Anir SurenI ran a little experiment. I opened one ChatGPT chat and, without switching tabs or starting over, walked it from "I have no idea what business to start" all the way to a landing page, sales copy, a payment link, and my first 30 pieces of content. Four prompts to plan it, one to build it. What it came back with genuinely blew me away, so I cleaned up the exact prompts I used and put all of them here, plus three bonus ones most people never think to run.
Here's the whole thing you can copy. Run them in order, in the same chat, and don't skip the setup notes on the last one.
Why it all has to happen in one chat
This is the part that makes it work, and it's easy to miss. When you keep everything in a single conversation, ChatGPT remembers your skills, the business it picked, the price, and the roadmap as you go. Each prompt builds on the last one instead of starting from a blank page. Open a new chat for the roadmap and you're introducing yourself all over again to something with amnesia. Same chat, top to bottom. That's the trick.
Prompt 1: the top 3 businesses you could actually start
This one does the thing most people are stuck on, which is just picking. Notice it's told to interview you first. If ChatGPT already has memory of you it'll use it, but forcing it to ask means the ideas are built around your real life instead of a generic list.
Act as a business strategist who has personally launched over 100 profitable one-person businesses. You are sharp, honest, and allergic to generic advice. First, ask me anything you still need to know: my skills, my personality, my interests, how many hours a week I have, how much money I can put in to start, and what I absolutely do not want to do. Wait for my answers. Then, based on everything you know about me, give me the top 3 businesses I could realistically build to $10,000 a month within 90 days. For each of the 3 ideas, give me: - The one-line pitch - Why it specifically fits ME and my situation - Exactly what I would sell and who buys it - A realistic price point, and how many sales a month it takes to hit $10k - The single biggest reason most people fail at it, and how I avoid that Rank them from highest to lowest probability of me actually reaching $10k, and be honest if my most realistic option is harder than it sounds.
Prompt 2: pick the winner and break it down
Now you make it commit. Instead of leaving you with three maybes, this forces it to choose the one it would bet its own money on and pressure-tests the offer, the buyer, and the price until they actually hold up.
Pick the single highest-probability business from the 3 you just gave me, the one you would bet your own money on, and break it down completely. Stay in character as the strategist. Give me: - THE OFFER: exactly what I am selling, in one clear sentence a stranger would understand - THE BUYER: who they are, where they already spend time online and offline, and the exact painful problem they will pay to make go away - THE PRICE: what I charge, why that number, and whether it is one-time, monthly, or a package - WHY THEY PAY: the real reasons someone hands over money for this instead of doing nothing or doing it themselves - THE MATH: how many customers at that price it takes to reach $10k a month, and whether that is realistic in 90 days - THE RISKS: the top 3 reasons this could flop, and what I do about each one now Do not be a cheerleader. If the price is wrong or the buyer is too hard to reach, say so and fix it.
Prompt 3: the full 90-day roadmap, day by day
This is where it stops being an idea and becomes a to-do list. The important line is the one that tells it to put you in front of customers in week one, so you're not building in a cave for a month before anyone can pay you.
Now build me the full 90-day roadmap to take this business from zero to $10,000 a month. Break it into 3 phases of 30 days each, then week by week inside each phase, then the specific daily tasks. For every week, tell me: - The one goal that week is built around - The exact daily tasks, kept concrete (like "message 10 people who fit the buyer profile," not "do marketing") - What "done" looks like by the end of that week - The number I should be tracking (leads, calls, sales, revenue) and the target for it Front-load it so I am talking to potential customers and making offers in week one, not building for 30 days before I show anyone anything. Assume I have limited time and limited money, so cut anything that does not directly move me toward paying customers. End with the 5 metrics I should check every single week to know if I am on track or off track.
Prompt 4: turn on agent mode and let it build
Here's the part that feels like cheating. Everything above was planning. This one actually produces the assets. You need to switch ChatGPT into agent mode first: click the + in the message box (or type /agent) and choose Agent mode. It's available on the Plus, Pro, Team, Business, and Enterprise plans. Once it's on, send this.
