A massive new dataset tracking 700 million ChatGPT interactions through July 2025 has uncovered a widening gap in how people approach AI—and why some extract enormous value while others barely scratch the surface.
The key difference isn’t about access to tools, budget, or technical skills. It comes down to mindset: are you treating AI like a task-finishing robot or as a thinking partner?
The Sophistication Split
Researchers found two primary usage patterns:
- Task Automation Users
These people rely on ChatGPT to generate emails, draft reports, and create slide decks. They want plug‑and‑play outputs that can be copied into workflows with minimal edits. - Decision Support Users
Instead of asking for finished work, this group uses ChatGPT to analyze data, brainstorm solutions, and sharpen their reasoning. They don’t outsource judgment—they amplify it.
The outcome? “Asking” interactions (seeking input, advice, or critique) score consistently higher on satisfaction than “Doing” interactions (direct task execution). Even more, “Asking” usage is growing rapidly while “Doing” usage declines.
Who’s Winning With AI
The data shows a clear pattern:
- Graduate degree holders are significantly more likely to use ChatGPT for decision support.
- Knowledge workers in higher-paying roles lean on AI as a research and analysis partner, while administrative roles focus on task completion.
- Users treating AI as a consultant create measurable value—estimated at $98 per month per user, translating to $97 billion in total U.S. economic value.
Interestingly, programming help has dropped from 12% to 5% of usage. Meanwhile, writing dominates—42% of all work-related use—but two-thirds of those conversations involve improving drafts, not generating from scratch.
The Real Competitive Advantage
The most sophisticated users concentrate almost entirely on two activities:
- Information Processing: organizing, interpreting, and contextualizing data.
- Decision Support: evaluating problems, weighing trade-offs, and thinking creatively.
Notice what’s missing: repetitive automation and shortcutting thought. The advantage comes not from what AI produces for you, but how it shapes your thinking.
Why Most AI Advice Gets It Wrong
Much mainstream productivity advice suggests: “find tasks AI can do for you.” The data says the opposite works better.
High performers ask: “How can AI help me think differently?”
That subtle shift explains why satisfaction scores are higher, adoption is faster, and professional users are pulling further ahead. The gap between “automation users” and “thinking partners” is widening—and it’s redefining the future of work.
Takeaway
Stop using AI to avoid thinking. Start using it to think better.
At Dynaimix, we see this play out daily with clients. Teams that learn to treat AI as a research partner—not just an automation tool—unlock deeper creativity, faster problem-solving, and higher‑quality decisions.
If you want to move beyond surface‑level automation and start training your team to use AI as a true thinking partner, check out our Intro to AI course or book a Dream Session with us. Let’s build your competitive edge before the gap widens further.