Section 1
What investors are actually underwriting
They are not buying the product. They are pricing a claim about the future that goes roughly like this: this market is about to reward something specific, this team has non-obvious information about what that something is, and the next tranche of money converts that information into evidence. So the story has to deliver four things in a form that can be repeated. The insight, meaning what you believe that most informed people do not, stated so plainly that a listener could disagree with it. The reason it is available now rather than five years ago, which is usually a change in cost, regulation, behaviour or infrastructure. The wedge: the narrow first market where you win outright rather than the large one you eventually address. And the evidence you have already gathered, whatever its size, presented with its size stated. A tight version of the same content, sized for a corridor rather than a boardroom, is covered in [Crafting Your Elevator Pitch as a Story](/blog/crafting-your-elevator-pitch-as-a-story).
Section 2
What consistently does not work
Top-down market sizing does not work, because everyone in the room has seen the same slide and knows the multiplication is arbitrary. A claim of no direct competitors does not work either, since the buyer is currently solving the problem somehow, and refusing to name that alternative reads as either inexperience or evasion. Hockey-stick projections without a mechanism do not work: the question is never whether you can draw the line, it is which input has to change and by how much. And selectively presented metrics do not work twice. Cohort data that starts at month three, a definition of active that is doing heavy lifting, a churn figure quoted on the wrong denominator. All of it is discoverable in diligence, and the discovery does more damage than the number ever would have.
Section 3
Building the version that survives retelling
Run your current deck through the model that follows. For each element, write the one sentence a partner would use in the Monday meeting. Anything you cannot compress to a sentence will not be carried into that room, and will simply be absent from the decision.
Section 4
Handling risk out loud
Name your three real risks before anyone asks, with the evidence that would resolve each and when you expect to have it. Concentration in one customer, a pending regulatory decision, a channel you do not control: these are already visible, and raising them yourself changes what the disclosure signals. The reason this works is not honesty as a virtue. It is that the investor's job in the room is to find the flaw. If you have found it first and can say what would settle it, you have moved the conversation from whether the risk exists to whether your plan for it is credible, which is a conversation you can win. Founders who conceal a known risk lose the meeting twice: once when it is found, and again when the partner reprices everything else they were told.
Section 5
The ask, sized to a milestone
Tie the amount to what it buys rather than to a runway figure. This much money, this specific set of hires and experiments, and the concrete claim that becomes provable as a result. An ask framed as eighteen months of survival invites the question of what happens in month nineteen. An ask framed as reaching a stated proof point invites a discussion of whether the proof point is the right one, which is the discussion you want. One more discipline: define your terms. Investors sit through pitches where platform, automation and intelligence each mean four different things in ten minutes. Plain definitions read as command of the subject rather than lack of ambition, and [What Is AI Automation? A Plain-English Guide for Founders](/blog/what-is-ai-automation-a-plain-english-guide-for-founders) is a useful model for how that sounds.