AI SEARCH · PATIENT COHORTS

Do ChatGPT-Referred Telehealth Patients Convert and Retain Better Than Google Patients?

David TerrellFounder, ClinicAdsNovember 18, 20266 min read

ChatGPT-referred telehealth patients may convert better than Google patients, but no telehealth study shows it, and nothing published measures whether they retain better. Cross-industry data puts the AI conversion edge anywhere from 1.26x to 4.4x organic search per visit. None of it tracks a paying subscription patient past the first month.

AEO, GEO, AI visibility, and AI SEO are four names for the same discipline: getting a brand named and cited when someone asks ChatGPT, Perplexity, Gemini, Copilot, or Google AI Overviews for a recommendation. This post uses AEO. The [plastic surgery version of this question](/blog/chatgpt-referred-patients-vs-google-search-conversion-plastic-surgery) judges a one-time consult-to-surgery decision, and the [med spa version](/blog/chatgpt-referred-clients-vs-google-search-conversion-med-spa) judges show rates and memberships. A DTC telehealth brand is judged on two numbers instead: how many completed intakes become paying patients, and how many of those are still paying at month 3. The studies here are a different set from both of those posts.

KEY TAKEAWAYS
  • No independent study found while writing this post compares ChatGPT-referred telehealth patients with Google search patients on intake-to-paid conversion or month-3 retention.
  • Cross-industry estimates of the AI conversion edge range from 1.26x organic search (Siege Media's 2026 median across 78 sites) to 4.4x (Semrush's estimate of AI search visitor value), and none of the datasets covers healthcare subscriptions.
  • Retention varies more by treatment line and cohort year than by any channel effect measured so far: Prime Therapeutics reported 31.5 percent one-year GLP-1 persistence for weight loss in its three-year claims study, while later tirzepatide starters reached about 62 to 64 percent.
  • A telehealth brand should judge AI-referred patients on payback, which needs both intake-to-paid rate and month-3 retention, compared within the same treatment line.

What does the evidence say about AI-referred conversion?

The evidence says AI-referred visitors usually convert at least as well as organic search visitors, but estimates of the gap differ by a factor of three or more. Semrush, Siege Media, and Seer Interactive each measured a different set of sites with a different method, and none of the three measured a healthcare brand or a subscription that renews monthly.

Semrush estimated that the average AI search visitor is worth 4.4 times the average traditional organic search visitor, based on conversion rate, in a study of more than 500 digital marketing and SEO topics. Siege Media screened 78 websites with usable conversion tracking in 2026 and found a median AI-to-organic conversion ratio of 1.26x, with AI matching or beating organic on 72 percent of sites and ecommerce closest to parity at 1.01x. Seer Interactive published a single-client case study in which ChatGPT traffic converted at 15.9 percent against 1.76 percent for Google organic.

Published AI-referral conversion findings as reported by each source, not independently verified by ClinicAds. None covers telehealth.
SourceWhat was measuredReported resultMain limit for telehealth
Semrush, 2025Visitor value across 500+ marketing and SEO topicsAI visitor worth 4.4x an organic visitorVendor benchmark, B2B-heavy topics
Siege Media, 2026GA4 key-event rates on 78 sitesMedian 1.26x organic; 72% of sites at or above paritySmall sample, single conversion event
Seer Interactive case studyOne client's ChatGPT vs Google organic traffic15.9% vs 1.76% conversionOne site, method not fully published

Is there any data on AI-referred telehealth retention?

There is no published data on whether AI-referred telehealth patients retain better or worse than patients from Google search. Every study above stops at the first conversion event. For a subscription telehealth brand, the first payment is the start of the measurement, because CAC is recovered over the months a patient keeps paying.

The retention baselines that do exist vary widely before any channel effect. Prime Therapeutics, a pharmacy benefit manager, reported in a three-year claims analysis that GLP-1 persistence for weight loss was 31.5 percent at one year, 15.1 percent at two years, and 8.1 percent at three. A later Prime analysis reported one-year tirzepatide persistence of about 64 percent for 2023 starters and 62 percent for 2024 starters. Those are insured-member figures, not cash-pay telehealth figures, but they show that drug, cohort year, and coverage move retention far more than a channel comparison is likely to.

Why could AI-referred patients convert or churn differently?

AI-referred telehealth patients could convert better because many have already asked ChatGPT about eligibility, monthly price, and whether a provider is legitimate before they reach the intake form. A Google searcher comparing three hair loss or TRT brands often does that research during intake, and drops off when the price or the clinician questionnaire appears.

