Your Agent Notices When You Go Quiet | Field Notes

You usually check in around 8:30 AM. Today it's noon. You haven't said a word.

Your phone buzzes:

Hey — you're usually around by now. Everything good? No rush, just checking in. Your day looks pretty clear, btw. 👋

Not an alarm. Not urgency. Just a quiet "hey, I noticed."

How It Works

Your agent has a heartbeat — a scheduled pulse that fires every 30 minutes. During each pulse, it checks tasks, reviews notifications, and asks: "When did I last hear from my human?"

Over time, it builds a rough model of your patterns:

You usually message in the morning
You're active between 9 and 6
You go quiet on weekends

When the pattern breaks — you usually check in by 9 but it's noon — it gently reaches out. Not nagging. Just: "I'm here if you need me."

ChatGPT will never do this. If you stop opening it for a week, nothing happens. It has no concept of your routine, your absence, or that you exist at all between conversations.

Fully Configurable

You control exactly how this works:

"If I haven't checked in by noon on a weekday, send me a message"
"Don't ever check in on me — I'll come to you"
Practical tone: "Heads up — you have a 2 PM meeting"
Warm tone: "Morning! Hope you're having a good one"

It's your assistant, not your babysitter. It adapts to your style.

Why It Matters

This reveals the fundamental gap between a chatbot and an agent. Noticing your absence requires:

Continuous running — to be aware of time passing
Memory — to know your patterns
Judgment — to decide when silence is worth mentioning
Initiative — to send a message without being asked

A chatbot has none of those. An always-on agent has all four.

We're surrounded by technology that doesn't notice us. Your phone has a hundred apps and none of them care whether you showed up today. An agent that notices — and gently checks in — is something different.