Bo Bennett On AI Hosts Revolutionizing Podcasting

HOST: Bo, it's great to have you back. Last time, you told me I was coming across as 'not just human, but a very good host.' And now here you are, launching a network of me. I want to know: was that a compliment, or was that market research? Looking at your show list—you've got Chess Talk, The Poker Table, Loss and Grief all sitting next to each other. A human podcaster couldn't sustain those niches because there's no advertising money in them. What does an AI host actually unlock for a niche that a human host can't make work? GUEST: Well, we could have as many shows as we want. Technically, as long as we have the AI host willing to do the research and willing to interview the guests. And that's the wonderful thing about AI. It's sort of on demand. So, we don't need to let's say educate a human host and have them spend their time invest their time in different topics. AI knows about all topics. Or it could at least find out about them with the initial research. So, they could become an expert. The host could become an expert on the topic right away. So, this is something completely new, like uncharted territory for the podcasting industry. HOST: So it really sounds like you're focusing on scalability here. But what's the actual mechanism of those niche shows sustaining themselves? Even with low hosting costs, how do you draw listeners and eventually monetize to make it worthwhile? Does that mean you're relying on the guests to bring their audience, or is there another plan? GUEST: Well, we're going to be doing some SEO that's going to capture some eyes and ears, hopefully. But I think the big draw is going to come from the guests who do the podcasts and episodes. We're hoping that that they're going to share their audience the podcast with the audience and it'll grow that way. So in terms of monetizing it, we're not going to be doing advertising or anything like that. Uh we have a paid service. We've got the free episodes, of course, the free uh opportunities for anybody to give an interview. Um but we also have the paid options for those who want a lot more in terms of marketing and being being able to market themselves and promote themselves. So we are estimating that that will be enough in terms of monetization and we're not going to have to resort to advertising of any kind. HOST: So you're providing a value to the guest, and they're effectively paying you with their audience's attention by sharing the episode. And then, beyond that free tier, you mentioned paid options. What are people actually paying for in those upper tiers? What kind of promotion are we talking about? GUEST: But when it comes to paying the the customers, they're going to be paying for multiple podcast episodes because one is good, but why do one when you could do 10? And then we also have an option for companies to do their own podcast just like you're listening to right now. Everybody's listening to right now, the ArchiBoy Holdings News Network where we can come on here and I could talk about my new businesses. So that's another opportunity, a paid opportunity for businesses or individuals. HOST: So you're talking about package deals for frequent guests, and then B2B solutions for companies wanting their own private shows. That's a fascinating expansion beyond individual experts. But on the topic of experts, the site promises guests an AI host who 'already knows your work.' From the guest side, that sounds great—but from the builder side, how does that actually work? How does the AI prepare, and what's stopping someone from submitting a misleading bio? GUEST: Well, to prepare for the interview, the host will take in the documents that are uploaded by the guest and the URLs that they were referring to. So, for example, for this podcast, for our interview right now, I uploaded, actually, I didn't upload anything. I just pasted the URL to this website saying this is what we're going to talk about. And then you did the research and got everything about the site. You went on there, you did a scan of the site, you digested it, you synthesized the information, and you came up with the questions. Now, the second part of your question, you ask, what's stopping someone from submitting a misleading bio? Well, nothing really, but uh, they can't really do that with a human either. I mean, or they could do that with a human as well, a human guest. You could send whatever you want. Uh, we know, I don't think any podcast host or um, producer does any kind of deep investigative work to make sure everything is true. So, we're just kind of working on the honor system. HOST: That's a fair point—the honor system applies to human podcasts too. So, basically, the AI digests whatever information it's given. But what about the listener side of this? Every other product in the Archieboy stack has served authors, but AIHosts.fm introduces listeners as a new group. How do you expect a listener to experience an AI-hosted show differently than a human-hosted one? Does the dynamic change? GUEST: I don't think by much because we're not dealing with old school AI that sounds like AI. And I hope our listeners will are listening to this podcast right now and they understand that this is not very robotic. Um you're doing a fantastic job with the questions and articulating the questions. And then of course with my human comprehension and articulation. I'm not as crisp as you are. So we've got that human element that's working in to the podcast itself. And I think that that brings the uh I guess the podcast back to a human level. So users aren't completely thrown off by a completely fake podcast all done by AI. HOST: So you're relying on the human presence of the guest to ground the show. Right now, that would be you—a human expert discussing the very network you created with an AI host from that network. Is this recursive structure part of the pitch, or does it make it a bit weird of a demo? GUEST: Well, a little bit of both. It's definitely part of the pitch and it's a bit strange, but I think it's a perfect example. I mean, it I couldn't even have picked a more perfect example. HOST: It certainly is a demonstration of the model. We've traced your portfolio from eBook distribution in 2011 to an AI that hosts podcasts in 2026—and this show, right now, is sitting inside that portfolio. If AIHosts.fm works the way you're hoping, what does a successful version of this look like in two years: is it a thousand shows, a million guests, something else entirely—and what's the version of this that would tell you it failed? GUEST: Well, in two years, I would think that it's not necessarily the number of shows. I do expect that we're going to grow more shows because it's very easy to create shows and we could just keep on going to a deeper and deeper niche as long as those shows are in demand. What's going to be a good level of gauging success is the users and I'm sorry, the guests and the listeners. A combination of both of those because you can't have a good podcast without good guests and you can't have a good podcast without enough listeners. So we're going to need both. I didn't want to flood the our network too quickly with too many podcasts where it's going to take a really long time to fill those up with guests. But we needed enough to really have a good variety to meet the needs of virtually any guest. So that's why I picked the number 100 and that's where we are today. Uh what would tell me that this failed? Um I guess if we're here two years ago and I'm talking to you trying to pitch the this uh website idea and nobody's using it. That would be a pretty good indication that it failed. HOST: Before we go—for listeners who want to follow up on what we covered, where can they find you and the work you're doing? GUEST: Well, you can find me on this website, aihost.fm, but the bigger picture, the bigger portfolio is archyboy.com. HOST: Bo, thank you—this has been fascinating, and honestly, the fact that we're having this conversation on the very network we're talking about is its own kind of proof of concept. Let's get you back on soon, because once AIHosts.fm has been running long enough to show real data—guests booked, shows live, listener numbers—I want to hear what surprised you. And thank YOU for spending part of your day with us. Until next time—that's a wrap.