The same patients could churn faster when the answer that sent them was wrong. AI answers can quote a monthly price the brand changed months ago, list a medication the brand no longer prescribes, or omit that a clinician must approve the prescription. A patient who pays on those expectations is more likely to cancel after the first refill. Which effect is larger is unmeasured, and it likely differs by treatment line.

Which matters more for payback: conversion or month-3 retention?

Month-3 retention usually matters more for payback than intake-to-paid conversion, because a telehealth brand recovers CAC from refills rather than from the first payment. A source that converts more intakes but loses half of those patients by month 3 can leave the brand with the same number of paying patients as a slower-converting source that keeps them.

The table shows two hypothetical sources per 100 completed intakes. Source A converts more intakes; Source B keeps more patients. By month 3 both have about 15 patients still paying, so a conversion-rate comparison alone would rank them wrongly. The figures are illustrative, not measured rates for AI or Google traffic.

Illustrative cohort per 100 completed intakes. Not measured data.
SourceIntake-to-paid ratePaying patientsMonth-3 retentionStill paying at month 3
Source A30%3050%15
Source B22%2270%About 15

How should a telehealth brand compare AI and Google cohorts?

A telehealth brand should record each patient's source at intake, store it in the intake or EHR system rather than an ad pixel, and compare cohorts within one treatment line after at least 90 days. Mixing GLP-1 patients from one source with hair loss patients from another produces a retention gap that reflects the drug, not the channel.

Referrer data undercounts AI-referred patients. Many people read a brand name in ChatGPT, then search the name or type the URL, so analytics records branded search or direct. ClinicAds has no published telehealth cohort comparing AI and Google patients, and no client data appears in this post. The steps below are the setup ClinicAds recommends, with source data kept server-side so no health information reaches an ad platform.

  • 1. Add an intake question on how the patient heard about the brand, with ChatGPT or another AI assistant as a listed answer
  • 2. Tag chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com referrals server-side and attach them to the intake record
  • 3. Track five steps per source: intake started, intake completed, clinician approved, first payment, still active at month 3
  • 4. Compare sources only within the same treatment line, such as GLP-1, TRT, or hair loss
  • 5. Wait for at least 50 paying patients per source in a treatment line before reading a retention difference

Should a telehealth brand move budget toward AEO on this evidence?

A telehealth brand should not move paid budget toward AEO on conversion evidence alone, because AI referrals are a small share of traffic and the retention side is unmeasured. AEO earns budget when the brand's own cohorts show AI-referred patients reaching payback at least as fast as paid patients, at a volume large enough to matter.

ClinicAds plans telehealth paid media at $10,000 to $50,000 a month against a CAC of $90 to $250 per paying patient, patient LTV of $600 to $2,000, payback under 3 months, and an LTV-to-CAC ratio of 3:1 or better. Those are agency averages, not guarantees, and they are the benchmark an AI-referred cohort has to beat. The [telehealth growth program](/telehealth) runs AEO inside the same CAC and payback reporting, not as a separate service.

  • Expand AEO work when AI-referred patients match paid month-3 retention and reach at least 5 percent of new paying patients in a treatment line
  • Hold steady when AI-referred patients convert better at intake but retain worse, and fix the pricing and eligibility facts AI answers repeat
  • Keep paid budgets on their own CAC and payback whenever AI-referred volume is too small to read
FREQUENTLY ASKED

Do ChatGPT-referred telehealth patients stay subscribed longer?

Nobody has published that comparison. Existing AI-referral studies stop at the first conversion. A telehealth brand has to measure month-3 retention by source in its own intake and billing data.

How much better do AI-referred visitors convert than organic search visitors?

Published estimates range from about 1.26x (Siege Media's 2026 median across 78 sites) to 4.4x (Semrush's visitor-value estimate). No estimate covers healthcare or telehealth subscriptions.

Can a telehealth brand track AI referrals without breaking HIPAA?

Yes, if source data stays in the brand's own intake or EHR system and is not sent to ad platforms with health information attached. A self-reported intake question and server-side referrer tagging both work without a third-party pixel.

Find out whether AI-referred patients pay back faster than paid ones

ClinicAds will review your intake form, source tracking, and cohort reporting, then set up a HIPAA-safe comparison of AI-referred, Google, and Meta patients on intake-to-paid rate and month-3 retention.