What 700 Million ChatGPT Users Actually Do with AI - And What It Reveals About Workplace Strategies
What 700 Million ChatGPT Users Actually Do with AI - And What It Reveals About Workplace Strategies
OpenAI published a massive study of ChatGPT consumer usage in September, conducted jointly with Harvard and Duke. ChatGPT currently has 700 million weekly users, representing approximately 10% of the world's adult population. The study analyzed a total of 1.5 million messages. While the research tells us nothing about organizational AI usage, it reveals fascinating insights into how people use personal AI services.
The study's key finding is that 49% of ChatGPT messages fall into the "Asking" category - requests for advice, learning, guidance, and decision support. 40% of messages are "Doing" messages, requests to perform specific tasks. The Asking category is growing faster and consistently receives better ratings from users - people want AI to support their thinking and decision-making.
Notably, in work-related usage, writing dominates - 42% of work messages relate to text production. Even more interesting is that two-thirds of these requests involve editing existing text rather than creating new content. People don't want AI to replace their creativity but to enhance it.
This raises important questions about workplace AI strategies. As Microsoft Copilot, ChatGPT Enterprise, and other business tools become more common, are they focusing sufficiently on what people actually want? Consumer ChatGPT usage strongly suggests that AI's value comes as much from supporting decision-making as from task automation.
From an organizational perspective, this opens intriguing opportunities for expanding AI strategy. While AI-powered automation solutions help employees with routine tasks, it might be worth considering what kinds of AI solutions support workers' decision-making and creative problem-solving.
How does your organization balance AI utilization between automation and decision support?
#AIStrategy #DecisionSupport #OrganizationalDevelopment
Reference: How People Use ChatGPT – OpenAI, Harvard & Duke (2025)
Marko Paananen
Strategic AI consultant and digital business development expert with 20+ years of experience. Helps companies turn AI potential into measurable business value.
Follow on LinkedIn →Related Insights
Shadow AI: Three Critical Risks Organizations Face
Employees increasingly share sensitive data with AI tools using personal accounts, creating governance and security challenges.
Autonomous AI Agents: Benefits and Hidden Risks
Autonomous AI agents are shifting work from task iteration to goal definition and evaluation. The risk: convincing but shallow output may flood decision-making.
AI Learns to Think Like Humans: The Hybrid Model Revolution
Nobel laureate Daniel Kahneman's 'Thinking, Fast and Slow' described how the human mind operates on two levels: fast, intuitive reactions to everyday situations and deep, analytical thinking for complex problems. Now AI is also learning to think like humans - sometimes fast, sometimes slow...
Interested in learning more?
Contact us to discuss your company's AI strategy.