Updated September 2026.
In September 2025, OpenAI published the largest study ever released on how people actually use ChatGPT, with data through July 2025. At the time, ChatGPT had more than 700 million weekly active users sending about 2.5 billion messages a day. Usage has kept growing since: in February 2026, OpenAI reported more than 900 million weekly active users.
At Amicatek, we work with AI solutions in partnership with Pareto, the company behind Tess AI, so we follow this data closely. Below we summarize the study (by OpenAI’s Economic Research team and Harvard economist David Deming, published as an NBER working paper) and what it means for businesses.
The Numbers that Impress
The study classified a sample of about one million de-identified messages from ChatGPT consumer plans, sent between May 2024 and June 2025. Highlights:
- About 70% of messages are personal, not work-related (73% in June 2025).
- 2.6 billion messages per day in June 2025.
- 483% growth in daily message volume between June 2024 and June 2025.
- Nearly half of all messages sent by adults came from users under 26.
- The gender gap has closed: users with typically feminine names went from 37% (January 2024) to 52% (July 2025).
- By May 2025, adoption was growing more than 4x faster in the lowest-income countries than in the highest-income ones.
The 3 Main Uses of ChatGPT
According to the study, about 77% of conversations (nearly 80%) fall into three categories: Practical Guidance, Seeking Information and Writing. This concentration reveals consistent patterns in how people actually use generative AI, which is useful for business strategy.
Practical Guidance (29% of messages)
What it is: The most popular category, including tutoring and teaching, how-to advice, creative ideation, and health, fitness and self-care guidance, all tailored to the user’s needs. Tutoring and teaching alone account for 10.2% of all messages, and how-to advice for 8.5%.
Key Features:
- Personalization: unlike generic information, it provides advice tailored to the specific context.
- Interactivity: allows follow-up and refinement throughout the conversation.
- Versatility: covers everything from personal fitness to business strategy.
Study Insights:
The study draws a clear line between the two closest categories: Practical Guidance is highly customized to the user and evolves through follow-up questions, while Seeking Information is about factual answers that should be the same for everyone.
Business Examples:
- Corporate Tutoring:
- Personalized onboarding programs.
- Adaptive technical training.
- Virtual mentoring for professional development.
- Business Guidance:
- Customized marketing plans.
- Specific sales strategies by segment.
- Guidance for operational problem-solving.
- Creative Idea Generation:
- Development of advertising campaigns.
- Innovation in products and services.
- Naming projects and initiatives.
Writing and Communication (24% of messages)
What it is: Text production, but mostly editing, critiquing, summarizing and translating content the user provides. Writing’s share fell from 36% of messages in July 2024 to 24% a year later, not because people write less with AI, but because other uses grew faster.
Key Finding of the Study:
About two-thirds of writing messages ask ChatGPT to modify the user’s own text rather than create new content from scratch. At work, AI acts much more like an editor than an author.
Business Use:
The study shows that writing makes up 40% of work-related messages on average (June 2025), making it the most common professional use. Among management and business professionals, the share rises to 52%, and it is also around half of work messages in education, health care and non-professional occupations (49% to 50%).
Strategic Applications for Businesses:
- Corporate Communication:
- Optimization of executive emails.
- Standardization of internal communications.
- Improvement of sales proposals.
- Content Marketing:
- Refinement of social media posts.
- Creation of impactful newsletters.
- Development of promotional materials.
- Business Document Creation:
- Procedure manuals.
- Internal policies.
- Executive reports.
Seeking Information (24% of messages)
What it is: Looking up specific information about people, current events, products and recipes. The study notes that this category works as a close substitute for web search.
Significant Growth:
It was one of the fastest-growing categories in the period analyzed:
- July 2024: 14% of all messages.
- July 2025: 24% of all messages.
- That is roughly 70% growth in share over 12 months.
Meanwhile, Technical Help (programming, math, data analysis) fell from 12% to about 5% of messages, with programming alone at 4.2%.
Competitive Advantages of AI vs. Traditional Search:
- Contextual Personalization:
- Responses tailored to the company profile.
- Consideration of specific business context.
- Synthesis of multiple sources into a single response.
- Operational Efficiency:
- Elimination of the need to filter multiple results.
- Direct and structured responses.
- Possibility of refinement through follow-up questions.
Strategic Implementation:
- Market Research: Example of use: “What are the main trends in [your industry] for the next 12 months, considering companies of the size [small/medium/large] in Brazil?”
- Competitive Analysis: “Compare the digital marketing strategies of [Competitor A] and [Competitor B], focusing on [specific relevant aspect].”
- Trend Monitoring: “What regulatory changes in [specific area] can impact companies like ours in the next 6 months?”
How People Use ChatGPT at Work
Only 27% of messages were work-related in June 2025, but that still meant about 716 million messages a day, more than triple the 213 million of June 2024. The study also classifies each message by intent:
| Intent | Share of all messages | Share of work messages | Definition |
|---|---|---|---|
| Asking | 49% | 35% | Seeking information or advice to make a decision |
| Doing | 40% | 56% | Asking ChatGPT to produce something concrete (text, code, a plan) |
| Expressing | 11% | 9% | Sharing views or feelings without requesting an action |
Asking and Expressing grew much faster than Doing over the year analyzed, and Asking messages receive the best user ratings. Mapped to the O*NET work activities, just seven activities (such as getting information, documenting information and making decisions) account for 77% of all messages and about 81% of work messages. The study’s conclusion: ChatGPT’s main economic value is decision support, especially in knowledge-intensive jobs.
How Amicatek Can Help Your Business Adopt AI
Based on these findings, Amicatek approaches AI adoption starting from the uses that create the most value. As specialists in WordPress and Artificial Intelligence, in partnership with Pareto (Tess AI), we cover the three usage patterns identified in the research: Asking, Doing and Expressing.
Our AI Solutions Based on Real-World Data
Strategic AI Implementation
Because 77% of usage is concentrated in 7 main activities, we organize our work into three tracks:
| Track | Activities covered | Who it is for | How we deliver |
|---|---|---|---|
| AI for communication | Documenting, interpreting, communicating | Management, marketing, sales | AI in WordPress: content, review and on-site support |
| AI for decisions | Solving problems, getting information | Executives, managers | Agents and assistants grounded in your company knowledge |
| AI for execution | Thinking creatively, working with computers | All industries | AI automation of repetitive processes |
Turn the insights from this study into concrete results for your business. Request a free assessment and we will map where AI creates the most value for your company.