Before you run these exercises
Chapter 8 is the most operationally dense chapter in the book — it's the one that tells you how to actually run a fundraise, not just prepare for it. There's a specific investor taxonomy here with a risk/patience matrix that changes how you should think about who to approach and in what order. There's a fundraising funnel with actual conversion numbers. And there's a section on talking big without lying that draws a precise line between ambition and dishonesty — one that most advice on fundraising gets completely wrong.
This chapter rewards close reading before you start the exercises. Several of the AI prompts below reference specific frameworks from the book — the investor ICP concept, the 3-week timeline, the "Step on the Gas" test — that will land differently once you've read the chapter and understand the logic behind them.
The chapter defines the three filters every investor must pass — right stage, right category, right geography. But before you can apply those filters, you need a list worth filtering. And the sequence matters enormously: AI-generated lists are hypotheses. Data you pull yourself is raw material. Founder-verified lists are signal. Use them in that order.
Start with structured data sources that reflect what actually happened — not what AI remembers from its training data. Then layer in human intelligence from founders who actually ran the process. Then use AI to analyze all of it. The mistake most founders make is asking AI to generate the list and then trying to verify it. That's backwards.
Step 1 — Pull raw data from real sources
Step 2 — Add human intelligence
You now have a raw dataset — firms from Crunchbase, names from LinkedIn lists, investors from founder conversations. None of it is verified yet, but all of it has more provenance than anything AI would generate cold. This is the right moment to bring AI in: not to create the list, but to analyze it.
The workflow is research first, then verify the things AI can't see, then synthesize. Treating AI output as the end product skips the due diligence that actually protects you from walking into a meeting with someone who left their firm six months ago — or who has capital but no authority to deploy it.
Research each name
Verify what AI can't confirm
The chapter gives you a specific framework for creating urgency without faking demand: a three-week structured process with explicit go/no-go milestones at the end of each week. The power isn't in the deadline itself — it's in the fact that you actually follow through. Investors who are interested move fast. Investors who are slow-rolling you are telling you something.
A deal with no deadline never closes. The chapter lays out the exact structure — Week 1 is high-level fit, Week 2 is deep dive, Week 3 is final alignment. Use Claude to customize this for your specific raise and draft the language you'll use to communicate it.
The chapter introduces a specific test for finding the right level of ambition in your projections — not what's conservative, not what's fantasy, but what's possible and defensible. The mechanic is simple: take your current projections and ask whether you could double them. If yes, do it. Then double them again. Keep pushing until you hit the point where you'd say to yourself: "No one would ever write us that check, and I wouldn't write it myself." That's your ceiling. The number just below it is what you should be pitching.
The chapter also draws a precise line between ambition and dishonesty — one that most founders get wrong in both directions. Some sandbag. Some overstate. The distinction is between selling possibility and claiming certainty. Read the chapter for the full argument, then use this to find your number.
The chapter is explicit that fundraising is just sales — and that means running it like a sales process, not a series of hope-filled coffee meetings. That means a pipeline with stages, conversion tracking, next steps with dates, and a systematic way to cut off what isn't moving. Granola captures what actually happened in each meeting. Claude turns that into pipeline hygiene.
After each investor meeting, the workflow is: Granola summary → Claude debrief → pipeline update. Over time, the pattern across meetings tells you where you're losing deals and what to fix.
Before the meeting — set your intentions
After the meeting — debrief against your intentions
Chapter 8 reframes fundraising as a process you run, not an outcome you hope for. The founders who close rounds aren't charming their way to yeses — they're working a funnel, managing a clock, and creating conditions where yes is the obvious answer.
Before you move on: have a qualified investor list with genuine ICP fit (not just "they have money"), a structured timeline you'll actually use, and a financial story you've pushed until it broke and then pulled back one step. And know which investor types are currently expressing interest — because misaligned money on your cap table is a problem that starts the day you take the check, not the day things go sideways.