Insights for Modern Publishing

Episode 71: How AI Book Companions Are Changing the Reading Experience

Written by firebrandtech | Nov 16, 2026, 2:39:20 PM

Wendy Kier is the founder of EmpowerAi™ and creator of the AI Book Companion™. Having developed more than 100 specialist AI agents, Wendy now works with authors to explore how artificial intelligence can extend the experience of reading a book. Her work focuses on developing AI Book Companion™ experiences that allow readers to ask questions, explore ideas and engage more deeply with an author's content, while respecting the author's voice, intellectual property and the boundaries of their expertise. Through developing and testing these experiences with authors and readers, Wendy has become particularly interested in the changing relationship between books, readers and AI, and the opportunities and responsibilities this creates for the publishing industry.

On this episode of the BookSmarts podcast, Wendy and Joshua discuss how interactive AI tools are transforming the way readers engage with books—from Amazon’s Kindle features to custom-built AI agents. They explore why human nuance, personal storytelling, and reader communities remain irreplaceable in the age of generative AI, how authors and publishers can establish strict boundaries and frameworks to keep AI companions accurate and safe, and how interactive book agents create new opportunities for reader connection and monetization.

Links Referenced:

https://wendykier.com/

https://tim.blog/2026/06/12/has-ai-already-killed-nonfiction/

https://goodereader.com/blog/kindle/here-is-all-of-the-ai-features-on-the-new-kindle-scribe

 

Digital Transcript

Joshua Tallent

Hello and welcome to the Booksmarts podcast, where we talk about publishing data and technologies, and send you away with some insights that will help you sell more books. I'm your host, Joshua Tallent. So this month on the Booksmarts podcast, I'm chatting with Wendy Kier, who is the founder of Empower AI and the creator of the AI Book Companion. Wendy has developed more than 100 specialist AI agents, and she now works with authors to explore how artificial intelligence can extend the experience of reading a book. Wendy, thanks for joining me on the podcast.

Wendy Kier

Thank you for having me.

Joshua Tallent

It's great to chat with you. AI obviously is a really big topic, and authors are kind of concerned and trying to figure out what's next and how do we adapt to it, or how do we push it away further. All those questions are obviously in the minds of a lot of authors today. So I want to start our conversation today by talking about a specific genre: in 2026, Tim Ferriss wrote this blog post about how AI chatbots have impacted his own nonfiction self-help book series and his sales pretty dramatically in just a very short period of time. AI can already answer a lot of the questions people once turned to self-help books for. What do you think that might mean for authors and their readers?

Wendy Kier

I haven't read the article, but it's good in one respect, and not good in another, that it's hitting his bottom line and his sales so significantly. I think when it comes to writing, we need to be crystal clear and be able to give the human experience—the human nuance of the experience—because AI cannot do that. We need to be able to tell the human side of the story, so we need to be crystal clear on who our audience is, what our niche is, what our market is, and also the personal stories that we can wrap inside that.

I read so many books, and sometimes you just can't get a grasp and a sense of who that individual is. It's almost like they're writing, but they're almost hiding at the same time. AI can be very good at—how can I say it?—it sounds right, but it's not quite right. One way to make a distinct difference between the two is to really give the human side of the story, to get deep down and deep personal. Obviously, don't leave people hanging there, but you want to really put as much of yourself into the story and the book as possible because AI can't actually do that bit for you.

Joshua Tallent

Yeah, and with self-help, you really want to have a conversation with the reader about whatever the self-help topic is from a personal perspective. If you can say, "I struggled with this area of my life, here's how I tried to approach it, how I fixed it, and what resources I went to," there is that personal approach, especially with self-help. Obviously, with AI tools, you can go talk to ChatGPT and have a conversation with the AI chatbot—even a verbal conversation—and talk about your feelings. All it's doing, though, is responding to you from the massive amount of data that it has, giving you responses that it thinks you might need to hear. That's very different than reading a book written by somebody who struggled with something or who is an expert in a certain area and is able to help guide and direct from personal experience as a human who is engaged with that.

Wendy Kier

Yeah, it's a really different experience. I've written a couple of books. One was written before AI was even about, and that was a personal story of my journey being brought up in the care system. That took me eight months to write, and I think the nice thing about when you write is that you go through a journey yourself as an author, and I think that's quite important.