You now have everything: the business, the offer, the buyer, the price, and the 90-day roadmap above. Go build the starting assets for me. 1. LANDING PAGE: build a simple, high-converting one-page landing page for this offer. Write all the copy and structure the sections (headline, subhead, the problem, the offer, proof, FAQ, and one clear call to action). Give it to me as a file I can publish, plus step-by-step instructions to put it live for free. 2. SALES COPY: write the main sales section that does the actual convincing, in a clear, honest voice that matches how my buyer talks. 3. PAYMENT LINK: walk me through setting up a payment link for this exact price using a free tool like Stripe, and tell me the precise settings to use. 4. FIRST 30 PIECES OF CONTENT: draft 30 pieces of content that attract my exact buyer, each with a hook, the main point, and a call to action that leads back to the offer. Number them and organize them so I can post one a day for the first month. Do as much of this yourself as you can. Where you genuinely need me to click something or enter my own account details, stop and give me the exact steps instead of guessing.
Agent mode actually goes and works, opening a little browser, writing the files, and drafting the content while you watch. Give it real time on this one. When it's done you're not looking at advice, you're looking at the starting pieces of a business.
Three bonus prompts most people skip
The four above are the core. These three are what separate the people who post about the plan from the people who actually get paid. Run them in the same chat.
Bonus 1: pressure-test the whole thing before you waste a day
Before you commit 90 days, make it try to kill the idea. This one has saved me from more bad plans than anything else.
Before I spend a single day on this, pressure-test the whole plan. Play three roles, one at a time. First, the harshest version of my ideal customer: tell me every reason you would NOT buy this. Second, a competitor who already does this: tell me how you would crush a beginner like me. Third, me in 90 days having failed: walk backwards and tell me the most likely reason it did not work. Then give me the 5 changes to make right now that kill the biggest risks, ranked by how much they matter.
Bonus 2: land your first 10 customers in 14 days
$10k a month starts with customer number one. This gets you a concrete outreach plan with the actual message to send, so you're not staring at a blank screen wondering how to start.
Give me a plan to land my first 10 paying customers in the next 14 days without spending money on ads. Tell me exactly where my buyers already gather, the outreach message to send them (write it word for word so it does not sound like spam), how many people I need to contact per day to get 10 sales, and how to handle the 3 most common objections for this specific offer. Include a simple, free way to book calls or take orders.
Bonus 3: your daily operator
Keep this one going every morning. It turns ChatGPT into the co-founder who tells you the three things that actually matter today instead of the twenty that don't.
Act as my daily operator for this business. Every morning I will paste in what happened yesterday: leads, replies, sales, and problems. You give me: - The 3 highest-leverage tasks for today - Anything that is falling behind the 90-day roadmap - One small experiment to try that could move revenue Keep it short and specific, no fluff. Ready when I am.
The honest part
I'm not going to tell you a chatbot hands you a running business while you nap. Here's the real version. ChatGPT plans a genuinely good business and, in agent mode, drafts the landing page, writes the copy, produces the 30 posts, and walks you through the payment setup step by step. That's a real head start, the kind that used to cost you weeks or a few thousand dollars in freelancers.
What it can't do is be you. It won't create your accounts, publish the page, take the payments, or do the outreach on its own, so you still connect the last mile. And the plan is only as good as what you feed it, so the more honest you are in prompt one, the better everything downstream gets. The $10k in 90 days is aggressive and depends entirely on you executing the roadmap, not on the AI. Treat this as the fastest possible way to go from nothing to a real starting line, then do the reps.
The hard part of starting a business was never the plan. It was the blank page. That's the part this quietly deletes.
Open one chat and run the first prompt tonight. Even if you never launch, watching it map a real business around your exact life is worth the ten minutes. And if you do launch, you'll be a month ahead of where you'd have started.
– Anir
If this is the kind of AI thinking you want working on your own business, apply to work with me. And to catch the next breakdown, join the newsletter.