The Alexa Podcast - Episode 7
Duration: 30 minutes, 01 seconds
Google Play Music
YouTube (+ closed captioning)
Bradley Metrock: [00:00:12] Hi, and welcome back to the Alexa Podcast, for November 30, 2017, Episode 7. My name is Bradley Metrock, I'm the CEO of a company called Score Publishing based here in Nashville, Tennessee. My co-host is Kevin Old. Kevin, say hello.
Kevin Old: [00:00:29] Hello Bradley. Great to be back.
Bradley Metrock: [00:00:32] Great to be back with you as well, Kevin. I'm looking forward to it. We have a great guest today as well. We'll get into that in just a second. But I want to give a shout-out to our sponsor for VoiceFirst.FM, VoiceXP. VoiceXP is taking the lead in developing the skills for some of the best brands in the world. With VoiceXP, all you have to do is say it to revolutionize your marketing strategy. Check them out at VoiceXP.com and look up Bob Stolzberg, he's doing great work in voice. If you need an Alexa Skill or Google Action, go see them. You'll be glad that you did. Our guest today is Amy Stapleton. Amy, say hello.
Amy Stapleton: [00:01:11] Hi everybody.
Bradley Metrock: [00:01:12] Amy, thank you very, very much for joining us.
Amy Stapleton: [00:01:14] You're welcome. I'm really looking forward to the conversation. Thank you for having me.
Bradley Metrock: [00:01:18] Absolutely. So Amy is the CEO and founder of Tellables. And she'll tell us what that is in just a moment, but she will also be joining us as a speaker and guest panelist at The Alexa Conference coming up in January, which is January 18th through the 20th in Chattanooga, Tennessee. If you are not registered....this is going to be a #VoiceFirst event not to be missed. You can get information on that at AlexaConference.com.
Bradley Metrock: [00:01:49] Amy, share with us, just to start off, what is Tellables?
Amy Stapleton: [00:01:54] Tellables is a company that I started with my co-founder Cory Boswell, and we are creating what we're calling interactive story games for voice assistants. So those are short, engaging, little stories that are actually told in the voice assistant's own voice, using text-to-speech and we're trying to make them interactive, using new interaction models that we're inventing.
Bradley Metrock: [00:02:23] How did you get into the voice space? Share with us some of your background and what led you to the point of wanting to start Tellables, and seeing that as an opportunity.
Amy Stapleton: [00:02:34] Sure. My last job....I worked at NASA actually for 14 years as a consultant, helping them with implementing ERP systems, and then I transitioned as a civil servant to be an IT manager for their enterprise applications. But I've always been interested in conversational technologies, and so as I came up to a place in my career where I knew I would be able to retire early, I started looking more and more into how I can get involved with voice technologies, conversational bots, that kind of thing. And I was at a conference, I think it was in 2013, and I believe it was Leor Grebler. I think you had him on your show before, right?
Bradley Metrock: [00:03:21] Yes. Yes, we have.
Amy Stapleton: [00:03:23] Yeah, so I saw him, and I'm pretty sure he was doing a demo of their UBI product. And I just immediately thought that there's going to be all these voice assistants....and again, I mean it was very early on, so we didn't know exactly what they were going to look like. But there's going to be voice assistants, there's also going to be these conversational companion bots, and I really thought that they were going to need stories to tell. It just kind of seemed like a great opportunity. I've always been interested in storytelling and I thought that these characters, these personalities, are going to need stories, and so I thought that might be kind of a cool thing to experiment with, and to just try to see how that was going to work. So that was the concept early on, and Leor actually gave Cory and me one of the early prototypes of the UBI. And we started to create some little stories, but there were some technical challenges because at the time, you could only broadcast I think like 60 seconds of text-to-speech before there had to be a customer interaction. So that wasn't really conducive for storytelling.
Bradley Metrock: [00:04:35] Leor is great. Yes, we did have him on the VoiceFirst Roundtable for an episode a few months ago, and he's just a fascinating individual who also will be joining us as a panelist as part of the Alexa conference, I might add. But Amy, both of us sort of have a mind for how voice technology will impact publishing and telling stories and sharing narratives in the future. And my question for you is, with what you're doing with Tellables and these stories, these interactive experiences you're creating....how are you looking to measure success with those? How are you either now, or into the future, looking to monetize that?
