I Trained an AI Agent to Deal With Boomers in Less Than 10 Minutes

Look, I love the boomers in my life. I really do. My old coworkers. The parents from my kid's basketball team. Salt of the earth, every one of them.

But God help them, they will click on anything.

"You are in this photo!" Click. "Reunion details here!" Click. "Your package is waiting!" Click click click.

And now? Their email accounts are compromised, and I'm the one paying for it. Every single day, my inbox fills up with phishing garbage, sent from people I actually know. Or at least, people whose names I know. The email addresses are different every time. karen.whitfield8847@sketchy-domain.ru one day, karenwhitfield.real@totally-not-a-scam.biz the next.

You can't block them. You block one address, three more pop up tomorrow with the same display name. It's email whack-a-mole, and my favorite inbox was turning into a dumpster.

I'm talking 5+ spam emails a day from the same five or six people. Bless their hearts, every one of them clicked a link they shouldn't have, and now their names are being used to flood my inbox with nonsense.

I tried filters. I tried blocking. I tried the "report spam" button so many times my thumb got tired. Nothing stuck.

The Realization

Here's the thing. I'm not a developer. I don't write code. I don't have a CS degree. What I do have is an AI assistant named Monstera that lives on a Mac mini.

I'd originally built Monstera (I give my agents houseplant names) for work. I was the social girl at a Web3/AI company, and I needed something that would watch X for trending keywords and ping me when a topic started heating up so I could jump on it. Then I got laid off a couple months ago, and Monstera became a very capable AI agent with nothing to do. Just hanging out on the Mac mini, waiting for me to put it back to work.

🌿 Monstera runs on 🦞 OpenClaw, and I talk to it through Telegram like I'd text a friend. It can read files, search the web, run code, send me alerts. Basically, it's a digital assistant that can actually do things, not just talk about them.

So one morning, staring at yet another "You won't believe these reunion photos!" email from Not-Actually-Karen-Whitfield, it hit me:

Wait. I have an AI that can run code on my machine. Email is just... a protocol. Can Monstera just... handle this?

Turns out: yes. In under 10 minutes.

The 10-Minute Setup

Here's exactly how it went down, step by step. I'm including this because I want you to see how easy it was. No terminal wizardry. No Stack Overflow deep dives. Just a conversation.

Minutes 1 to 3: Gave Monstera access to my email.

I'm being intentionally vague here because the specifics depend on your email provider, but the gist: I generated the credentials Monstera needed to connect to my inbox and sent them over in our Telegram chat. Most major email providers let you do this in a couple of minutes from their security settings page. If you have two-factor authentication on (and you should), there's usually an "app password" or "app-specific password" option.

Minutes 4 to 6: Told Monstera what I needed.

I literally just texted Monstera on Telegram:

"Hey, I need you to connect to my email and delete all the phishing spam from these people: Karen Whitfield, Brian Hollings, Doug Patterson, Linda Marsh, Gary Stenson. They keep sending me junk from different addresses but always the same names." (Fake names, obviously. To protect the innocent.)

Minutes 7 to 9: Monstera did the thing.

Monstera connected to my inbox and started scanning. Here's where the magic happened. Instead of trying to match email addresses (which change constantly), it searched by sender display name. That's the key insight. Spammers can spoof a different email address every day, but they have to keep using the same name if they want their target to recognize it and click.

It found 180 emails. Mostly phishing attempts.

I told it to nuke everything. It did.

Minute 10: Set it and forget it.

I asked Monstera to run this check automatically every day. It set up a recurring scan. Once a day it connects to my email, searches for those names, and nukes anything it finds. No input needed from me.

That's it. Ten minutes. Most of that was me clicking through my email provider's security settings and making my lunch.

The Result

That was a few weeks ago. Since then, Monstera has caught and deleted dozens more phishing emails. They just show up and disappear before I ever see them. My inbox is free of boomer mistakes for the first time in months.

The beautiful part? The spammers can keep changing email addresses all they want. It doesn't matter. Monstera searches by display name. As long as they keep pretending to be Karen Whitfield, they're getting caught.

(Sorry, real Karen, if you ever actually email me. You're collateral damage now. Maybe stop clicking links that say "you are in this photo" and we can talk.)

The Bigger Picture

Here's why I'm writing this: not because email filtering is revolutionary, but because of how it happened.

I didn't learn to code. I didn't read any documentation. I didn't set up a cron job or configure a server. I had a conversation with my AI assistant, in plain English, on my phone, and the problem got solved.

This is what AI agents are actually good for. Not writing your novel. Not replacing your job. Just... handling the annoying stuff. The boring, repetitive, "I can't believe I have to deal with this" stuff that eats up your time and energy.

Think about all the little tasks in your life that are technically solvable but not worth the effort to solve manually:

• Cleaning up your inbox ✓
• Organizing files that pile up in your Downloads folder
• Checking if your kid's school posted any schedule changes
• Monitoring a website for price drops

Each one is a 10-minute conversation away from being automated. You don't need to be technical. You just need to be able to describe what you want in plain language.

That's the unlock. AI agents aren't just chatbots. They're agents. They live on your computer. They can run code, access your stuff (with your permission), and do real work in the real world. The difference between ChatGPT and an AI agent is the difference between asking someone for advice and asking someone to actually do the thing.

I built Monstera for one purpose and ended up using it for something completely different. That's the other lesson here. Once you have an agent set up, every problem starts to look like something you can hand off.

Want One?

I use OpenClaw because it runs locally on my Mac mini. My data stays on my machine, and Monstera is available 24/7. I talk to it on Telegram, but it also works with Discord, WhatsApp, slack, whatever.

If you've got a computer that's always on, you can set this up. You don't need to be a developer. You just need to be willing to talk to your AI like it's a capable assistant rather than a search bar.

And if you've got boomers in your life who keep clicking phishing links?

Well, you could try to teach them to stop. Or you could spend 10 minutes training an AI agent to clean up after them.

I know which one's more realistic. 🌿