Hi Reader!

I probably owe you an apology.

Following the principle of "practice what you preach," I've been automating parts of my work process: connecting platforms, adding tools to Claude, streamlining workflows. One of those automations added all my Calendly contacts to this newsletter.

If you subscribed before, great. But if you suddenly received an email from me last Thursday out of nowhere, I'm sorry.

The thing is, I didn't fail to check. I didn't know what to check—or in this case, when to check. As in the majority of business automation cases, the problem is not with the tool but with the process.

Control won't help if there's no understanding.

The Flan and the Pancake

You might want to check the flan and pancake stories I shared a few newsletters back. Two AI-powered recruiting tools, both following instructions. One sent a flan recipe to a candidate. The other sent a pancake recipe—but only after it escalated to a human and asked for permission to send it.

The "pancake" automation worked. The "flan" automation broke.

And the difference wasn't the AI. It was whether someone had mapped the process before automating it.

My Calendly mistake was the same. I connected two platforms, hit "go," and assumed it would work as intended. But I didn't ask myself:

What are the actual steps in this workflow?
What could break?
And how would I know?

The result? Well, you know.

This Is the Automation Trap

We talk a lot about automation bias—the tendency to trust automated systems more than we should. But the deeper issue isn't that we trust AI too much. It's that we automate workflows we don't fully understand.

When you hand a process to AI without mapping it first, you lose visibility into what's happening in real life. You can't troubleshoot what you didn't design. You can't catch errors you didn't anticipate.

And over time, you delegate not just the task, but the judgment around it—without ever deciding to.

That's how over-automation leads to disengagement. People still care, but eventually they stop knowing where they're supposed to be involved.

The Fix: Map Before You Automate

In one of my previous newsletters, I wrote about the Human Agency Scale—a framework from Stanford University that forces clarity around a single question: where does human judgment live in this workflow, and who's accountable for the outcome?

Before you connect two platforms and hit "go," stop. Map the process:

What are the steps? Write them out. All of them. (Or at least those that involve people outside of your organization.)

Where does judgment matter? What decisions are being made, and who should be making them?

What could break? And how would you know if it did?

If you can't answer those questions, you're not ready to automate yet. You're just hoping the AI figures it out. (That'd be awesome, right?)

What I'm Taking From This

I teach leaders to build teams who keep their judgment in an AI-driven world. And this week, I learned the lesson again myself.

Automation is powerful. But it's not a shortcut around understanding your own work. It's a tool that amplifies whatever process you feed it—clear or chaotic, thoughtful or rushed.

The goal isn't to automate less. It's to automate with intention.

And that starts with knowing your workflow well enough to decide what you're actually handing over.

So if you feel like you don't want to receive these anymore, just click the Unsubscribe button at the very end of this letter. But if you want to give it a chance—that would make me super happy.

See you next Thursday. (I hope.)

Daria

P.S. I can't be the only one who's over-automated something and regretted it. Hit reply and tell me your AI flop story—I'd love to hear what went wrong and what you learned.