There's a pattern emerging in how people actually use AI tools at work — not the use cases in the press releases, but the real ones. It's less about automation and more about what we'd call working modes: the specific cognitive postures you adopt depending on what kind of problem you're in.
A writer using AI to break through a blank page is doing something fundamentally different from a product manager using it to pressure-test a spec, or a support lead using it to triage tickets at 2am. The tool is the same. The working mode — the relationship between the human and the output — is completely different in each case.
Why it matters
Organizations that treat AI adoption as a single initiative tend to get undifferentiated results. The ones getting leverage are the ones — often implicitly, sometimes explicitly — matching the tool to the mode. They're not asking "how do we use AI?" They're asking "what are we actually trying to do here, and is AI the right shape for it?"
That's a harder question. It requires people to have enough self-awareness about their own work to identify which parts are bottlenecked by generation, which by judgment, and which by context that the AI simply doesn't have.
The practical version
We've started asking every guest some version of this: when you reach for an AI tool, what mode are you in? The answers are almost always more specific and more interesting than "I use it to save time." Time is rarely the real constraint. Attention, clarity, confidence, context — those are closer to it.
The most effective AI users we've spoken with treat these tools less like productivity software and more like a second working session with a thoughtful generalist who has no ego about being wrong. That reframe changes everything about how you interact with them.