Damian Nomura

<start of post>

The first AI workflow you automate should not be the one you understand least.

I watched a team spend three weeks building a beautiful system around a process nobody had mapped properly.

The dashboard was clean. The prompts were clever. The output looked impressive.

The inputs were wrong.

That is the part people skip when they talk about AI adoption. They start with the tool because the tool is visible. The messy handoffs stay invisible.

Someone exports the data. Someone changes the fields. Someone makes a judgment in the middle and forgets to write it down. Someone else cleans up the result before it reaches the person who needs it.

By the time the AI enters the picture, the workflow has already become a small archaeological site.

So we slowed it down.

Not to make the team document everything. Not to create another process nobody follows. Just to find the places where the work changed shape as it moved from one person to the next.

That was where the useful automation lived.

Not in replacing the decision. In removing the waiting around it.

The team did not need a bigger tool. They needed the same work to arrive in a form the tool could actually use.

This is why I am running a small working session next week for teams figuring out where AI belongs in their workflows.

No slide deck about the future of work. No generic list of use cases.

Bring one process that currently depends on too much copying, checking, chasing, or fixing.

We will map it, find the judgment inside it, and build one small improvement together.

Thursday, 17:30, online. Small group. Link in the comments.

What is the workflow in your company that everyone knows is broken, but nobody has had time to look at properly?

<end of post>