AI Teammates Are Asynchronous by Design — and I Love It

When I work with a human teammate (it is wild and vertigo-inducing that we now have to specify human), I don’t stand over their shoulder waiting for them to finish every task before I give feedback and move on. That would be a terrible way to work.

But that’s exactly what working with a chatbot is like today.

With Claude, ChatGPT, or any chat-based AI, the interaction is synchronous by default. You prompt, it responds, you respond, it continues. The model is working, you’re waiting, thirty seconds pass, and you’re back in the window. So what ends up happening is that you run three or four parallel chat sessions to stay productive between responses, or you do a quick task and jump back before the chat goes cold. The result is a work experience that’s constantly fragmented and reactive. Context switching all day. Not great.


AI Teammates work differently, and I didn’t fully appreciate it until a conversation with @Bastien_Siebman today.

Because Teammates operate through comments and task assignments rather than chat, there’s a natural delay built in. I assign work, move on, and come back not when the Teammate finishes — but when I’m ready to revisit. That’s a fundamentally different dynamic.

When one of our Teammates completes a task, it surfaces in my Asana inbox alongside everything else. I schedule it into My Tasks. I get to it on my terms. Not when the tool is ready for me — when I’m ready for it.

That’s exactly how I work with the humans on my team.


I’m calling this “the delay is a feature,” but what I really mean is that the asynchronous structure of AI Teammates is a forcing function. It pushes you toward the way you probably already want to work: self-directed, organized, deliberate — rather than reactive and tethered to a chat window.

Many of us have spent years trying to build a more asynchronous work culture — specifically to protect heads-down time, reduce interruptions, and give people space to think. Chatbots came along and blew all of that up.

For anyone who’s noticed that their AI productivity time tends to feel frantic, this might be the alternative worth exploring.

8 Likes

I love your take on that! cc @Arthur_BEGOU @Filippo_Baj_i.DO

Love this take!!

Yes, this is a great distinction to call out. Nice, @Liohn_Sherer (and that other guy you were talking to).