When I first saw this news, the first thing that came to mind wasn’t a hospital—it was our own group chat
I’ve been stuck here too: when a project gets messy, my first instinct is to manage the team with more spreadsheets, more reminders, more check-ins. Later I realized that once people feel watched all the time, they may move faster, but their judgment gets worse, and the relationship breaks down first.
What this is: it’s not that AI is too smart—it’s that it has started “watching people” for us
In this report, nurses at Kaiser are talking about something very real: AI and workplace surveillance haven’t made the job lighter; they’ve made the care experience worse. That hit a nerve for me, because a lot of small teams are heading down the same path: using systems to track response speed, online hours, and script compliance. It reminded me of 9 p.m. last Wednesday in a coffee shop on Wantang Road in Hangzhou, when Lin Wen, who runs a knowledge-product business, was replying to student messages while telling me that after she installed a tracking plugin, her assistant started caring only about “replying fast” and no longer dared to pause and make careful judgment calls. I’ve gotten this wrong too, thinking that more data meant steadier management. But in some kinds of work, the most valuable part is exactly that human pause, judgment, and consideration.
What it costs to copy today: start with “less monitoring”—it almost costs nothing
Money: 0-99 yuan; Time: 30 minutes; Technical barrier: no coding needed, just basic backend settings; First step: open the admin panel for the project management, customer service, or attendance tool you’re already using, and find “notifications / analytics / member visibility.” Switching from “real-time watching” to “results review” is enough to start. For example, look less at online hours and more at weekly delivery; look less at the seconds per reply and more at whether the customer’s issue was actually solved. Not everyone needs this kind of tool, and it’s fine not to try it right now. In my experience, writing the rules clearly often helps more than adding one more monitoring plugin.
If I were choosing at different stages, here’s how I’d do it
Just starting out: I wouldn’t rush into monitoring. Chat records and a task list are enough for now; the key is getting clear on what “done well” actually means.
With 1-2 clients: if missed messages are starting to worry me, I’d add automatic reminders first, not employee tracking. I’ve stepped on that rake myself.
Scaling up: if I were leading 3-5 people, I’d use AI to organize meeting notes, summarize issues, and spot repeated mistakes—not to watch who replied at what time. What we want is more peace of mind, not a team that feels more tense.