Finding the Slow Parts

What this is for

Every talk has sections that move and sections that drag, and the person who wrote it is usually the last one to notice which is which. This guide uses AI to read for pacing specifically - not content, not structure, just where the energy of the talk rises and where it flattens.

When to use it

Use this on a full draft, after the structure and the honest read (guides 20 and 30) are done. Pacing problems are easiest to fix once you know the content is right - you're adjusting rhythm, not rewriting ideas.

The prompt kit

I want you to read this talk draft purely for pacing - where does it move and where does it slow down. Don't comment on content or ideas yet, just energy and momentum. Ask me anything you need about how I usually deliver a talk - pace, pauses, whether I use slides - before you start.

Here's the draft: [paste it]

Go section by section and flag anywhere the pacing likely drops: too much setup before a point lands, back-to-back sections with the same rhythm, a story that takes longer to tell than the point it's making is worth. For each flag, tell me why that section is likely to feel slow, not just that it is.

For the sections you flagged, what's the fix - cutting length, breaking up a long passage with a question to the audience, moving a slower section earlier or later in the talk? Give me one option per flagged section, not a full rewrite.

Now look at the whole talk end to end. Where are the two or three moments where energy should peak - the emotional high points? Do they land where they should in the running order, or are they buried behind slow stretches that use up the audience's attention first?

What good output looks like

Good pacing feedback names a specific reason for a slow flag - "this story has three setup beats before the payoff" is useful, "this part feels slow" is not. The fix suggestions should be small and targeted, not a request to rewrite the section. And the end-to-end pass should tell you something about how the whole talk is built, not just list individual weak spots - where the peaks are, and whether the talk earns its way to them.

Watch out for

Confusing slow with wrong. A slow section isn't automatically a bad one - some of the best moments in a talk are quiet on purpose, right before or after a peak. Ask the AI to distinguish "this is paced to be quiet" from "this is dragging with no payoff."

Fixing pacing by cutting content instead of restructuring it. The fastest fix isn't always the right one. Sometimes a slow section needs to move earlier in the talk, not disappear. Check guide 23 if the real problem is length, not placement.

Trusting a pacing read on a script you haven't spoken yet. Text pacing and spoken pacing aren't the same thing. Use this guide to find candidates, then confirm them by reading the draft aloud or recording a rehearsal pass - guide 40 picks up from there.