The Recall Simulator - Use AI to Predict What Your Audience Will Remember | The So-What

Ultimize. I'm about to hand you the tool I'm most excited about in this whole series, and it rebuilds the most powerful thing I've ever taught, except now you can do it in the time it takes to pour a cup of coffee. For 30 years, my favorite test for any presentation has been brutally simple. Before you give it for real, you record yourself giving it and you send that recording to 3 or 4 people who resemble your actual audience. You ask them to take it in once. Then you call them and ask them one question. What do you remember? Not, did you like it? Not, was I good? Just, what stuck? If they hand you back the messages you intended, you're ready. If they hand you back a blank stare and say something like, you seem confident, you've got work to do. And better to learn it now than in front of the people who sign your budget. That test is never lost. Because the audience is never wrong. If they didn't remember your point, it isn't their fault, it's yours, and it's fixable. The only problem with the test was friction— finding 3 or 4 honest people, imposing on their time, waiting a day or 2 for them to circle back. It was so much work that almost nobody actually did it. They agreed it was brilliant, and then they skipped it most of the time. That friction is now essentially gone, and that changes everything. Here's the move. Take your presentation, the script, the outline, the transcript, whatever you've got in words, and paste it into an AI tool. Then, and this is the part people skip, describe your audience in real detail. Not my team. Say something like, my audience is 8 mid-level operations managers. Most of them skeptical of new software after a painful rollout last year, who care most about not adding to their own workload, and who'll be taking this in at 4:30 on a Friday afternoon when they're already fried. The richer the description, the more useful everything downstream becomes. Now give it the test. Ask, you're a person in that audience. You've experienced this presentation once. It's Monday morning. What are the 3 things you actually remember? And what do you think I wanted you to do? Now read the answer carefully because you're about to learn something you can't learn from inside your own head. You wrote the presentation, so you already know what it's supposed to say, which makes you the one person on earth who can't hear it fresh. The simulator can. And when it hands back 3 things that are not your 3 things, you've caught a failure days before it would have happened for real, at zero social cost, with nobody thinking less of you. Sometimes what comes back is close and you tune. Sometimes what comes back is a polite fog. The presenter discussed several operational topics. And that fog is a gift. Because it's telling you the whole thing is landing as an undifferentiated blur. Either way, you now know before it counts. I do have to be straight with you, and I'll be much more so in a later lesson. This is a simulation. It's not your real audience. Machine gives you a plausible human, not the actual humans in your actual room. It's a rehearsal, not the performance. So we keep the original test too, the real recording sent to real people as the final check before you go live. But now you can run the simulated version 10 times first, fixing the obvious failures cheaply so that when you finally spend a real person's time, you're refining something good instead of discovering it's broken. So here's your assignment. Take a presentation you're working on, write a rich, specific description of your audience, Put real effort into that part. Then run the recall test and compare what comes back against your intended messages. The gap between what you meant and what a listener remembers has always been the whole game. For the first time, you get to see that gap before you step into the room. Use it.