Bo Bennett On Building Bookyawards.com

HOST: It's good to have you back, Bo. Last time, you mentioned that the only real failure in your pipeline is producing a wonderful book that gets no sales. BookyAwards.com feels like the first tool in this series that's an answer to that, by providing external validation. But, if an AI wrote the book and an AI is judging it, what exactly is the Booky certifying—quality, or just AI capability? GUEST: Well, thanks for having me on again. The uh Book Awards is probably not the first tool in the series that uh that answers that question about getting more sales, but it's certainly one of the many. Uh you ask about an AI that writes the book and an AI that judges it. Well, keep in mind that AI isn't a singular thing. There's not just one AI, and that's where this program really shines. Well, there are many different AIs out there, and the type that we use to judge book judge books is different than the type that we use to write books. So, even in the I guess I would call it a rare case where an AI writes the book, and then they submit it for uh judgment. Then um it it's still a different AI doing the judging. So, it's going to compare that writing against other books that that it has in the platform and other books that it has read. So, there's there's value to it whether AI wrote it or not. HOST: Okay, so we're distinguishing between different types of AI here, not a singular entity, as you say. Can you tell me about the specific specialist AI judges you have on BookyAwards though, like Eleanor Whitcombe for literary fiction? Why have genre-specific judges in the first place? GUEST: Well, it was important to train the AI on the different genres, and rather than just having one judge that is a master of everything, it seemed better to have multiple agents that can each specialize in a certain group of genres. Therefore, we could train the agent. It doesn't have as large of a memory file, and it won't get as confused, and it could process more information, so it could make better judgments. So, that was our goal, and that was the purpose of doing it and setting it up this way. HOST: I get it, so less confusion for the AI by having them specialize. But, genre lines can be blurry, like literary thrillers or something weird fiction. How does the logic handle something that doesn't fit cleanly, and what happens if it gets it wrong? GUEST: It's not a big deal if it gets it wrong, uh, in unless it gives an award to the wrong category, in which case, if the author says something, we can change that category to make it a little bit more specific. But for the most part, it does a good job, and we do have a a generalist judge. So if AI really can't find a reasonable category, it puts it in that category, the generalist, and the generalist will do its best to to judge it, uh, accordingly. HOST: Okay, so there's that generalist judge to fall back on, kind of like a safety net. Now, moving to that "honest or free" guarantee, you're promising a refund if the judges can't find an award they'd stand behind. What percentage of submissions do you expect to fail that screen, and what does a refund-earning book look like? GUEST: Well, so far we're guessing at it's probably around 25%. Of of books that submit will will get I I don't want to use the word rejected. That's kind of harsh, but let's say that it AI could honestly say, yeah, this this isn't something that we could actually um give an award for. So, that's where that's probably the percent. And what does a book look like? I mean, honestly, it's not a bad book. I've seen some of the books that have been rejected, and I thought it was kind of harsh. Uh but that's that's the kind of quality level we're going for, and I'm okay with that. So far 25% is is reasonable uh within that 25% range. If it were to be any harsher, I would tell it to um ease up a little bit. If it were letting too many in, I would say, okay, you're letting too many in, change that uh algorithm a little bit to to be a little bit tighter on that criteria. But I think where we are right now is pretty good. HOST: Okay, so around a quarter of submissions are getting a pass because of that level of strictness, which you feel is about right. You mentioned that sometimes those rejections can feel harsh. Is that because the AI is prioritizing very specific elements like grammar over, say, character development or creative approach? GUEST: I think it takes a holistic approach and it looks at the quality of the cover, the quality of the writing, grammar, spelling, and then also the the prose itself and the writing. And and for example, if one of the categories is most interesting character, it may find something, but maybe due to all the other characters it takes off all the other characteristics of the book, it takes off the points and the book will not end up getting award. So it is important to have an overall good quality book. Everything does matter. HOST: So it's not just about one thing, whether it's grammar or a particular character, but the whole package: cover, writing quality, everything contributes to the final score. Now, looking at where this fits in the pipeline, after production and marketing tools like BookGist or Promoto, where does a Booky badge specifically work? Does it improve conversion on retail pages, or solve a different problem? GUEST: It does. It gives the book a little bit of credibility, um, social acceptance, sort of like reviews do. It's in that same category. Um, so that's where it's important in the the beginning of the process before the book is heavily marketed, it's good to have that award, um, not necessarily on the cover, but at least like on the web page promoting it, maybe mention it in the description. Um, that gives the the book credibility and helps the reader to move past that point of just reading the description versus actually clicking the buy now button and and buying it. HOST: Credibility and social acceptance, sort of like reviews, pushing the reader past simple consideration toward the purchase. Now, you've built tools that essentially let authors create the book itself, edit, design, and get it out there. With BookyAwards, you've built something that asks an outside evaluator—even an AI one—to say if the book is actually good. Is that the most honest thing you've built, or is it the most dangerous thing for authors who need the validation to be real? GUEST: Oh, I uh I like to think that everything I build is honest. Um in terms of the most honest, I I I don't know about that either. But let's say it it is an honest product in terms of the evaluation. And yeah, there always is that that push and pull between wanting to please the authors and also wanting to make money because if we don't accept their book for evaluation, then we don't get any money. So, that's not a good thing. Um but again, if we accept low quality books, then the the value of the award would be very low. And we don't want that either. That's the more strategic long-term plan. So, that's why that's more important than some quick cash. Uh in terms of the validation, I think that um I think that authors do appreciate when they when they say, "Hey, wow, this this super intelligence has evaluated my book and really came to the conclusion that these are pretty amazing characters." Or that this book deserves an award because of the suspenseful ending or whatever the category may be. And that is definitely validation. And and well worth it. HOST: It's definitely validation, and well worth it, as you say. Balancing long-term perceived value against short-term revenue by maintaining quality makes perfect sense. 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: As usual, archuboy.com or specifically for this site, we're talking about bookieawards.com.As usual, archuboy.com or specifically for this site, we're talking about bookieawards.com. HOST: Good to know. Thanks again, Bo, for joining us today and breaking down the philosophy behind BookyAwards.com. I especially appreciate you highlighting how its transparent process and genre-specific AI judges are designed to build trust in a world of black-box operations. And thank you out there for spending part of your day with us. Catch you on the next one—that's a wrap.