So here’s a little experiment I’ve been running behind the scenes: I’ve been using ChatGPT as a kind of virtual productivity manager. Think: personal assistant, accountability buddy, and to-do list — all in one.
I set up a new project inside ChatGPT called Productivity, gave it a little bit of instruction (basically: “you’re my productivity agent — help me manage my day”), and then I started treating it like an actual assistant.
Here’s how it works (and how you could do it too):
1. Create a dedicated Productivity project
Inside ChatGPT, I spun up a new project thread and used the custom instructions feature to shape its behavior. I told it:
“I want you to be my productivity coach. I have a lot of balls in the air, and need help getting stuff done. Specifically helping me make lists and prioritize what needs to be accomplished.”
That way, I could always come back to this thread and pick up where we left off.
2. Build your daily rhythm
I started each day with a quick check-in:
“Good morning. What’s on the cards today?”
And ChatGPT would give me a breakdown based on:
- My recurring tasks (like clearing email)
- Anything I told it during the week (like meetings or errands)
- Projects I was working on (like a custom app build)
This was key: I wasn’t inputting everything at once. I just dropped notes into the thread throughout the day. Things like:
- “Had a call with Sarah, need to follow up with pricing.”
- “Remind me to get Builders Plastic on Thursday — we’re doing floors on Friday.”
- “Meeting with Nathan tomorrow — prep some notes on the marketing plan.”
No formal structure. No app-switching. Just a rolling chat log of what I was doing, what I needed to do, and what I wanted to remember.
3. Use natural language — ChatGPT gets it
One of the best parts is that I don’t need to follow some strict task syntax or use a specific app interface. I just talk to it.
I can say:
“I won’t get to this today — bump it to tomorrow.”
Or:
“Scrap all the current dev tasks, let’s focus on the UI this week.”
And it actually remembers.
Even better, it connects the dots across other threads. So if I’m also chatting in a separate “Portal Dev” project, it pulls that context into the Productivity thread and reminds me about outstanding items — like brainstorming a new deployment flow or fixing a login issue.
4. Drop in screenshots, bookings, anything
Planning a trip? I just screenshot my flight confirmation and hotel details and dump them into the thread.
Need to pack gear for an install or a client visit? I tell it what I’m doing, and it spits out a packing checklist — complete with “don’t forget the PoE switch.”
This feels like the most underrated part of using ChatGPT this way. You’re not just typing tasks — you’re building context. The AI helps synthesize it all.
5. But here’s where it still drops the ball
It’s not flawless. Sometimes it gets a bit ahead of itself.
Example: I told it that “clearing email” was a daily task. And some afternoons, it would cheerfully inform me that email was already done — when I hadn’t even opened my inbox.
Or take the Builders Plastic saga: I told it on Monday to remind me on Thursday to buy it. And it reminded me on Tuesday. And Wednesday. And Friday. But not Thursday. (Classic.)
That said, it did nag me so much about the plastic that I couldn’t forget it — so maybe that’s a win?
6. Why it works (especially if you’ve got a chaotic brain)
Traditional productivity systems are too rigid. Calendar blocks don’t work if your day shifts constantly. And to-do lists become a graveyard of half-finished intentions.
What I like about using ChatGPT is that it feels conversational. I can ask:
- “What’s still outstanding?”
- “What should I prioritize next?”
- “What do we need to prep for tomorrow?”
And it answers like a colleague who’s been following along the whole time.
You’re not staring at a wall of overdue tasks. You’re just… chatting your way through your day.
Bottom line?
If you’ve tried every productivity method under the sun and none of them stick — give this a shot. Start a new project, call it Productivity, give it a simple job description, and then treat it like a PA who never sleeps.
It’s not perfect. But it’s surprisingly good.
Thanks for reading,
Shawn
PS. If you find this AI malarkey useful or interesting, consider jumping into our Skool community where I'll be building courses and more soon.