At 9:15 on a Monday morning, a new generative artificial intelligence user asks a large language model (LLM) to polish an email, reads the answer twice and checks a search engine before sending it. A few desks away, a colleague with more than one year of gen AI experience has a dedicated AI platform open next to two other models. That second user summarizes a meeting, pressure-tests a budget and organizes a project plan. The first user is testing a feature. The second has built a workflow.
That contrast frames the ninth Consumer AI Benchmark Report from PYMNTS Intelligence. The data shows that 61% of workplace gen AI users have used the technology for at least one year, while just 12% started within the past six months. Personal use has a younger mix, but 51% of those users also have at least one year of experience.
The central finding is that at work, time with gen AI changes both tool and task choice. Longer-tenured users move toward dedicated platforms, add specialized tools, use AI for more activities and rely on it for harder decisions. Growth is increasingly tied to retention, user skill and workflow depth, while personal use still offers room to recruit and shape newer users.
Five things to know
- Most users are experienced. Roughly 61% of gen AI workplace users have used the technology for at least one year, and only 12% are newcomers. Personal use has about 25% more newcomers than workplace use, at 15% versus 12%.
- Heavy use develops over time. Three-quarters of power users have used gen AI for at least one year, compared with 42% of light users and 50% of mainstream users.
- Tenure expands the job list. Personal users with at least one year of experience complete an average of 11 tasks with gen AI, 68% more than the 6.6 tasks completed by newcomers.
- Experienced users bring AI into harder decisions. The share who consider gen AI essential for managing finances and banking is 129% higher among those who have used the technology for one year or more than among newcomers.
- Time produces a wider tool stack. Experienced personal users are 158% more likely than newcomers to use Claude, while power users average 2.9 platforms, more than twice the 1.4 used by light users.
The gen AI on-ramp has matured
Workplace use predates personal use. Sixty-one percent of workplace users have at least one year of experience, including 26% with two years or more. The comparable shares for personal use are 51% and 20%. Newcomers account for 12% of workplace users and 15% of personal users.
That gap describes the current mix in each setting. It doesn’t prove that consumers generally tried gen AI at work before using it personally. The firm conclusion for now is that workplace use is more established, while personal use continues to account for a larger share of new users.
Tenure also overlaps with usage intensity. PYMNTS Intelligence uses a proprietary scoring system to measure how deeply a consumer uses AI across activities to create a scale that divides AI users into power users, mainstream users, light users and holdouts. We take every AI task performed in a given month and add up points assigned to each (1 point for simple, low-risk tasks like drafting an email, up to 3 points for complex, higher-risk ones like investment advice or sharing card details). A higher total means more frequent use of more demanding tasks. Power users have historically made up roughly 9%–12% of consumers, mainstream users around 27%–34% and light users about 10%–13%, with holdouts making up the remainder.
Seventy-five percent of power users have at least a year of experience. That falls to 50% for mainstream users and 42% for light users. The measures are related but distinct: Tenure indicates how long someone has used gen AI, while usage personas indicate how broadly and deeply they use it. Heavy use usually takes time to develop.
The longer trend supports that reading. From September 2025 through May 2026, personal-only use became the largest active group, reaching about 37% of consumers in May. Use across both work and personal settings reached roughly one-quarter, while work-only use remained in the single digits. Expansion is occurring mainly through personal and dual use.
Experience makes gen AI use deeper, broader and more selective
The clearest change in tool choice is the move toward a dedicated AI platform. Among personal users with less than six months of experience, 22% name a dedicated platform as their most helpful AI tool. That rises to 35% among users with at least a year of experience. Reliance on a mobile phone assistant falls from 20% among newcomers to 11% among longer-tenured users.
The workplace shows a similar pattern. Longer-tenured work users are about 24% more likely than newcomers to identify a dedicated platform as most helpful and about 46% less likely to choose an AI search summary. Familiarity moves users away from the nearest AI feature and toward the tool they deliberately open for the job.
Search engines feel the largest substitution effect. Twenty-six percent of personal newcomers say they use search engines less because of gen AI. That share rises to 38% for users with at least a year of experience, a 47% increase relative to the previous share. Pullbacks from specialized sites and shopping sites are smaller, while social media use is almost flat across tenure groups.
