Is AI Killing Traditional Org Structures?

Posted on In Views:

Traditional silos limit the productivity gains unlocked by AI workflows

As I mentioned in my previous article, I recently made a career shift. I moved from being a pure software architect to a hands-on staff role. This means I am now deeply involved in implementation.

I made this adjustment because there were things I wanted to verify. For instance, I wanted to see how much productivity an AI-powered workflow could unlock. The results proved that the efficiency gains are significant.

This led me to think about another question.

Can traditional organizational structures still serve their purpose?

The “traditional” structures I am referring to fall into two categories:

  1. Organizations divided by function, such as PM, ENG, QA, etc.
  2. Organizations divided by Cross-functional teams. For example, a Scrum team has a PO, developers, and QA. This might seem new rather than traditional to some people. Nevertheless, I believe this is still a traditional setup from the era before AI entered our workflows.

I raise this question because AI empowers us to increase productivity to a dramatic extent. However, the silos created by traditional organizational structures work against this boost.

There are many reasons for these silos. We have functional silos between roles. We have cognitive silos between different feature modules. And we have responsibility silos between departments. These boundaries used to seem natural. Now, AI breaks them down easily.

Let’s look at an example. We often used domain boundaries or microservice boundaries to split organizations. When Team A’s changes touched code owned by Team B, it required a lot of communication and coordination. Even development and deployment created dependencies.

Waiting is inevitable in this process. This constitutes waste.

But if this traditional organizational structure is broken, how should we define boundaries? Or simply put, how do we clarify responsibilities?

I don’t have the answer yet. I am still looking into it.

Originally published on Medium