TLDR: DeskTime data from 50,000 users shows ChatGPT’s workplace AI time share plunged from 99.91% in 2023 to 74.71% by early 2026. Gemini and Claude surged as Copilot stayed near 1%, pressuring leaders to keep up.
Key Takeaways:
- DeskTime tracked power users logging at least 26 hours of AI work annually, using anonymized data across 50,000 people over three years.
- ChatGPT fell from near total dominance in 2023 to 74.71% in the first four months of 2026, while Gemini reached 14.38% and Claude hit 8.56%.
- Copilot stayed stuck around 1% and smaller rivals like Perplexity and Mistral failed to break through, suggesting workplaces are spreading across tools.
When your product stops being the default, “innovation” becomes a speed contest, not a feature list. If Gemini and Claude keep converting daily habits, ChatGPT may need more than upgrades to win back workflow gravity.
When your product stops being the default, “innovation” becomes a speed contest, not a feature list. If Gemini and Claude keep converting daily habits, ChatGPT may need more than upgrades to win back workflow gravity.
Q&A
If DeskTime’s numbers are based on one tracking service, what could make the real competitive picture look even worse or better for ChatGPT?
Usage patterns differ by job role, browser behavior, and what counts as AI time, so results can tilt toward tools favored by certain workflows. A multi measurement approach would confirm whether the slide is universal or concentrated in specific office use.
Why might Gemini and Claude be winning share specifically inside productivity workflows rather than just casual experiments?
They likely match common office tasks with faster iteration cycles and stronger fit for writing, summarizing, and knowledge work. Winning repeat usage depends less on one spectacular demo and more on day to day reliability.
What happens to security and IT governance when companies diversify from one AI interface to several?
Procurement, logging, and data handling policies must cover multiple vendors, which can slow rollouts even as productivity rises. Teams may adopt separate controls for each tool or push for an enterprise layer to unify permissions.
Why would Copilot remain near 1% even as AI adoption accelerates at work?
Stagnation can signal strong brand familiarity without enough workflow stickiness, or that usage is already capped by how users interact with Microsoft apps. It may also reflect measurement gaps if AI time is logged differently across environments.
Could ChatGPT reverse the curve, and what would “reversal” actually require in practical terms?
To reverse share loss, it would need to reclaim repeat daily use, not just attract trials. That usually means deeper workflow integration, tighter enterprise controls, and measurable speed or accuracy gains that users feel in recurring tasks.
Top in AI
AI giant surges 1,200% as valuations cool
Lumen Technologies jumps as AI hyperscalers quietly reroute demand
Illinois Legislature OKs SB 315, forcing AI third party audits
Pymetrics AI screening rejects Black and Asian applicants more
No comments yet. Be the first to share your thoughts!