A shot at the funnel: What millions of GLP‑1 convos reveal about AI search
We analyzed millions of AI conversations about GLP‑1s at the start of 2026. The headline: every stage of the funnel now plays out inside AI chat interfaces—often in non‑linear, surprising ways. If you market in the GLP‑1 space (or any fast‑moving category), this creates new opportunities to measure what matters, spot gaps, and influence decisions earlier.
Below you’ll find what we learned about GLP‑1 behavior in AI, which signals to track, and how to act on them using Scrunch.
The short version: AI isn’t just replacing search—it’s reconceptualizing the journey. It’s a companion from first question to ongoing care.
AI funnel archetypes for GLP‑1s
We clustered conversations into five audience archetypes based on jobs‑to‑be‑done. Four of the five are similarly sized; “Access Seeker” is smaller but highly commercial.
Knowledge Seeker (~24% of conversations)
Who: Predominantly male, youngest segment
Behavior: 1–2 questions, then bounce; awareness with no declared intent
Prompts: “What’s the difference between GLP‑1 and GIP?”, “What is Tirzepatide and how does it work?”
Content plays: Beginner’s guide; myth‑busting; short explainer video; interactive quiz
Active Evaluator (~20%)
Who: Most likely to switch intent on a single answer
Behavior: Revisited across the journey; highest conversion moment
Prompts: “Mounjaro vs. Wegovy?”, “Are compounded GLP‑1s as effective?”
The full funnel lives inside LLMs. Awareness, consideration, evaluation, purchase, and retention all happen in‑chat at real volume.
The journey is non‑linear. Users bounce between stages as they check new symptoms, compare options, or revisit dosing—often with the same AI thread.
AI is a trusted partner, not a list of links. People use it to think through decisions, role‑play expert advice, and apply tight constraints to their realities.
“What users ask” isn’t always expected. Alongside brand and dosing, we saw lifestyle questions like “glow‑up stacks” and travel‑safe protocols—ripe for visibility wins if you address them directly.
The AI search metrics that matter (and how to use them)
Traditional SEO measured rankings and traffic. AI search (often called AEO) measures visibility and influence. Four signals now tell you if you’re winning:
Brand presence (aka share of answer)
What it is: How often your brand is mentioned in answers across tracked prompts and where it’s placed in the response.
Why it matters: Inclusion equals validation. In GLP‑1, being named in “best telehealth services” or “Mounjaro alternatives” drives familiarity and trust.
How to improve: Create answerable, structured content for priority prompts; align tone/format to how LLMs summarize.
Citations
What it is: Domains the AI cites beneath its answer.
Why it matters: Citations shape what models say and which brands they trust. For GLP‑1s, think clinical trials, regulatory pages, medical publishers, and reputable reviews.
How to improve: Publish original data and explainers, earn coverage on trusted third‑party sites, and ensure your medical pages are technically accessible.
LLM referral traffic
What it is: Clicks from AI engines (e.g., ChatGPT, Perplexity) to your site.
Why it matters: Lower volume, higher intent. Visitors are pre‑qualified by the model.
How to measure: Scrunch’s AI Referrals connects to GA4 to surface AI‑sourced sessions, top landing pages, and conversion rates. Most impact is still indirect (brand search/direct), but direct clicks are a strong signal. See how it works in this FAQ: Does Scrunch track AI referral traffic to my website?
Agent traffic
What it is: Visits from AI retrieval bots (GPTBot, ClaudeBot, PerplexityBot).
Why it matters: Early indicator that your content is being considered in answers. In healthcare categories, these visits often cluster on dosing, side‑effect, and access pages.
How to measure: Scrunch surfaces bot visits, agent diversity, and the pages they hit. Integrations with Akamai, Cloudflare, Vercel, and WordPress make this simple. Track AI traffic on your site with Scrunch
Analytics leaders: expect your funnel metrics to look different. Insights we shared with Lenovo at eTail London 2025 hold up here: bounce rates can rise on AI‑referred sessions (visitors arrive with high intent and specific expectations), and conversion rates can spike as users drop into your site mid‑journey.
