AI sounds like a shortcut until your “great” ideas flop in the wild, and that is exactly what Jay Schwedelson digs into here. He breaks down a simple 100,000 person prompt that quietly flips AI out of generic creative mode and into deeper statistical thinking, then shares what happened when he A/B tested it on subject lines and CTAs. Plus, a surprisingly relatable question about hating phone calls turns into a legit strategy for staying connected without awkward calls or endless text threads.

ㅤ

Best Moments:

(01:00) Alyssa from Denver asks why AI generated ideas keep underperforming, and Jay flags the hidden problem with how most of us prompt these tools.

(01:47) Jay introduces the 100,000 person prompt and shows how to ask AI to simulate huge audiences reacting to subject lines, landing pages, and offers.

(03:05) Jay explains how framing prompts around 100,000 real people forces AI out of pure creativity and into deeper statistical mode for sharper recommendations.

(03:45) Jay shares A/B test results where 100,000 person prompts beat generic AI suggestions over 70 percent of the time and even double click throughs on CTA buttons.

(05:15) Jared from Dallas admits he hates phone calls and finds texting shallow, and Jay gets real about how easy it is to drift into isolation.

(06:00) Jay reveals his go to solution of voice memos as a low pressure, high connection way to maintain relationships without live calls or walls of text.

ㅤ

Prompts mentioned:

• Predict how 100,000 real subscribers would respond to each of these subject lines, rank them by expected open rate, and explain the psychology.
• Evaluate this hero section as if 100,000 new visitors landed on this landing page. Identify confusion points and drop-off risks.
• How would 100,000 consumer buyers or 100,000 B2B buyers interpret this offer? What is confusing, what is strong, and what is missing?
• We want you to simulate 100,000 people interacting with this call-to-action button. What should the language be?

ㅤ

Check out Jay’s YOUTUBE Channel: https://www.youtube.com/@schwedelson

Check out Jay’s TIKTOK: https://www.tiktok.com/@schwedelson

Check Out Jay's INSTAGRAM: https://www.instagram.com/jayschwedelson/

Jay: we are back for Ask us anything from the do this, not that podcast. This is our short episode. We're all week long. We get in questions, we getting in work, questions. We get in ridiculous questions. We try to tackle one of each. And if you wanna submit a question, well that would be cool. All you gotta do is go to jay sch shettleston.com.

Jay: There's a button that says podcast, another one that says, ask us anything. And that's where we get our questions and awesomeness and all the things. Let's do the work question first. Okay. We got a question from Alyssa from Denver. I, I'm such a bad skier. You have no idea. I am like Green Hill, but scared on a green hill.

Jay: That's not the point of your question. What's the question? Alyssa? JI keep getting ideas back from AI tools, uh, that I use, and then I run the test of what they give me back and the results are not great. What am I doing wrong? Well, that's an interesting one, and we've uncovered something that I don't think enough people know about to do, which is kind of a game changer and can radically increase the performance on all of your marketing by doing one simple thing.

Jay: It doesn't matter if you're using Claude or Gemini or Chate or Perplexity or whatever. So what happens is when you go to any of these AI tools, you might be asking it to give you or helped you write a blog, helped write an email, a subject line, A A a A, a podcast title, anything. You go to the AI tools and you say, Hey, can you help me with blah, blah, blah.

Jay: Okay? And then it gives you back an answer. That's not what you wanna do. Here's what we've uncovered. When you go to any AI tool, instead of just saying, Hey, can you help me do this? You say, you use what I, what we call the 100,000 person prompt. When you use the 100,000 person prompt, you get exponentially better things back and we've ab tested it and the stuff that you get back performs exponentially better.

Jay: So what do I mean by that? For example. Let's say you wrote this to one of the AI tools. Predict how 100,000 real subscribers would respond to each of these subject lines, rank them by expected open rate, and explain the psychology or evaluate this hero section as if 100,000 new visitors landed on this landing page.

