I Asked AI to Build My $50M Plan. It Took Seconds.

I gave it a goal. It gave me the whole plan.

If you're anything like me, your biggest goals live in your head as a feeling. Not a plan. A feeling.

Mine was a number. There's a business unit I help run at my day job, and the goal was to take it to $50 million. I knew roughly where the money came from. I did not know the exact path.

So I tried something. I connected Claude to my email and our CRM (the system where every client and every deal lives) and asked one question: how do we get to $50 million?

It read years of history in seconds. Then it handed me the plan. The path. The data. The statistics. The playbook.

The kind of analysis a consultant bills for over weeks. I had it in less time than it takes to pour a coffee.

The whole move in five lines

1. Pick one big number. Revenue, a raise, a department target. One.
2. Connect AI to where your history lives. Email, CRM, Shopify, QuickBooks, in a protected, approved account. Or just upload a spreadsheet export.
3. Give it a real job title. Not "assistant." Chief strategy officer.
4. Ask it to reverse engineer the goal. What has to be true, and by when.
5. Ask for the playbook. The next 90 days, week by week.

Steal this prompt

Speak it if you can. Voice mode gets you more context than typing ever will.

Act as a chief strategy officer who has spent 20 years taking businesses from where mine is to where I want it to go. You can see my data. My goal is to hit a specific number by a specific date, and I'll tell you both.

Before you build anything, interview me. Ask one question at a time, five total: what counts toward the goal, what's off limits, my team and budget, what has worked before, and what I'm most worried about.

Then go through the last two to three years of my data and give me:

1. Where the money actually came from, ranked (customers, products, channels).
2. The gap between today and the goal, broken into quarterly targets.
3. What has to be true to close it: how many deals or sales per month, average size, conversion, and how many customers we keep.
4. The three biggest levers and the two biggest risks.
5. A 90 day playbook, week by week, including what I personally do.

Show your math. Flag anything you're guessing. Then tell me what would make this plan fail.

Before you connect anything

If you're bringing in customer data, financial data, or anything about your business, do it in a protected environment.

• Work for a company? Check your company's AI guidelines and talk to IT before you share anything sensitive.
• Run your own business? Use a paid plan with business privacy protections, not a free personal account.
• Everyone: never put other people's personal information into an AI tool you aren't authorized to use.

What came back

The path. Quarterly targets that actually add up to the number. No more "we'll grow a lot."

The data. Where the revenue really came from, ranked. A few surprises in there, the kind you only see when something reads every email at once.

The statistics. The math working backward from the goal: how many, how big, how often.

The playbook. What to do this week. Then next week. Then the week after.

Here's the part that got me. None of it was magic. The answer was already sitting in my inbox and our CRM. I just never had the hours to read three years of it.

Your number, your version

You don't need a $50 million goal for this to work. You need a goal and some history.

You run your own business. Connect Shopify, Stripe, or QuickBooks plus your email. Ask: what does it take to double revenue next year? Which products, which customers, what has to happen every week?

You're going for the promotion. Connect your email and calendar. Ask it to pull your last twelve months of wins, projects, and praise, then build your case for the title and the raise. With numbers.

You run a department. Connect the CRM or project tool your team lives in. Ask where deals or projects stall, who your best accounts are, and what the 90 day plan looks like to grow 30 percent.

You're building something on the side. Skip the connections. Export a spreadsheet, upload it, and ask the same reverse engineering question.

Same prompt every time. Only the number changes.

The honest miss

The first plan was too clean. It assumed every big client would renew and every deal would close on schedule. Anyone who has ever run anything knows that's not how a year goes.

The fix was one line: "Assume 20 percent of the pipeline slips. Rebuild the plan." The second version was the real one. Tighter, more honest, and a lot more useful.

And it can't read the room. It sees the emails, not the relationship behind them. That part is still your job, and it's the part you're best at.

Insider stuff

• Check your company's AI rules first. For work data, use the AI tool your company approves. Describe the work, not the deal.
• Ask for the math, not just the answer. A plan you can't check is a plan you can't defend in a meeting.
• Make it argue with itself. "What would make this plan fail?" is the most valuable question in the prompt.
• Keep the same chat. Once a month, tell it what actually happened and ask it to update the plan. It gets sharper every round.
• Put the playbook on your calendar. A plan in a chat window is a nice idea. A plan in your calendar is a quarter.

The real win isn't the plan. It's the clarity. You stop carrying the goal around as a feeling and start working it as a list. That's time and headspace you get back, at work and at home.

New to connecting tools? Start with how I use Claude, then see why I use Claude and ChatGPT together. Want more time back? Make AI your chief of staff.

The quick version

• One number, one date.
• Connect your data safely, or upload an export.
• Chief strategy officer, interview first.
• Reverse engineer it, show the math.
• Stress test it, then calendar it.

FAQ

Do I have to connect my email and CRM?

No. Connecting is fastest, but a spreadsheet export works. Upload it and use the same prompt.

Which AI is best for this?

I use Claude because it connects to my tools and handles long, complicated data well. ChatGPT and Gemini can do versions of this too. Try your plan in two and compare.

Is it safe to connect work data to AI?

Only in a protected environment. If you work for a company, follow its AI guidelines and check with IT first. If you run your own business, use a paid plan with business privacy protections. Never share other people's personal information you aren't authorized to share. For personal and family data, here's how I think about AI safety.

What if I don't have years of data?

Start with what you have. Even three months shows a pattern. Ask it to fill gaps with industry benchmarks and to label which numbers are yours and which are estimates.