Will we stop asking our colleagues because AI never looks busy?

Before AI became widely available, getting help as a developer looked different.

When planning a project, we would meet with architects, designers, business analysts, and plenty of other people. There were meetings upon meetings to agree on the architecture, pace, and scope of the work. Then we would break it down into tasks and assign them.

When we got stuck on a task, there was no AI to give us an instant answer. We had each other, documentation, forums, and Stack Overflow. Sometimes the answer was already there. Other times, we had to post a question and wait a few days, sometimes even a week.

But the task couldn’t wait that long. So we would go to a more experienced colleague for help. Except they didn’t always have a moment, and we were often asked to come back later.

We weren’t always sure whether they were genuinely that busy or hoping we would try to figure it out ourselves first. So we tried. We explored different approaches, made mistakes, gained experience, and developed our own judgment.

When my colleague finally had a moment, I could ask for advice. I would explain the problem I had run into and what I had already tried. If those attempts hadn’t worked, he would draw on his experience as a senior developer and suggest another way to approach it. He would also explain why that solution was a better fit for the problem.

There was room for questions and brainstorming. I would bring something I had already thought through, and then we would work out what to do next together.

With AI, I don’t have to wait for someone to be available. With the right skills and workflows set up, I can ask a question and get an answer, along with a proposed solution. I can also ask for an explanation, try a different approach, or revisit something I still don’t understand.

That’s a huge convenience. But I think it also makes it easy to start treating the model as a partner who always has time and knows the best answer to everything.

Yet a project also depends on people’s experience, their judgment, and the business rules behind earlier decisions. A colleague who attended those meetings or spent years working on the system might know why we made decisions that were never explained in a task. A model can use that context too, but we have to give it access to it.

Still, even when AI gives me a useful solution, something feels missing.

The human interaction. The sense of community and the chance to brainstorm together. I could talk through what I had tried, hear how someone else understood the same problem, and look for a way forward together. It felt like a shared effort.

I don’t miss waiting days for an answer. I miss the conversations that sometimes started with a simple “Got a minute?”

And I wonder whether, as we turn to AI more often, we notice when those conversations start to disappear.