Jay: Identify confusion points, drop off risks, whatever. Or it could even be about an offer. How would 100,000 consumer buyers or 100,000 B2B buyers interpret this offer? What's confusing, what's strong, what's missing? And here's what we found, which is pretty wild when you ask any of the AI tools to take on this view of a hundred thousand people.

Jay: Okay? Interacting with whatever it is you're asking to create. What happens is it switches the AI tool switch from creative mode to statistical mode automatically, and it's when it simulates a hundred thousand people reacting to something, it goes deeper in its research, deeper in its thinking. And what we have found, which is pretty wild, is for example, subject lines.

Jay: We took subject lines. It gave as a result of. Just asking it normally versus subject lines. It gave as a result of simulating a hundred thousand readers and we, and then we would send both out AB tests and we found over 70% of the time, the one that was simulated with the a hundred thousand readers did better than just the generic one.

Jay: Okay. The same thing worked, for example, for call to action buttons when we asked. Any of the AI tools, uh, hey, what is the right call to action button? Words to put on the call to action button for this offer. And then we said, uh, we want you to simulate a hundred thousand people interacting with this call to action button.

Jay: What should the language be? When we tested it head to head, we found that the a hundred thousand person prompt generated a two x increase in click-through rate. Versus the prompt, uh, response that it got when we didn't ask it to do the a hundred thousand thing. So whatever it is that you're asking AI to do, if you ask it to predict how a hundred thousand real subscribers predict how a hundred thousand real readers.

Jay: Simulate a hundred thousand people, blah, blah, blah. It radically changes the depth and the accuracy and the level of information that you're getting back, and it doesn't cost you anything. So if you've never tried that, I strongly recommend it. It's kind of cool. All right, let's get on to the ridiculous question.

Jay: Okay. Jay, what do we got? I am Jay. What's going on? This is a question from Jared from Dallas. Your cowboys stink. Although they did win last week, so, but they stink. No offense. Sorry. I'm mean, I hate for that, but they kind of stink. Anyway, Jay, I hate talking to people on the phone by the way. So do i. I hate talking to people on the phone, but texting is annoying and hard to have relationships that way.

Jay: I don't want to be isolated. What can I do? Oh wow. This is like a deep question. Uh, I relate to this and I actually have a solve for this. I don't like talking to people on the phone. Either and texting is annoying 'cause I never know how to end the conversation. Right. You know, we go back and forth, back and forth and I really, I'm not able to get enough words out.

Jay: I don't feel like typing. It's all very, very annoying to me. So I have migrated something this year, which has been awesome and I recommend it to everybody. Which is voice memos. So I'm a big voice memo person. That's when you click on that like little orange button on your phone, and then you can talk as if it's a voicemail and then you send it to the person.

Jay: And I would say that 90% of the conversations I have now are texting. I'm sending voice memos to people and they could be two minutes, three minutes, four minutes, whatever, and then they voice memo me back. Okay. And the cool part is, it's like. Being able to have a phone conversation that you can end and also you feel like you're having a relationship with the person because you're hearing their voice and all the things and it's very long and you listen to it on when you wanna listen to it.

Jay: I am telling everybody out there, if you are not on Team Voice Memo. This is awesome, especially if you, I mean, some people like talking on the phone. I don't know. I get in my car, my wife's always like, call your friends on the way home from work when you're driving home. I'm like, I don't wanna do that. I wanna listen to a horrible podcast or some cheesy eighties music.

Jay: I don't wanna call anybody. And then I wind up not calling people, but the voice memo is the happy middle ground. So. No one's gonna do it. No one's gonna listen to me. Everyone's gonna say I'm a dork. Uh, but I'm telling you, this is pretty cool. What's not cool is me. I'm not cool. What else do I have to tell everybody?

Jay: I don't know. Uh, I was just thinking about Stranger things. Am I gonna watch Stranger Things? That's a hard pivot. Um, I don't know. I gave up after season two and everyone's telling me, hurry up and watch season three and four. I don't think I care at at all. It went off the rails. What am I talking about? I don't know.

Jay: Anyway, appreciate you being here. Try out that a hundred thousand thing. It's pretty cool, and uh, I'll check you later.