The other book I've written has been written with the help of AI, and it's a very different type of experience. I think when you're really writing a self-development book, it's important for the author to go through the journey, the insights, and learning how they're actually going to express that through the book, because from that, they will start to learn what their own frameworks are. This is how the IP starts to develop—their own frameworks, their methodologies—and AI won't have it from the personal human side of the individual. It will be theoretical, and it will lack the nuance. AI at the moment can't do nuance. I don't know about later on, but it definitely can't do it now.

Joshua Tallent

Yeah, and there's also the aspect here of the engagement with the book itself, right? If you read one of Tim Ferriss's books—The 4-Hour Body was a book that really changed my understanding of how physical engagement with your body really matters—if you look at how a single book can have an impact, that book sprouted up a bunch of Facebook groups and Reddit groups and people talking about how they're implementing the ideas he had in the book, what kinds of vitamins they're taking, or what kind of exercise they're doing. All of those things actually had an impact on people that brought them together into a community in a way that actually made the book more successful.

I think that's the thing about self-help especially, but even other genres and other types of books: the goal is not just to have someone buy a book and read it. The goal is to sprout conversation, to make it inspiring, to bring people together, and to make them want to tell their friends about the book because their friends then want to go buy it and read it. Building community around reading actually is a big part of the reading experience, and something that has a connection to the humanity of the writing as well.

Wendy Kier

Yeah, you've touched on a really important point, and that is: how can you help people to follow through? If they've engaged with the book and they're reading it and learning something, how do you help them then to make that next step?

Joshua Tallent

Yeah, and that IP matters too, because the opinion you have, or the approach you have, or the scientifically developed process you've come up with gives direction and connection to the people. If everybody's just asking their own chatbot, "Hey, what should I do about my own personal health?"—which is not a bad thing to do—that's very different than having a community of people who come alongside you, and you come alongside them, and we build something together, working together toward the same goal. It's a very different experience.

Wendy Kier

Yeah, many people give us multiple perspectives, doesn't it? You get that common connection you can't get from AI: "Oh, you feel like that too? I thought it was just me." AI might be able to mimic it, but it's different from hearing it from someone else who's struggling with something that you're struggling with.

Joshua Tallent

Yeah. And beyond self-help, there are obviously other parts of publishing, other genres, and other books that benefit from the same idea, even if it's not for the same purpose. At the same time, there's a benefit in having AI tools that help us with the process of reading a book. Amazon introduced some interactive AI capabilities on the Kindle platform earlier this year, including Story So Far and Ask This Book, and your AI Book Companion is also helping readers engage with books using AI.

So what happens when we give readers that ability to have a conversation with the book that they've been reading? You still have the human element of whoever wrote the book, but I, as a reader, want to understand the book more effectively, implement its ideas in my life, or engage with and understand the story. What do you think happens when we give readers the ability to do that kind of conversing with the book itself?

Wendy Kier

It's absolutely fascinating. I created my first book companion from my first book by accident. I was just thinking, "Oh, what would happen if I turned this into something that I can actually talk to and have a conversation with?" And then I suddenly realized, actually, this is really cool.

I read many books, and I'm dyslexic, so I'll get stuck on some terminology or some wording. Now, with my own companion, what I can do is go and ask questions so I can understand what the author has said, rather than putting the book down and maybe never going back to it. Normally, what I would do is read on in the hope that I would understand what the author was actually saying. With a companion, you can actually quiz the book. You can ask it to explain: "What does the author mean by this? What does it mean in context to what I'm experiencing? How does it fit in with the challenge that I've got?" It's a completely different experience, based solely on what the author has created within the book, which then becomes molded to the individual that is having the conversation with the book.

Joshua Tallent

That's interesting too. With this concept of having that conversation, Amazon's Story So Far feature allows you as a reader to engage with what you have already read, right? So it's progressive to the book content itself—it's not going to reveal something that's going to happen later. I just finished the Mistborn series, all seven books in the Mistborn series, and as you're going through, things are being revealed little by little, and you don't want to have something come up in a chat response that's going to happen at the end of the book, right?

I want to be able to say, "Hey, I haven't listened to the audiobook here in a week or so. I've been busy. Give me a recap of where we stand, or what's been the journey of this character." That kind of approach is really interesting because it means that I can come back and re-engage a book that I had to put down. Maybe it was a nonfiction book that was a little heavy or too much for me to keep up with during a busy time in my life, and when I come back to it later, I can just ask, "Give me a recap, what's going on?" or "What about this topic that I'm reading about?"

