I write better business plans than AI can, and it’s not even close.
A founder sent me his business plan last week. Eleven pages. Clean formatting. Not a single typo.
He’d generated it in two minutes with ChatGPT.
I read it twice. Market sizing, check. Competitive landscape, check. Go-to-market section that used the word “synergy” unironically — check, unfortunately. On paper, it looked like a business plan.
It wasn’t one.
The Reality Check: AI didn’t write a business plan. It wrote about a business plan. Those are two different documents, and only one of them raises money.
Here’s the thing nobody tells founders before they open a fundraising round: a business plan isn’t really a plan. It’s a written record of a hundred decisions you already made in your head, long before you opened the laptop. Your CAC assumptions aren’t a number you looked up — they’re a number you’re braced for, because you watched one blow up a company you ran. Your churn projections aren’t optimism. They’re the specific memory of the month everything went sideways and you had to explain it to your board.
AI doesn’t have that memory. It has the internet’s memory, which is a different thing entirely. It can tell you what a good SaaS churn rate looks like. It can’t tell you what it feels like to watch yours double in a quarter and have to decide, in real time, whether to cut headcount or cut runway.
I spent 20+ years in banking and fintech. I’ve raised $800M for companies. I’ve launched over $300M in products. That’s not to brag, it’s to make a point. My first startup had members in 45 countries and I shut it down two years later — GoGrabLunch, if you want the name attached to the scar tissue. Every plan I write now has that failure sitting quietly underneath it, whether I mention it or not. It’s why I know which assumptions are the ones an investor will actually poke at. Not because I read a framework. Because I’ve been on the other side of that exact question, in that exact meeting, with worse numbers than the founder across from me thinks he has.
AI can regurgitate what’s been written about banking and fintech. It has never had to sit in a room and defend a plan it built that convinced a company to fund the largest product build in its 50+ year history. That’s a real thing I had to do. It has never had to live with the plan after the round closed, when the assumptions stop being slides and start being a payroll run.
That’s the gap. Not intelligence. Implementation.
So here’s what I actually tell founders: use AI for the parts of the plan that don’t require having lived your life. Market research — let it pull the data. Structure — let it organize your sections so you’re not staring at a blank page. Editing and grammar — let it catch the sentence you’ve read so many times you can’t see it anymore. That’s real, useful work, and I use it too.
But the story — the reason this business needs to exist, told in a way that makes an investor lean forward instead of skim — that has to come from you. Nobody can outsource the translation of the thing that’s been living in your head for two years onto a page in a way that convinces a stranger to write a check. That translation is the actual skill. Everything else is formatting.
I think about this the same way I think about training. You can hire a coach to build your program. You can have someone plan every set and rep down to the minute. Nobody can do the reps for you. The fight doesn’t care how good your plan looked on paper — it cares whether you actually did the work the plan was describing. A business plan is the same. It’s not the pitch. It’s the receipt for work you’ve already done.
So when a founder tells me AI can write their business plan in two minutes, I don’t tell them it’s wrong. I tell them it’s unfinished. The document exists. The experience behind it doesn’t — yet. And investors can always tell which one they’re reading.
If you want a plan that raises money, don’t ask what a good business plan sounds like. Ask what you’ve actually lived that nobody else in that room has.
Start there. Let AI clean up the rest.