How Scrunch helps you monitor and improve AI visibility
Monitoring & Insights
Multi‑platform tracking: Monitor brand presence, citations, and answer placement across supported AI platforms including ChatGPT, Claude, Gemini, Perplexity, Google AI Mode, Google AI Overviews, Meta AI, and Microsoft Copilot (Grok coming soon). See the full list in supported platforms.
Prompt intelligence that mirrors SEO thinking: Bring your keyword sets and right‑size them into prompts, topics, and personas. Helpful primer: From keywords to prompts.
Competitive benchmarking: Compare your share of answer and citation share against category peers to spot where you’re winning or invisible.
Content Gaps: Automatically find high‑value prompts where your audience is asking and you don’t yet have an answer. Prioritize by impact. See how Content Gaps works.
AI Referrals (GA4): Attribute AI‑sourced traffic and conversions; compare AI vs. non‑AI performance.
Agent traffic: See which bots crawl your site, how often, and which pages they hit.
AXP (Agent Experience Platform)
Deliver AI‑optimized content to visiting agents—a parallel, bot‑friendly experience that improves crawlability, context, and consistency while your human UX stays intact. Ideal for critical GLP‑1 pages (dosing, side effects, access) where accuracy and structure matter most.
Built for SEO and content pros
The UI uses familiar concepts—prompts, topics, tags, competitor sets, and shareable dashboards—so consultants and in‑house teams can get value quickly without retraining the whole stack. Deep‑dive explainer on the new KPIs: From SEO to AEO: new metrics that matter in AI search.
Credibility in the community
Scrunch shared the main stage with Lenovo’s Global SEO & Operations Lead at eTail London 2025, unpacking how LLMs are reshaping discovery and measurement. Key takeaways from that session—AI redefining the funnel, “bounce rates are back,” and SEO teams evolving into behavioral research hubs—are embedded in the platform’s approach. Our team regularly publishes practical guidance and data‑backed takes on AI search trends and tactics across the Scrunch blog.
A GLP‑1 playbook you can run this week
Set up measurement
Map 50–200 prompts across the five archetypes (awareness to regimen). Include obvious queries and “unexpected but consequential” prompts (e.g., symptom management, travel protocols).
Add top competitors and credible third‑parties you want to influence.
Connect GA4 in Scrunch’s AI Referrals to baseline AI‑sourced sessions and conversions.
Turn on agent traffic monitoring to log GPTBot/ClaudeBot/PerplexityBot activity.
Find and fix gaps
Use Content Gaps to identify high‑impact prompts where you’re absent.
Ensure AI can read your site: unblock AI user agents, add clean sitemaps, reduce JS‑dependent rendering, and structure content for summarization. Practical checklist: Page‑level monitoring and technical audits.
Influence citations
Publish original data/FAQs; earn mentions on trusted medical and review domains; keep regulatory pages clear and up to date. Learn how to track them: How to track citations in AI search.
Optimize the agent experience
Use AXP to serve structured, current facts to agents on critical pages (dosing, side effects, coverage), reducing misinterpretation and improving answer quality.
Measure outcomes and iterate
Watch share of answer and citation share versus competitors.
Treat agent visits as demand signals; pair with AI Referrals to see which prompts and pages convert.
Interpret bounce with context: higher intent means visitors may skim for a specific proof point and leave satisfied—or convert quickly.
We scored conversations with a behavior‑first taxonomy (user state, task, assistance mode, personal context, decision proximity) using structured LLM outputs, then clustered one‑hot encoded labels (k‑means, k=5) validated by silhouette and ARI on held‑out splits. Demographics were joined and weighted to AI‑active populations for projection. Access Seeker estimates are directional due to a smaller base.