Finding the Talk Only You Can Give

What this is for

Most speakers on a given topic say roughly the same thing, because they're drawing from the same research and the same handful of case studies everyone quotes. The talk that gets remembered is the one built from something only you lived through. This guide uses AI to help you dig that up - not to hand you a topic, but to ask you the questions that surface it.

When to use it

Use this early, before you've committed to a title or an outline. It works well when you have a topic area in mind but the angle feels generic, or when you're staring at a call for proposals and don't yet know what you'd say that's different from the last ten people who spoke on that stage.

The prompt kit

I speak on [topic area]. I want to find the specific angle only I can give - not a generic take on this topic.

Ask me questions, one at a time, to surface what's actually mine: a moment I lived through, a mistake I made, a belief I hold that goes against the grain, or a pattern I've seen that most people in this field haven't named.

Don't suggest topics yet. Just ask questions until we find the raw material.

Once you've surfaced two or three candidates:

Here's what came up: [list the moments, beliefs, or patterns from the last conversation]

For each one, tell me honestly: is this something only I could say, or is it something anyone in my field could say? Don't be diplomatic - if it's generic, tell me it's generic and why.

When you've picked a direction:

I want to build a talk around: [the angle you picked]

Ask me the questions a skeptical audience member would ask about this - the holes, the "yeah but" moments, the places where I need a real example instead of a claim. Don't answer them for me. Just ask.

What good output looks like

The AI should be asking more than it's telling. If three prompts in you still haven't answered a single specific question about your own experience, it's guessing at content instead of pulling it out of you. Good questions sound like: "What did you do differently the second time?" or "What's the belief here that most people in your field would push back on?" Those questions should make you stop and think, not just fill in a form.

Watch out for

Settling for the first angle. The first thing that comes up is usually the most obvious one, which means it's probably the one everyone else has already used. Push through two or three rounds of questions before you commit.

Letting the AI suggest a topic instead of asking about your experience. If it starts pitching talk ideas before it's asked you five real questions, stop it and redirect: "Ask me more before you suggest anything."

Confusing "interesting to me" with "only mine." A story is only yours if someone else in your field, given the same prompt, couldn't produce the same answer. Test every candidate angle against that bar.

Skipping the anonymizing step on client stories. If the moment that surfaces involves a client, employer, or coworker, strip identifying details before you type it into the chat. See guide 00 for the rule on what never gets pasted in.