3 min. reading

What 700M ChatGPT Users Reveal About How We Use AI

OpenAI published the first major study on how people actually use ChatGPT. Researchers analysed 2.5 billion messages daily from 700 million active users. For e-commerce, there are seven concrete patterns that can be used right now.

Katarína Šimčíková Katarína Šimčíková
Partnership Manager & E-commerce Content Writer, Ecommerce Bridge EU
What 700M ChatGPT Users Reveal About How We Use AI
Source: ChatGPT

ChatGPT reached 10% of the global adult population in less than three years. The study mapped user behavior from May 2024 to July 2025 through automated classifiers – no human saw message content during analysis to protect privacy.

Writing Help Dominates Work Usage

40% of all work-related messages involve writing. But here’s the catch – two-thirds of these messages aren’t about creating new text. Users want to modify what they already have.

“Edit this product description to be more technical” is a more typical message than “Write me a product description from scratch.”

Asking Beats Doing

The study introduced a new classification:

  • Asking (seeking information to make decisions)
  • Doing (creating output)
  • Expressing (expressing feelings).

Result: 49% of messages are Asking, 40% Doing, 11% Expressing.

More interesting: Asking is growing faster and has higher user satisfaction. For work-related messages, Doing is at 56%, but Asking messages get better ratings from users.

Line chart showing ChatGPT usage trends from July 2024 to June 2025. Asking grew to 51.6%, Doing declined to 34.6%, Expressing rose to 13.8%.

Soure: cdn.openai.com

Practical Guidance is The Most Common Use Case

29% of all ChatGPT messages fall into Practical Guidance – tutoring, how-to advice, creative ideation. Another 24% is Seeking Information (replacing Google search) and 24% Writing.

These three categories cover 77% of total usage.

10% of all messages are direct tutoring requests. People don’t just want to buy a product, they want to learn how to use it.

Search Behavior is Changing

Seeking Information grew from 14% to 24% in one year. Conversational search is replacing keyword-based search.

Mobile and Multi-channel Reality

36% of users express loyalty through brand apps. But communication preferences are divided: email 69%, SMS 43%, direct mail 43%, mobile push 30%.

Emerging Markets Grow Fastest

Disproportionate growth is in countries with GDP per capita $10,000-40,000. Low-to-middle income countries are adopting ChatGPT faster than developed markets.

Treemap showing ChatGPT usage categories: Practical Guidance 28.3%, Writing 28.1%, Seeking Information 21.3%, with detailed subcategories breakdown.

Source: cdn.openai.com

What This Means Practically

The study shows one fundamental fact: the majority of ChatGPT usage is non-work related, and this share is growing rapidly. Non-work messages increased from 53% (June 2024) to 73% (June 2025).

Home production – shopping, decision-making, learning – is becoming a bigger part of the economic impact than workplace automation.

Computer programming? Just 4.2% of messages. Relationships and companionship? 1.9%. Most people use ChatGPT for practical decisions and information seeking.

The gender gap has closed – in June 2025, there were more active users with typically female names. The age gap persists – 46% of messages from users under 26.

About the study: Methodology – the study analyzed consumer ChatGPT users (Free, Plus, Pro plans) from May 2024 to July 2025. Messages were classified using automated LLM classifiers – no human saw message content during analysis to protect privacy. The sample excluded users under 18, those who opted out of data sharing, and Business/Enterprise accounts.

Share article
Katarína Šimčíková
Partnership Manager & E-commerce Content Writer, Ecommerce Bridge EU

Partnership Manager & E-commerce Content Writer with 10+ years of international experience. Former Groupon Team Lead. Connects European companies with Slovak and Czech markets through partnerships and content marketing.

Similar articles
Building AI? This Booking.com Leader Says Start Small and Solve One Problem Well
8 min. reading

Building AI? This Booking.com Leader Says Start Small and Solve One Problem Well

What does it take to build AI responsibly at a global platform? Our expert Marija Ristovska talked with Marina Angelovska from Booking.com to find out—covering everything from managing uncertainty to why your first AI project should start small and focused.

Marija Ristovska Marija Ristovska
E-commerce Project Manager | Marketing and PR consultant and Strategist, E-commerce Macedonia Association
Marketers Now Create Ad Campaigns in Hours, Not Weeks
3 min. reading

Marketers Now Create Ad Campaigns in Hours, Not Weeks

Marketers no longer need weeks to create advertising campaigns. A new system combining AI tools can produce quality creatives in hours instead of weeks. Marketing expert Liana Hakobyan tested the methodology on campaigns for Dior with remarkable results. Traditional ad creation is broken. While teams waste time in endless revision cycles and approval processes, competitors […]

Liana Hakobyan Liana Hakobyan
Marketing Strategy Lead | TEDxSpeaker | Microsoft Startup Finalist, Iris.ai
Alibaba Launches Its Most Advanced AI Coding Tool
2 min. reading

Alibaba Launches Its Most Advanced AI Coding Tool

Alibaba dropped their latest AI tool Wednesday – Qwen3-Coder, which they say beats local competition and matches GPT-4 in some areas. It’s free, open-source, and designed to write code without human help.

Katarína Šimčíková Katarína Šimčíková
Partnership Manager & E-commerce Content Writer, Ecommerce Bridge EU