Usage patterns show that most people encounter conversational AI through a single dominant platform and a shrinking set of mobile assistants shaped by operating system defaults, according to new data from PYMNTS Intelligence.
The findings draw on surveys of consumers in the United States who reported using either dedicated AI platforms or AI assistants embedded in smartphones. The data measures whether respondents used each tool at least once, capturing reach and early habit formation rather than frequency or depth of engagement. Even at that level, usage patterns point to a market that is already becoming directional.
One Platform Accounts for Most First Encounters
Among consumers who used a dedicated AI platform for at least one task, 83% reported having used OpenAI’s ChatGPT. By comparison, 48% reported using Google’s Gemini, and 30% reported using Microsoft’s Copilot. This portion of the survey included 840 respondents.
The gap matters because it reflects where usage begins. ChatGPT reaches most consumers at the point of first interaction, while Gemini and Copilot reach smaller subsets. Rather than trying several comparable tools, most users appear to start with one platform and stop there.
Because the data measures whether a consumer has used a platform at least once, it reflects exposure rather than sustained engagement. Still, exposure at this scale shapes behavior. In consumer software, the platform that introduces users to a category often becomes the reference point for how the category works.
Advertisement: Scroll to Continue
The data suggests that conversational AI is already following that pattern. ChatGPT’s reach exceeds Gemini’s by 35 percentage points and Copilot’s by 53 points, indicating that trial itself is concentrated. Users are not dividing their attention evenly. They encounter AI through a single primary interface, with others remaining peripheral.
That usage pattern reduces the likelihood that consumers will actively compare tools unless a clear functional gap emerges. It also raises the bar for secondary platforms, which must persuade users to leave an interface they already recognize.
Mobile AI Usage
Usage patterns shift on smartphones, where AI assistants are embedded at the system level. Among respondents who reported using a built-in AI assistant for at least one task, 40% used ChatGPT, 40% used Google Gemini, and 37% used Google Assistant.
The narrow spread between these figures does not reflect equal consumer preference. It reflects distribution. On mobile devices, users often interact with whichever assistant the operating system surfaces by default.
Google Assistant’s continued reach illustrates that effect. Its usage reflects years of default placement on Android devices rather than renewed consumer adoption. Many users encounter it because it is present, not because they sought it out.
That distinction becomes more important as Google prepares to retire Google Assistant and replace it with Gemini this year. The overlap between Assistant and Gemini usage suggests that much of Gemini’s future mobile usage may come from migration rather than discovery.
As Gemini inherits Assistant’s position in the interface, usage is likely to rise without requiring users to change behavior. The assistant changes, but the habit remains.
Apple faces a different usage challenge. While the iPhone remains one of the most widely distributed consumer computing platforms, Apple has yet to translate that distribution advantage into meaningful conversational AI usage. Siri has not emerged as a primary AI entry point, highlighting how default placement alone is no longer sufficient if capability lags expectations.
Apple has sought external help, including discussions with Google, as it works to deliver a long-awaited AI upgrade to Siri.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
