AG1 Spends Millions on Marketing. It Scored 0 out of 100 on AI Agent Readiness.

Health & Wellness ranked fifth out of 10 industries in our State of Agent Readiness 2026 report, with an average score of 37.3 out of 100.

Health and wellness is a trust-driven category. Consumers want to know what's in a product, whether it's third-party tested, what the dosage is, and what real users have experienced. That's exactly the kind of detailed, structured information AI agents need to make recommendations. And yet: AG1 scored 0 out of 100. One of the most-marketed supplements on the internet is completely invisible to AI shopping agents.

We used the Pacestack Agent Readiness Scanner to audit the top 25 health & wellness websites across 20+ signals, from Schema.org markup to MCP server readiness.

The Numbers

• Average score: 37.3/100 (Grade D)
• Highest scorer: Noom at 63/100
• Lowest scorer: AG1 at 0/100
• 44% of brands scored below 50, meaning AI agents will struggle to recommend their products

Grade Distribution

Only Noom broke into B territory. More than half of brands sit in C, and a third are in D, meaning agents can read fragments of their sites but miss most of the product information that matters.

The Top 5

Noom leads the category, likely benefiting from a content-heavy, structured approach to its product pages.Whoop and Ro both perform relatively well, possibly because their products require detailed specification data (biometric tracking, treatment protocols) that lends itself to structured markup.

The Bottom 5

AG1 (Athletic Greens) spends enormous amounts on influencer marketing and podcast advertising. That spend drives human awareness, but AI agents don't listen to podcasts. They read structured data. And AG1 provides effectively none. Levels, a CGM-based metabolic health company, is in the same position.

What's Going Wrong

The three most common failures across health & wellness brands:

1. Product Schema (92% failed): Without Product markup, agents can't identify what you sell, what it costs, or whether it's available. For supplements and health products where ingredient details and certifications matter, this is a critical gap.
2. Product Schema Completeness (92% failed): Even the few brands with some markup are missing fields like ingredient lists, dosage information, certifications, and review data that agents need to make informed recommendations.
3. Breadcrumb Navigation (88% failed): Without breadcrumb markup, agents can't navigate from "supplements" to "protein powder" to "plant-based protein under $40." Category structure matters for discovery.

Category Breakdown

Structured Data (avg: 8.1/30). Below the overall average. Health brands tend to have rich product information (ingredients, dosages, certifications) but almost none of it is structured in a way agents can parse.

Agent Accessibility (avg: 19.1/25). Strong. Most health brands allow AI crawlers. The door is open, but the store shelves aren't labeled.

MCP Readiness (avg: 1.5/20). Nearly zero. No real MCP implementations. Health and wellness is a category where real-time data matters (subscription availability, personalized dosing), making this gap especially costly.

What This Means

Health and wellness is one of the fastest-growing categories in e-commerce, and AI agents are well-positioned to accelerate that growth. Consumers increasingly ask AI for personalized recommendations: "What supplements should I take for joint health?", "Best fitness tracker for sleep monitoring?", "Which protein powder has the fewest additives?"

The brands that provide structured, machine-readable answers to these questions will be the ones agents recommend. The brands relying purely on influencer marketing and podcast ads will find their reach doesn't extend to AI-mediated shopping.

McKinsey's research shows that AI-generated product recommendations convert at 4.4x the rate of traditional search. For health products, where trust and specificity drive purchases, that multiplier could be even higher. But only for brands whose data is actually readable.

← Back to the full State of Agent Readiness 2026 report

Methodology: All scans were performed on 2026-02-12 using the Pacestack Agent Readiness Scanner, which evaluates 20+ signals across Structured Data, Agent Accessibility, MCP Readiness, and AI Perception. Learn more about our methodology.