How Many Telehealth Patients Research in ChatGPT Instead of Google?

Do telehealth patients research in ChatGPT instead of Google?

No independent survey found while writing this post measures telehealth patients who research in ChatGPT instead of Google. The evidence supports a narrower claim: roughly a third of consumers now ask AI chatbots about health, and a telehealth patient is a consumer asking health questions. Treat AI as a second front door beside Google, not a replacement for it.

That distinction drives budget. A brand that reads the headline numbers as proof Google is finished would cut the search and paid work that still produces most intake starts. A brand that reads them as proof AI does not matter would miss the comparison stage, where patients pick which two or three providers to price.

How many consumers use AI chatbots for health information?

Rock Health's consumer adoption survey, fielded in December 2025 and reported in March 2026 by Health Populi, found 32 percent of consumers use AI chatbots for health information, up from 16 percent in 2024. Among those seekers, ChatGPT was the most used brand at 23 percent, followed by Gemini at 15 percent.

The post reporting these figures does not state the sample size, and its headline cites a different 36 percent figure for a broader AI measure, so a brand should quote the 32 percent number with its source and check Rock Health's own report before using it in client-facing material.

• ·32 percent of consumers used AI chatbots for health information, up from 16 percent in 2024
• ·ChatGPT was the most used chatbot brand among AI health seekers, at 23 percent
• ·The same report says 84 percent of AI health users had used an app or virtual care program in the past year, against 42 percent of non-AI users
• ·These figures cover health information generally, not choosing a telehealth provider

Why does the virtual care overlap matter for telehealth?

The overlap between AI health use and virtual care use is the closest signal to a telehealth audience. Rock Health data, as reported by Health Populi, shows 84 percent of AI health users had used an app or virtual care program in the past year, double the 42 percent of non-AI users. People who ask chatbots about health are already comfortable with care delivered online.

It is a correlation from a survey of consumers, not a measure of providers chosen inside ChatGPT. It does suggest the people a telehealth brand most wants to reach are over-represented among AI users, which raises the cost of being absent from the answers they read.

How many U.S. adults use ChatGPT at all?

Pew Research Center's 2026 survey of 5,119 U.S. adults, fielded February 17 to 23, found 44 percent use ChatGPT, up from 34 percent in 2025. Pew's series runs 18 percent in 2023, 23 percent in 2024, 34 percent in 2025, and 44 percent in 2026. ChatGPT led every other chatbot, with Gemini at 24 percent.

Use is heavily age-skewed: 61 percent of adults 18 to 29, 55 percent of ages 30 to 49, 37 percent of ages 50 to 64, and 19 percent of those 65 and older. Pew changed the 2026 question to cover other chatbots too, so the jump from 34 to 44 percent is not perfectly like for like. Most telehealth categories skew toward the under-50 groups where use is now a majority.

What does OpenAI say about health questions?

OpenAI reported at the January 2026 launch of ChatGPT Health that more than 230 million people worldwide ask health or wellness questions on ChatGPT every week, and a separate OpenAI report put daily health askers above 40 million. These are company-reported usage figures, not audited or survey-based, and they count questions, not provider decisions.

A health question is a long way from a purchase. A user asking how a medication works is not the same as a user picking a brand to start intake. The number shows the channel carries health demand at scale; it does not say how much of that demand reaches the brand-comparison stage.

Why do the adoption numbers differ so much?

The numbers differ because each source defines the behavior, the population, and the window differently. Rock Health counts consumers who use chatbots for health information. Pew counts any adult who uses ChatGPT. OpenAI counts questions asked worldwide. None is the share of a telehealth brand's patients who used ChatGPT to choose it.

A brand should never quote one national number as the share of its own patients. The defensible statement is a range with each end tied to its definition, such as about a third of consumers using chatbots for health in Rock Health's data and 44 percent of adults using ChatGPT in Pew's, plus a figure measured from the brand's own intake form.

What should a telehealth brand measure and do about it?

A telehealth brand should measure its own position in AI answers and its own patient sources before moving budget. National data says the channel is large enough to test. A monthly prompt log says whether the brand is named, and an intake question says whether it matters for acquisition.

This sits inside the [telehealth marketing program](/telehealth) alongside paid media and speed-to-lead, not as a separate service. ClinicAds plans telehealth paid media at a $90 to $250 CAC per paying patient with payback under 3 months, agency averages, not guarantees. Paid keeps running, because nothing in the data supports dropping it.

• ·Run 10 condition-plus-treatment prompts monthly, such as best online provider for hair loss, and log every brand named
• ·Publish price, inclusions, clinician licensure, and state coverage on crawlable pages that AI models can quote
• ·Add an intake question asking whether an AI assistant helped the patient choose
• ·Track month-3 retention and CAC for AI-assisted patients separately for 6 months