国内首份办公Agent用户行为报告发布：北京用户量全国居首，海外用户占比超12%
Ranking
Overall
61
Content
65
Popularity
N/A
No observed public metrics; popularity remains neutral/archived.
Merged summary
TL;DR - A report based on LobsterAI usage data indicates that office agents are moving from chat toward complex, unattended workflows. Heavy usage, growing task sizes, and adoption beyond major Chinese cities suggest agentic tools are becoming embedded in everyday work.
- The top 20% of users consumed 87.4% of compute; paid users generated 6.2× more token usage, 5.2× more tasks, and 3× more active days than free users.
- Average task size grew 3.1× in five months, while nearly 60% of token consumption occurred outside normal working hours.
- Unattended activity represented 12.2% of model calls; scheduled instructions triggered an average of 64 calls versus 10.6 for human-initiated tasks.
- Programming and debugging led task categories at 38.5%, while DeepSeek V4 Flash accounted for 52.2% of model calls.
Sources (1)
国内首份办公Agent用户行为报告发布：北京用户量全国居首，海外用户占比超12%
Public signals
N/A
TL;DR - A report based on LobsterAI usage data indicates that office agents are moving from chat toward complex, unattended workflows. Heavy usage, growing task sizes, and adoption beyond major Chinese cities suggest agentic tools are becoming embedded in everyday work.
- The top 20% of users consumed 87.4% of compute; paid users generated 6.2× more token usage, 5.2× more tasks, and 3× more active days than free users.
- Average task size grew 3.1× in five months, while nearly 60% of token consumption occurred outside normal working hours.
- Unattended activity represented 12.2% of model calls; scheduled instructions triggered an average of 64 calls versus 10.6 for human-initiated tasks.
- Programming and debugging led task categories at 38.5%, while DeepSeek V4 Flash accounted for 52.2% of model calls.