Experienced users also do much more. Personal newcomers average 6.6 gen AI tasks. The average jumps to 11 for users with at least a year of experience. Replacement of older routines rises more gradually, from 33% of newcomers who say gen AI has fully or mostly replaced their previous method to 38% of one-year-plus users.
Tenure changes the AI task mix and platform stack
New users tend to begin with familiar, low-friction tasks. Among those with less than six months of experience, 47% have used gen AI for writing and 39% for discovering what to buy. Their use is lower for learning and travel (both at 26%) and for finances (at 23%).
After a year, the mix fills out. Writing rises to 59%, product discovery to 52%, everyday planning to 52%, learning to 50%, health and wellness to 49%, travel to 46%, entertainment to 46% and finances to 39%. Experience adds volume and narrows the gap between simple tasks and those that require more context, judgment or sensitive information.
The strongest evidence comes from tasks users call essential, meaning they could not complete them without gen AI or would find them much harder to complete. Among one-year-plus users, 31% say gen AI is essential for managing finances and banking, compared with 13% of newcomers. The experienced group is also more likely to call it essential for learning (35% versus 26%) and for health and wellness (32% versus 23%).
Product discovery reveals a gap. Experienced users are more likely to use gen AI to find what to buy, but only 20% say it’s essential, compared with 28% of newcomers. The internal review interpreted this as a performance issue. Seasoned users keep trying AI for shopping, yet many do not find the experience strong enough to rely on. Merchants and commerce platforms have a product challenge around comparison quality, product data, ranking transparency and the handoff from recommendation to purchase.
The platform mix widens at the same time. ChatGPT use is high across all tenure levels, reaching 74% among one-year-plus users. The growth appears in the second and third tools. Among experienced personal users, 52% use Gemini, 28% use Microsoft Copilot, 16% use Claude and 12% use Grok. Newcomer shares are 45%, 18%, 6% and 4%, respectively. Work users show the same broadening, with one-year-plus users 48% more likely to use Gemini and 61% more likely to use Claude than newcomers.
Implications and opportunities
Providers should compete for depth. Most workplace users already have meaningful experience, so growth will come from getting existing users to complete more tasks, use the tool more often and bring it into higher-value workflows.
Personal use remains the younger frontier. Its larger newcomer share gives platforms, banks, FinTechs and merchants an opening to shape habits through strong onboarding, useful templates and clear demonstrations of value.
The “most helpful” position is the strategic prize. Experienced users increasingly choose a dedicated platform, but they also assemble wider stacks. Winning requires becoming the default tool for a defined set of jobs, rather than assuming a single platform will own every task.
Financial services providers have a specific opening. Reliance on gen AI for finances rises sharply with tenure, increasing the value of trustworthy data, permission controls, explainable outputs and smooth escalation to a human. Commerce providers face a different signal in product discovery: usage is broad, but perceived necessity is weak. Better product data and a cleaner path from research to transaction could convert repeated trials into dependable use.
Conclusion
Consumer gen AI has entered a more mature stage. Most workplace users have a year or more of experience, and that experience changes behavior. Users move toward dedicated platforms, build broader stacks, complete more tasks and give AI a larger role in complex decisions.
The next phase will reward providers that help users progress. Better models contribute to that progress, as do faster results and easier interfaces. User skill contributes too. Tools that teach people how to get better outcomes, remember useful context and fit naturally into work and personal routines can deepen trust over time. Personal use will continue to add newcomers. Workplace use already offers a large, seasoned base. Across both settings, the durable growth opportunity lies in turning occasional assistance into a dependable habit.
Read More
PYMNTS Intelligence is the leading provider of information on the consumer trends driving innovation in consumer finance, digital payments and financial inclusion. To stay up to date, subscribe to our newsletters and read our in-depth reports.
Methodology
“From Trial to Habit: How Time on Gen AI Deepens Consumer Use,” the ninth installment of the Consumer AI Benchmark Series, is based on a PYMNTS Intelligence survey of 2226 U.S. adults conducted in May 2026. The report examines how tenure relates to task breadth, platform choice, substitution and perceived value. Samples were balanced to match the U.S. adult population by age, gender, education and income.
