A Make It Ryne Playbook

The Overnight Ops Manager

Build a ChatGPT agent that researches your market every night and hands you a three-point morning brief before your alarm goes off. Setup time: about 15 minutes.

What you'll build

By the end of this playbook you'll have a scheduled ChatGPT agent that runs while you sleep. Every night it opens its own browser, hunts down new reviews of your business, checks competitor pricing, and scans for customer complaints. Then it writes you a brief with the three things to fix first.

No code. No Zapier. No new tools to learn. Just ChatGPT's built-in agent mode plus one prompt you'll customize once and never touch again.

Heads up: agent mode requires a paid ChatGPT plan. Plus gets you roughly 40 agent runs per month, which covers a nightly run with room to spare. Pro gets you around 400. The free tier won't cut it for this one.
1

Open agent mode

Log into ChatGPT on desktop (browser or app, either works). In a new chat, type:

In the message box/agent

Select agent mode from the menu that pops up. You can also find it in the tools dropdown next to the message box. Either way, you'll see the chat switch into agent mode, which means ChatGPT can now browse the web on its own, click around, and work through multi-step tasks without you babysitting it.

If you don't see the option, check that you're on a paid plan and that your app is updated. Agent mode rolled out to Plus, Pro, and Team plans, so it should be there.

2

Paste the overnight ops prompt

This is the exact prompt from the video. Copy it as-is first, then customize it in Step 3.

The promptAct as my overnight ops manager. Search for new reviews of my business, competitor price changes, and anything customers are complaining about, then write me a morning brief with the three things to fix first.

Hit send and watch it work. The agent will open a browser, run searches, visit pages, and pull everything together. First run usually takes a few minutes. Let it finish before you judge the output.

3

Make it yours (this is where the magic happens)

The generic prompt works, but a specific prompt works ten times better. The agent can't guess your business name or who your competitors are. Feed it the details:

Customized version (fill in the brackets)Act as my overnight ops manager for [BUSINESS NAME] in [CITY]. Every run, do these three things: 1. Search Google, Yelp, and Facebook for new reviews of [BUSINESS NAME] posted in the last 24 hours. Quote anything negative word for word. 2. Check the websites of [COMPETITOR 1] and [COMPETITOR 2] for price changes, new offers, or promotions. 3. Search for recent complaints or questions about [MY SERVICE TYPE] in [CITY] on Reddit and local Facebook groups. Then write me a morning brief with: - The three things to fix or act on first, ranked by urgency - One opportunity I should jump on this week - Links to everything you found so I can verify it

Swap in your real business, your real competitors, your real service category. The more specific you are about where to look, the less time the agent wastes wandering and the better your brief gets.

Pro move: the agent's output is a starting point, not gospel. It cites its sources, so click the links on anything you're about to act on. Trust, then verify.
4

Put it on a schedule

When the agent finishes its run, look for the clock icon near the response (you can also ask it directly: "run this every night at 2am"). Set it to repeat nightly.

That's the whole trick. ChatGPT's scheduled tasks feature will now fire this agent every night on its own, and the finished brief will be sitting in your chats when you wake up. You can also turn on notifications so it pings you when the brief is ready.

To manage or edit your scheduled runs later, go to your profile settings and look for Tasks (or just ask ChatGPT "show me my scheduled tasks").

5

Build the morning habit

An unread brief is worthless. Here's the routine that makes this pay off:

  • Read the brief with your coffee. Two minutes, max. It's already ranked by urgency.
  • Act on item one immediately. Bad review? Respond before 9am. Competitor cut prices? Decide if you match, beat, or hold.
  • Log the wins. Keep a running note of every catch the agent makes. In the video, mine spotted a competitor price cut before I'd heard a word about it. That's the kind of edge that compounds.

After a week, tune the prompt. If the brief keeps surfacing stuff you don't care about, tell the agent to drop it. If you keep manually checking something it missed, add it to the list. The prompt is a living document for the first two weeks, then it just runs.

6

Level up: from watching to doing

The nightly brief is level one. It watches and reports. Level two is agents that actually execute: reading your inbox, drafting replies, pricing quotes off rules you write down once. That's the stack I run with Claude and n8n, and fair warning, the first few weeks of building it were ugly. I broke it more than it worked.

Start with the brief. Once you trust it, you'll know exactly which task you want automated next, because it'll be the item that shows up in your brief every single morning.

One mindset shift before you go: the business doesn't need you awake. It needs the rules written down. Every prompt you write is a rule your business can run without you. Start writing them down.

The recap checklist

Total setup: 15 minutes tonight. Payoff: a morning brief waiting for you at 6am, every day, forever.

Want the rest of the stack?

This playbook is one of 50+ in the free AI Automation Starter Stack. Grab all of them and build the full system.

Get all 50+ free playbooks