On a recent Saturday morning, a typical American consumer we’ll call Maya had a familiar problem. She had a vacation flight to book, a household budget to organize and a nagging sense she should “eat healthier.” A year ago, Maya would have opened a Google browser on her desktop, bounced between search results and bookmarked a few articles she’d never return to. This time, she tapped the ChatGPT icon on her phone and asked for guidance. The artificial intelligence app remembered her last trip, suggested an itinerary and drafted a message she could send to a hotel. Later, it pulled her grocery shopping list and turned it into a meal plan.
Most U.S. adults have now tried consumer AI tools for help with routine, everyday tasks. That means the conversation has now shifted to which AI models they’ll keep using. Some tools are becoming go-to utilities, opened by default the way people once pointed their mouse cursor to a search bar. For a growing number of users, AI is replacing search engines as their first stop for getting things done.
New PYMNTS Intelligence data shows that consumer AI usage has crossed the “try it” stage. Adoption hit 54% of adults in January, up five points month over month. Mainstream users, defined as those using AI for a smaller number of mostly low-complexity tasks and to complement their typical search methods, now represent just over one in three consumers. ChatGPT is the dominant entry point, but consumers are exploring other alternatives. Claude’s personal-use base grew 22% in a single month, and Gemini rose 8%. As people use AI more often and for more involved tasks, AI app use is rising, while browser use is falling.
These are just some of the findings detailed in this edition of the Agentic AI Report series, “AI Becomes a Daily Habit: The Consumer Shift From Trying Tools to Living With Them,” a PYMNTS Intelligence exclusive report. This edition examines how consumers use AI and draws on insights from a survey of 3,497 U.S. adult consumers conducted from Jan. 6, 2026, to Jan. 30, 2026.
Over the Halfway Mark
Adoption of consumer AI tools hit 54% of U.S. adults in January, up five percentage points month over month.
More than one in two U.S. adults used AI for personal reasons in January, up from 49% last December. Mainstream users now represent 34% of consumers. But the more important story is who is showing up and how their behavior is changing.1
First, the growth did not come from one age group. It came from every generation, each moving in the same direction at the same time. That matters for banks, payment firms and digital commerce players because it suggests that AI is not stuck as a niche tool for early adopters. Instead, it’s becoming part of everyday problem-solving across all life stages.
Second, the “mainstream” segment is expanding fast. Mainstream users now constitute 34% of adults, and Generation Z, the oldest of whom is now roughly 29, is leading that expansion. Among Gen Z consumers, 47% now qualify as mainstream users. Their share of light users fell from 15% to 9%, a sign that many are moving from dabbling with AI to using it for more involved, repeatable tasks. 2
Third, which large language model (LLM) tool a consumer uses is fluid. ChatGPT remains the dominant entry point, but consumers are exploring alternatives as the technology rapidly evolves. Claude’s personal-use base grew 22% in a single month, and Gemini grew 8%. To note, this report tracks AI as used for personal tasks (shopping, finances, health, travel and similar use cases), not workplace adoption.
ChatGPT dominates consumer AI use, but the field is crowded.
Right now, ChatGPT sits in a league of its own. With 83.2% of AI users having tried it at least once, it’s essentially the default entry point into AI for most consumers, matching December 2025’s level and suggesting that its reach is, at least for now, holding steady.
Google Gemini comes in second with roughly 47.8%–51.7% of users employing it, a meaningful gap behind ChatGPT but still a strong showing given Google’s ability to push Gemini through its existing products like Search and Gmail. Microsoft Copilot and Meta AI cluster just below that at roughly 29%–31%, likely buoyed by their deep integration into Office 365 and WhatsApp/Facebook rather than standalone appeal.
The middle tier—Perplexity, Grok and DeepSeek—lands in the 9%–15% range for usage. Perplexity’s search-focused approach has found a niche, while Grok and DeepSeek benefited from significant media attention (Grok from Elon Musk’s X platform, DeepSeek from its viral moment in January 2025, when it released its R1 reasoning model demonstrating performance comparable to OpenAI’s most advanced offerings at a fraction of the cost). That DeepSeek managed 9.3%–11.7% so quickly after launching is notable.
Consumer use of Claude sits at 10.1%–12.3%, just ahead of DeepSeek, which reflects its stronger foothold among professional users and developers rather than mainstream consumers. Brave Leo and Poe trail at the bottom, each with less than 3.5% of consumers using them, suggesting they remain niche tools.
The big takeaway is that most AI users have tried more than one platform, but ChatGPT remains the undisputed starting point. Everyone else is competing for second place.
Browser Woes
ChatGPT’s largest single access point is its smartphone app, used by 38% of its users. Google Gemini follows a similar pattern, at 36%.
There is a meaningful difference between a consumer who opens a browser, navigates to an AI platform and types a question, and one who taps an app icon they installed weeks ago, picking up a conversation where they left off. Both are using AI. But one of them has made a commitment.
What we know about habit formation suggests that the shorter the distance between a cue and a behavior, the more likely that behavior is to become automatic. Apps maintain conversation history in a way that creates a self-reinforcing loop far more naturally than a browser session does.
