Spent the last few months building decks for agencies and founders, and there's a pattern that keeps showing up: people obsess over the design software and completely skip the thinking part.

A pitch deck isn't a design problem. It's a compression problem. You have 90 seconds to get someone from "who is this" to "I want to know more" — and most decks try to say everything instead of the one thing that matters.

The framework I keep coming back to when building decks (with or without AI):

1. One slide, one idea. If you need two sentences to explain a slide, it's two slides.

2. Lead with the tension, not the solution. Every good pitch has a "wait, that's a problem?" moment before the "here's how we fix it" moment. Skip the tension and your solution feels irrelevant.

3. Numbers need context, not decoration. "$2M ARR" means nothing without "growing 40% quarter over quarter with 3 people." Context is what makes a number persuasive.

4. The last slide is the ask, always. Sounds obvious. Most decks bury it or forget it entirely.

Been using Claude to help generate first-draft content and visual direction for decks lately, and the biggest unlock isn't speed — it's that AI is brutally good at telling you when your narrative doesn't hold together. It doesn't have ego about your idea, so it'll surface the gaps you've been ignoring.

Curious what frameworks other people here use for structuring pitches, especially for client/sales decks vs investor decks — feels like a different game.