I read a lot of theology books, so being able to go back and say, "I'm interested to know how the author sees this area, and if there's anything in the book that discusses this specific verse or area of conversation"—sure, you could go to an index and hope that the indexer put that verse reference in there or has a scripture index, but it's very different when you have that AI to be able to interpret the entire book.

Wendy Kier

Yeah, I guess the AI companion that I've created works slightly differently to that. These companions are based on personal development books, and one of the ways that people or clients use them is they have the book, they're doing different sections or doing a course with a client, and then during the week, they can dip into the book almost as extra support. If they're struggling with something, they may not necessarily have read all the chapters, but they can go back in and say, "What does Sarah think of this? I'm struggling with this—what are her thoughts around this?" So I guess what we're going to start to see developing is very different types of companions for very different types of books, which is great.

Joshua Tallent

Yeah, and that also keeps it from being just a general-purpose chatbot. Sure, I could go ask ChatGPT, "What do you know about certain books that I'm reading, and can you maybe intuit what the author might say about this topic?" But unless OpenAI has scanned that book or imported that book into their general knowledge category—which is a very different subject altogether—it's going to be guessing or basing it off of other publicly available information like reviews. So being able to engage with the book itself: how do you think AI-powered book experiences stay connected to the author's actual content rather than becoming that general-purpose chatbot?

Wendy Kier

When you create the agent—when you create the companion—you have to give it very tight boundaries. When I work with a client, we have to think about what we want it to do, and what it is that we don't want it to do. One of the things I do is create a job description with them, so we're very clear about what the book's role is. It goes through tone of voice and the boundaries of what happens in different scenarios. We give examples of how it should work. So we really lock it in and contain it in its own environment, so it's not actually using any other source of reference for the book—it's just using the book itself and any supporting material that belongs to the book.

These types of agents are more conversational and more intuitive. Depending on how you've created it, you can have them be more inquisitive—they'll ask questions and take the person back to something they might have said earlier on. A standard chatbot is a question-and-answer sort of scenario: you ask a question, it gives you an answer. This isn't that. This is much more fluid, much more organic, but you have to contain it when you're creating it. If you don't contain it, it's just going to go all over the place.

Joshua Tallent

Yeah. So this is also an area of concern for people as they're having a conversation with a chatbot or an AI companion related to a book: "I'm talking to this about my personal life, about very sensitive questions or sensitive issues in my life." What responsibilities come along with making self-help content an interactive book, especially when you've got those kinds of sensitive questions?

Wendy Kier

Yeah, the books we create companions for are really sensitive. So I'm very clear with my clients and everybody that you need to have a contract that clearly explains what the creator's role has been, what the AI's role is, and what their role is as well. It's a partnership between the three, rather than just saying, "You need to be careful about what information you input into the AI." I think it needs to be spelled out and crystal clear what has been created and how you use it.

So I get my clients to do contracts with their companions. It's not like Amazon—I wouldn't let these ones go wild. I guess it comes back to containment again and the types of books and companions that we're creating, being crystal clear about each level of responsibility.

Joshua Tallent

That makes sense, and those requirements can hopefully keep the companion from going off the rails. It's very important when you're building these things out to have the right approach to the sensitivity. But what about the other side of sensitivities—the concerns that publishers and authors have around AI engines accessing their books? Obviously, if it's a firewalled system, then they say it's not going to be used for training. Who knows? So, what about those very real concerns for authors and publishers?

Wendy Kier

It absolutely is a concern. So much of this we don't have control over, because people will be taking pictures, uploading things, uploading ebooks. It feels like there's no sense of control where we can say, "You can't do this, this, and this," and that will prevent it. It almost feels like if someone's got the ebook, they will upload it. I mean, I say in my book, "Don't upload it to AI," but I'm not sure anyone really takes any notice of that.

The platform I use says it doesn't include any of the books inside any AI training data. I'm not entirely sure 100% what that actually means, but they tell me that it's not being used to train any of the models, and that it's contained within its own environment. I'm not sure what the answer is, because my brain's trying to figure out a way of thinking, "How can you lock it down?" But I don't think you can now. I think what you need to do is actually step into and take ownership of what it is you have created, and protect those methods as much as possible—so your name's on everything, and you try and protect your own writing as much as possible.

