someone asks ChatGPT for a recommendation, it names 5-7 brands and moves on. The brands it chose? Overwhelmingly Next.js: Peloton, Strava, Calm, Headspace, GoodRx, Equinox, Dexcom. WordPress sites were largely invisible. These are companies that invested in modern web architectures, and ChatGPT rewards that investment by citing them first.

Claude was the most generous with citations overall, mentioning 103 study sites across 25 queries — but even with a wider net, Next.js sites still dominated with 72% of mentions. WordPress sites that did get cited tended to be large publishers with massive content libraries. Smaller WordPress sites were almost entirely shut out.

Google AI Overviews showed the strongest skew of all. When the AI Overview panel appeared (we tested 10 queries), 78% of study site citations went to Next.js properties. WordPress sites struggled to break through even when they ranked well in traditional organic results. Google's AI clearly favors the kind of fast, well-structured sites that Next.js produces out of the box.

Why Next.js Sites Win in AI

We dug into the data to understand why this gap exists. It comes down to four things.

1. Next.js Powers the Brands AI Trusts

The companies building on Next.js in the health space tend to be the ones investing heavily in product, content, and brand. Think Healthline, GoodRx, Calm, Whoop, Peloton, Dexcom. These aren't small operations — they're category leaders with deep content libraries, millions of users, and strong brand signals across the web.

AI models learn from the internet's collective knowledge. When Healthline shows up in thousands of medical articles, Reddit threads, and "best of" lists, AI systems internalize that authority. When a local med spa has a WordPress site with good Yoast SEO scores but limited off-site presence, AI has far less signal to work with.

The takeaway: Next.js has become the default choice for ambitious, tech-forward health companies. That selection bias means the Next.js ecosystem is stacked with exactly the kind of brands AI engines prefer to cite.

2. WordPress's SEO Plugins Don't Help with AI

Here's the uncomfortable truth for WordPress loyalists. WordPress sites in our study actually had better traditional SEO scores than Next.js sites. WordPress averaged a 5.64 SEO score (out of 9) compared to 4.73 for Next.js. WordPress sites had more schema markup (1.42 average schema types vs 0.65), more Open Graph tags, and more Twitter cards — largely thanks to Yoast SEO doing the heavy lifting.

None of that translated into more AI citations. Not even close.

This exposes a fundamental problem with the WordPress approach to discoverability. Plugins like Yoast optimize for a search paradigm that's being replaced. Schema markup, meta descriptions, and structured data are table stakes for Google's organic results, but AI citation engines don't care about your Yoast score. They care about brand authority, content depth, and real-world recognition — signals that no plugin can fake.

If your GEO strategy starts and ends with "install Yoast and add more schema markup," you're optimizing for yesterday's algorithms while your Next.js competitors own tomorrow's.

3. WordPress Is Slow — and AI Penalizes Slow Sites

Page speed isn't just a UX metric anymore — it's an AI visibility signal. The average WordPress health site in our study loaded in 4-6 seconds on mobile. Next.js sites averaged under 2 seconds. That's not a rounding error. It's the difference between a site AI engines can efficiently crawl and index, and one they deprioritize.

Google has explicitly tied Core Web Vitals to ranking signals, and there's growing evidence that AI systems inherit those same biases. A slow site signals low investment, poor maintenance, and a worse user experience. WordPress's reliance on heavy themes, render-blocking plugins, and unoptimized PHP makes speed an uphill battle. Next.js ships fast by default — server-side rendering, automatic code splitting, image optimization, and edge caching are built into the framework, not bolted on as afterthoughts.

When AI engines evaluate which sites to cite, they're drawing from the same crawl data that penalizes slow pages in traditional search. WordPress sites carry that weight whether they know it or not.

4. Next.js Is Built for How AI Evaluates Sites

Next.js gives development teams capabilities that directly produce the kind of web presence AI engines favor: server-side rendering for fast, crawlable pages; API-driven content architectures that scale across channels; superior Core Web Vitals that improve user engagement; and the flexibility to build rich, interactive content experiences that keep visitors on-site longer.

WordPress, by contrast, fights against its own architecture to achieve these things. Plugin bloat tanks performance. PHP-rendered pages are slower to crawl. Theme constraints limit content flexibility. Every optimization requires another plugin, another workaround, another layer of technical debt.

Next.js sites don't just perform better — they create the conditions for building the kind of authoritative, high-performance web presence that AI engines learn to trust. It's the difference between a platform that enables growth and one that constrains it.

What This Means for Your GEO Strategy

If you're a health industry brand thinking about AI visibility, here's what our data suggests:

If you're on WordPress: Every month you stay on WordPress is a month your competitors on Next.js are pulling further ahead in AI visibility. A migration won't instantly generate citations — you still need brand authority and content depth — but staying on WordPress means fighting uphill with an architecture AI engines aren't built to favor. If you're serious about being discoverable in the AI era, the rebuild conversation needs to happen now, not next year.

If you're building something new: Next.js isn't just the better choice — it's the only serious choice. The health companies dominating AI citations are overwhelmingly on Next.js, and building on the same foundation positions you alongside those brands from day one. You'll attract better development talent, scale faster, and build the kind of performant, content-rich experience that AI engines reward. Starting on WordPress in 2026 is starting behind.

About This Study

This analysis is based on original research by Driftless AI, testing 99 health industry websites across 25 standardized queries on ChatGPT, Claude, and Google AI Overviews. Sites were classified by CMS platform, brand authority tier, and SEO/schema markup level. Complete data and methodology are available in the companion analysis spreadsheets.

Want to understand where your brand stands in AI search? Get in touch with Driftless AI for a 100% free GEO/AEO audit.