ChatGPT and Gemini look like mobile-native products. ChatGPT’s largest single access point is its smartphone app, used by 38% of its users. Gemini follows a similar pattern at 36%. These are not platforms people visit occasionally out of curiosity. Instead, they behave like utilities, the kind of tools consumers reach for reflexively, just as they check a weather app or send a message. That reflexive quality is exactly what habit looks like in practice.
Perplexity and Claude show a more balanced picture. Claude splits almost evenly across smartphone app, smartphone browser, desktop app and desktop browser. Each channel captures roughly one-quarter of usage. That balance can signal flexible, deliberate usage across different contexts. But it can also signal that the platform hasn’t yet become the default for any given moment in a user’s day. Having no clear home can mean being everyone’s second choice.
Copilot tells a different story. Sixty percent of its usage happens on desktop or laptop. Just 23% of users access it through a smartphone app. Even in personal use contexts, Copilot behaves more like a one-off productivity tool than a daily companion.
Downloading an AI app is the gateway to frequent consumer usage, but not to app dominance.
The platforms becoming daily utilities are the ones living on a consumer’s home screen. And crucially, installing an app is itself a behavioral signal. It’s a small but deliberate act of commitment that primes users to return.
As users move from light experimentation to high-intensity and high-complexity engagement with AI models, browser reliance tends to decline while app and mobile usage rise. Across most platforms, more than six in 10 power users primarily accessed AI via a smartphone or tablet (ChatGPT: 69%; Gemini: 62%).
Significantly, this contradicts the common assumption that mobile apps are primarily for lightweight tasks. In several cases, the most sophisticated users are more likely to operate inside apps, particularly those on smartphones. The implication is that the app environment may signal more than convenience. It may reflect trust, workflow continuity and habitual integration.
Millennials Emerge as Power Users
Millennials use more consumer AI models—nearly three—than any other generational cohort.
Even at the demographic level, consumers are trying, comparing and switching across assistants rather than locking into just one.
Generationally, the center of gravity sits with adults still in their working or earning careers. Millennials and bridge millennials use the most models (on average 2.8 and 2.9, respectively), slightly ahead of Gen Z (2.5) and well ahead of baby boomers and seniors (1.8). This is notable because it challenges the assumption that the youngest users are driving the most diversified AI behavior. Instead, prime working-age consumers appear to be integrating AI into more aspects of daily life, likely spanning productivity, financial management, planning and personal organization. Older consumers participate, but with narrower experimentation.
The sharpest divide, however, is behavioral, not demographic. Power users average 3.7 models, compared to just 1.4 among light users. Power users are building AI stacks, switching between tools depending on the task and embedding AI more deeply into their daily routines. Light users, by contrast, remain closer to single-tool or occasional engagement. Multi-model usage is therefore a defining marker of intensity, not just access.
Finally, depth of usage scales with behavioral replacement. Consumers who say AI has fully or mostly replaced old methods use 3.0 models on average, compared to 2.1 among those who only lightly complement traditional approaches. Replacement appears to be a tipping point. Once AI displaces legacy behaviors, consumers expand their toolkit rather than consolidate it.
Taken together, the story is clear: Multi-model AI behavior is now mainstream, but true diversification in which model is used is concentrated among power users, millennials and consumers who have meaningfully replaced their older methods. The future expansion of AI ecosystems will depend less on basic adoption and more on whether mainstream users move toward deeper replacement and multi-tool engagement.
Read More
PYMNTS Intelligence is the leading provider of information on the trends driving consumer use of AI in their daily lives. To stay up to date, subscribe to our newsletters and read our in-depth reports.
Methodology
“AI Becomes a Daily Habit: The Consumer Shift From Trying Tools to Living With Them” is based on a survey of 3,497 U.S. adult consumers conducted from Jan. 6, 2026, to Jan. 30, 2026. The report explains how consumers are using LLMs for personal tasks.
Addendum
This research uses a consumer task framework spanning 54 activities across nine broad categories (including shopping, finances, health, education, travel and more). Each activity is weighted based on three simple ideas:
- How much AI needs to reason and be accurate
- What could go wrong for the consumer if the AI gives inaccurate outputs
- How sensitive the consumer’s personal data is
The task list is also grouped by complexity and risk. At the low end are tasks like drafting messages, making lists and finding shopping deals. In the middle are travel planning and product research. At the high end are tasks that can affect health decisions and financial well-being, such as understanding credit reports and loan terms, reviewing bank and card fees, comparing financial products or building a household budget.
This framework matters because it lets us track not just whether people use AI but whether they are moving into repeatable, higher-stakes routines.
1. PYMNTS Intelligence defines power users as using AI to perform 27 or more distinct tasks each month. These tasks include higher-complexity ones such as personal investment management. Mainstream users perform an average of eight tasks per month, most low complexity, like comparison shopping. Light consumers perform an average of two tasks a month.↩
2. PYMNTS Intelligence uses the following approximate birth dates and approximate age ranges in 2026 for generational cohorts: baby boomers: born in 1964 or earlier and now aged 62 or older; Generation X: born between 1965 and 1980 and now aged 46–61; millennials: born between 1981 and 1996 and now aged 30–45; bridge millennials: born between 1978 and 1988 and now aged 38–48; zillennials: born between 1991 and 1999 and now aged 26–35; and Generation Z: born in 1997 or later and now aged 29 or younger.↩