Joshua Tallent

Yeah, that makes sense. Like you said, it's hard for any of us to know how to contain—and if it's even possible to contain—book content that's out there. Obviously, books that have been published before are being scanned; all these companies are buying up old books, chopping up the binding, scanning them, and throwing them away. That's sad to see happening, but it's also not something we can control—it's the way things are working.

So the question then becomes: how do we responsibly engage with AI, and how do we responsibly engage with these tools in a way that will benefit readers, authors, and publishers as well? You've been working through building out the AI companion for a while, and I'm just curious about what experiences you've had that might be useful for publishers who are considering some sort of AI interactive content. What are the things you've learned that you'd like to pass along to others?

Wendy Kier

I think it's really important to have a framework and a process when you're creating any type of companion. It's the same as when you write a book: you want to create your own methodology and have a step-by-step guide of what the role of the companion is, what restrictions you want to put on it, and what the responsibility is of the author, the AI, and the client. Who's held responsible for what?

So I think the first thing would be a very clear framework of how you want the agent to work, and that will also protect you from things that can potentially go wrong, because who knows what AI is going to do in the future? You want to protect the author, yourself, and the client at the same time, so it's about becoming responsible when you're creating those companions. That's really key—not just thinking you can put it into AI and that's it, you've created it. You need to really define what the companion is about and what sort of results the client and the author want the companion to create, and then build from there rather than just guessing and throwing content at it. It will give you answers and you will think it's doing a good job, but the difficulty is, as I mentioned earlier, what it says sounds right, but it may not necessarily be right. So it's making sure it's right in context for the author and the person that is using it.

Joshua Tallent

I imagine this would be even more effective for an author who has multiple books on the same topic—who has been working on this field of study for a couple of years, 10 years, 20 years, or 50 years. That amount of content makes a big difference. If you could take all of the content an author has written about a topic and make it part of the companion, then you're really specializing in a way that even a single book won't necessarily be able to specialize.

Wendy Kier

Yeah, it's funny—the job description just pinged in my head, which might seem a bit of a strange thing to say, but as long as that job description is clear, it's clear how those books connect to one another, and the AI can see the connection between the books, you'll be okay. If that bit isn't figured out—like cross-referencing and mapping between the books—I think at the moment it will struggle. It's almost too much content, and too much can go wrong because one book is very different from the second book. Even though the story or topic might be very similar, they might need a very different type of framework.

Joshua Tallent

Yeah, and as things progress and as AI engines get more robust than they already are, we'll start to see things change. I think it'll be interesting to see the impact on reading from a consumer reader's perspective: "How am I engaging with my books? What are the tools available to me?" If Amazon has their thing, other systems I'm sure will come out with their own as well. It'd be very surprising if Siri wasn't expanded into this at some point on the Apple platform. So I think there's a good chance we're going to see more and more of this. Publishers being aware of it, thinking about it, and considering what their options are—and authors knowing about it and being aware that maybe their book will be engaged with in that way at some point in the future—is pretty important.

Wendy Kier

The other thing I wanted to mention that I haven't mentioned yet, which is good for publishers, is that you can monetize the book companion. Because it's an agent, you can actually charge a subscription, or you can charge a one-off fee, which is great because it means you get the readers' details and you can see what they're engaging with.

Joshua Tallent

Yeah, that makes sense, and that's again another way for a publisher and author to find some value in their book and have more ongoing engagement with it. Well, Wendy, thanks for joining me today on the podcast. I really appreciate your time and your expertise in chatting about how AI companions can work.

Wendy Kier

Thank you.

Joshua Tallent

Is there any place that people can find you online? Where should they go if they're interested in learning more about the work you do?

Wendy Kier

Yep, they can just hop across to https://wendykier.com/.

Joshua Tallent

We'll put a link in the show notes.

Thanks for joining us today for the BookSmarts podcast. If you like what you've heard, please leave a review or rating in Apple Podcasts or Spotify or wherever you listen to the podcast, and please share this podcast with your colleagues. And if you have topic suggestions or feedback about the show, you can email me at joshua@firebrandtech.com. Thanks for joining us and getting smarter about your books.