I Gave My AI Agent a Monday Morning Job

There is a task I have done every Monday morning for two years.

It isn't hard. It isn't even particularly time-consuming, really. It's just always there. The research pull. The weekly briefing compile. The "okay, what's relevant this week" ritual I used to run with three browser tabs open, a notebook, and one eye on Slack.

Most Mondays, I half-did it.

Not because I don't care about the work. I care deeply about it.

But Mondays have about seventeen competing priorities trying to get my attention before 9am. The research task always felt like something I should have done Sunday night. And Sunday nights belong to something else entirely.

Here's the thing I don't hear enough people say out loud: it's the small tasks that grind you down. Not the big projects. The recurring, background-of-the-brain, takes-20-minutes-if-I-rush-it tasks. They don't feel like a crisis. They feel like a constant, low-grade drain. The kind that you don't notice until one Tuesday morning you realize you've been half-present at your own Monday briefing for months.

I kept waiting for a solution that felt technical enough to justify using. A workflow. An integration. Something I could feel sophisticated about.

And then I just described what I wanted.

This is what I actually typed into Claude: "Every Monday morning, I need a briefing on leadership clarity, AI adoption in retail, and operational trends for mid-market leaders. Pull from public sources. Summarize in under 300 words. Highlight anything connecting to decision frameworks or team psychological safety."

I hit return.

Three minutes later, I had a briefing.

Not a perfect one. Not a finished product. But a starting point that was more useful than the half-done version I'd been producing for myself under time pressure.

I want to be honest about what happened next, because I think the overpromise version of this story does more harm than good.

Week one, I edited it a lot.

Week two, I edited it less.

Week three, I mostly just read it.

By week four, something small had shifted. Monday mornings felt different. Not dramatically different. The dread around that one specific task was gone. And what replaced it was an extra 30 minutes I wasn't spending on something I never wanted to own in the first place.

Here's what the editing process taught me:

The AI doesn't know what I mean by "relevant." I had to teach it. Not through code. Through conversation.

"When you find a trend, tell me why it matters to a leader who has 200 employees and no technical background. Don't summarize the technology. Summarize the implication for the people managing the work."

That one instruction changed the output entirely.

This is what I keep coming back to when people tell me AI is too complex: you're not programming a machine. You're briefing a very fast, very literal collaborator who needs specific direction. That is a skill most leaders already have. They just haven't applied it here yet.

The technical barrier to using AI agents in 2026 is genuinely lower than it has ever been. Claude, ChatGPT, and Microsoft Copilot (which you probably already have in your Office suite, whether you've opened it intentionally or not) have all built agent-like behaviors into their core interfaces. You can describe a task in plain language and get a working starting point.

What I keep seeing hold women back isn't capability. It's the quiet impostor syndrome around whether they're "supposed to" use these tools. Whether they need a technical background. Whether someone will ask "how did you do that?" and they'll have to say "I just asked it."

You're allowed to just ask.

Recent data is clear: women non-technical leaders are adopting AI tools faster than men right now. The gap isn't skill. The gap is confidence that the skill they already have is enough.

Here's what I wish someone had told me earlier: you don't need to understand how the technology works. You need to understand what you want done. That is leadership. That is what you do every time you brief a colleague, onboard a contractor, or write a job description.

AI agents are a capable, fast, very literal team member who needs clear direction. You already know how to give clear direction.

We spend a lot of conversation about AI talking about what will be automated in the future. What roles will change. What skills will become obsolete.

Very little of that conversation is aimed at the people who are already running operations, managing teams, and doing meaningful work right now. The non-technical leader who just wants her Monday mornings back.

The access point isn't a technical training. It's permission.

Permission to describe what you want in plain language and trust that's enough to start.

This week, pick one recurring task that takes 20 to 45 minutes and that you consistently half-do under time pressure.

Write a description of what you actually need from it. Not what the task is called. What you need it to produce. For whom. In what format.

Give that description to Claude (claude.ai, free account works fine) and see what comes back.

Don't expect perfection. Expect a first draft that's better than nothing. Edit it once. Note what you'd change. Brief it again with those notes added.

That's the whole process.

The feeling you're chasing isn't "wow, AI did it perfectly." The feeling is: I just got 30 minutes back on a Monday morning, and that's 30 minutes I used for something I actually wanted to do.

Chase that feeling.

See you next Tuesday.

— Jonelle

P.S. What's the recurring task you've been meaning to get off your plate? The one that's always there, always slightly underdone? Reply and tell me what it is. I read every single one.