The old-style business plan — 30–50 pages of tables, five-year forecasts, competitor analyses — is useless for a startup-stage founder. It's written for a bank, not for yourself. No one reads it twice.

But: a document that forces you to think through the essential questions is still one of the best tools. In the AI era, producing it is easier than ever — but quality collapses if you delegate to AI badly.

What an early-stage business plan needs

The minimum that survives self-challenge + possible investor scrutiny:

Problem and target group (1–2 pages)
Solution (1 page)
Market (1–2 pages)
Business model (1 page — Lean Canvas or BMC)
Competition and differentiation (1–2 pages)
Team (½ page)
Roadmap and milestones (1 page)
Financials and funding need (1–2 pages)

Total 8–12 pages. Enough for yourself, enough for grant applications, enough for an angel investor's first meeting.

What's NOT needed

Most things a "business plan template" asks for are pointless:

5-year forecast (nobody believes it, and you don't know)
Detailed SWOT in 4 boxes with abstract bullet points
Competitor table with 20 rows of which 18 aren't really competitors
"Mission statement" you wrote in 5 minutes this morning
Risk analysis that lists every theoretical risk without weighting

A good workflow with AI

Step 1: One-sentence story

Before writing anything else, write one sentence: "I help [target group] achieve [outcome] through [differentiator], because [existing problem with current solutions]."

E.g.: "I help first-time technical founders validate their concept in days instead of weeks through AI-assisted critical sparring, because existing tools are either too heavy (full consulting) or too light (random ChatGPT chat)."

If you can't write this sentence, don't continue with the rest. Go back to sparring.

Step 2: Lean Canvas first

Before writing long text, fill in the Lean Canvas. 9 boxes, one A4. In Innovaidor this is one of the methods.

If the Lean Canvas doesn't feel coherent — don't continue to the text. Fix the issues first.

Step 3: Write in human language

AI helps with drafting. But then write in your own words. Not "synergistic platform that catalyzes". More like "a tool with which you validate the idea with your customer in one afternoon."

Step 4: Numbers

Finance is the part where AI helps structurally but numerical accuracy is yours. A good early-stage finance section:

This year's monthly burn rate
This year's predicted monthly revenue (with base assumptions)
Break-even point or funding need until then
One alternative scenario (if growth is slower than assumed)

Step 5: Stress test

Give the finished document to AI and ask: "Act as a skeptical investor. What are the five hardest questions you'd ask based on this business plan?"

Answer the questions. If the answers are weak, fix the plan.

Pitfalls

1. AI writes, you publish. A reader (investor, advisor) recognizes AI text in seconds. It signals laziness. Use AI as a drafting tool, finish yourself.

2. Too-big document. Over 15 pages → no one reads. Keep to 8–12.

3. Numbers without assumptions. "Revenue in year 3 is €1M" → not believed. "Assumes a 0.5% conversion from a target group of 200,000 at an average price of €25/month." → believable or at least debatable.

4. Just one version. Build two: 2-page executive summary + 8-page full. Send the 2-pager first, give the 8-pager only when requested.

Closing

A business plan isn't a document for funding — it's a tool for your own thinking. Writing forces you to clarify what you've kept in your head. AI accelerates writing, but the thinking has to be yours.

Start: fill the Lean Canvas in Innovaidor this week. If any box is empty or unclear → there's a weakness to fix before writing a full business plan.