Amy Stapleton: [00:05:24] Well, I think the only way to really measure success for us right now is how often our voice apps or our Skills get accessed, and how often they get used, and then obviously, too, if they're repeat customers. So we try to look at those analytics. Just to step back a little bit....what we're trying to do is perhaps a little bit different than what some of the other interactive storytelling companies are focusing on. I kind of have this vision of the voice assistant and/or the companion robot or whatever these things are going to evolve into. My vision is that they will tell stories in their own voices. So what we're trying to do, or what we decided to do....we kind of went down the path of using text-to-speech as opposed to human narration. And I think if you are a storytelling company, that's kind of the first decision you have to make. Are you going to use text-to-speech, or are you going to use human narration? Audible.com is very popular, and it works well also on a voice assistant platform. But it's a different experience than what we're thinking about trying to provide. We're actually trying to create stories that the character can tell in their own voice. But obviously, then you run into the current constraints of text-to-speech. It's a different tactic than other companies are taking.
Bradley Metrock: [00:06:58] So you're going to have, as Tellables grows, different authors, or different content creators, different opportunities, to tell different stories. And if you're choosing between story A and story B to devote Tellables resources to....walk us through how you make that decision.
Amy Stapleton: [00:07:23] It's tough right now, because we're so early. What we're trying to do is invent interaction models that are that are fun and engaging, and it's really hard to figure out how to do that, because we don't really have access to the audience. I mean, for trial purposes. As an example, we created our first storytelling voice app for the Alexa platform, which was, at the time, and I think it's still called listening comprehension practice, and that's kind of a strange name. But the reason is because we started out with just a simple storytelling Skill. We got some feedback from Amazon, and this was early 2016, that it wasn't interactive enough and so we added some comprehension questions to the end of it, so that the child could hear the story and then respond to the questions. But that interaction model, just based on how often that skill was used, we kind of got the sense that maybe that interaction model is not that engaging, right? I mean something like The Magic Door, which is kind of a Choose Your Own Adventure, that seemed to be a more engaging model, because you could just see a lot more people were using it and they were going back to it. So we were like okay, well, what's an interaction model that might work better?
Amy Stapleton: And I really didn't want to do a classical Choose Your Own Adventure story. I wanted to do something different. So we came up with the concept of Tricky Genie, which has been pretty successful, and it seems to resonate, and it's kind of a different interaction model, it's more of a problem solving type of story experience. So I think what I'm looking for is, these interactive interaction models that are going to work. But right now, it's more like just throwing stuff against the wall to see what people use and what they don't use. And it's kind of hard right now, because there's not that many other people that you can talk to about this kind of thing, because it's so new. I mean, if you talk to people that are storytellers, they don't really have a sense for this new platform, and so you can't really engage them in a meaningful conversation about interaction models for voice assistants. You know what I'm saying?
Bradley Metrock: [00:09:31] I do. I do. It's funny you would say that, because, as you know, we just acquired the Digital Book World business, in that which is centered around this big annual conference, and one of the things that we made very clear is that we're going to involve voice technology and what we call #VoiceFirst technology in that conference moving forward, because how could it not be? Anybody publishing any type of content has to be aware of voice assistants and what they can and cannot do. They have to be aware of the shifting way that we're using computers to more voice-centric, #VoiceFirst approach. And when you have that type of conversation, just in this example, with someone in the publishing industry, who has been in the publishing industry for 30 years, there's a lot of groundwork you have to lay first, before you can even get to the point of having a conversation on that. And it's just that people don't see the future yet. I know that you do. And so it's interesting to hear you talk about how you're experimenting with different interaction models and you’re just laying the groundwork for what's to come with your business. It sounds like that's what I'm hearing you say.
Amy Stapleton: [00:10:57] Yes, to me, it's exciting to try to come up with a different way for someone to engage both with a character, or a personality, a virtual personality, something like your voice assistant device. So Alexa, in a way, is a personality. But when you're engaging with that personality, you're also engaging with an author that has created an interesting story, or an imaginative world. So I think a lot of people are gravitating towards podcasts and listening to audio content. Now how do you take that desire for an audio experience, and then turn it into a kind of interactive, quicker, more engaging experience that you can have with a voice assistant? That's what I'm trying to figure out. I haven't really cracked the nut on that yet, but I imagine a future where you're interacting a lot with your voice assistant. You want to hear something interesting. And you have a favorite author and there is a quick, interactive, type of experience that you can have with Alexa, or whatever the voice assistant is, that is provided by that author. So you're immersed in the world of that character, and it may be kind of a Choose Your Own Adventure thing, it may be something a little bit different, but still it's a fun, concise, engaging, experience, and it makes you want to read something by that author later. So it's kind of a gateway into other types of mediums that were created by that author.
Bradley Metrock: [00:12:42] I am right there on the same page and I think about it....my framework is sort of the DVD bonus features. I always think about DVDs and bonus features and how you can go watch the movie....if you're going to watch a movie, you can go watch it in the theater and that has certain attributes and pros and cons to it. Or you can wait until it comes out and you can rent it on demand. So that's got certain attributes as well. Or you can buy the DVD, and one of the reasons to do that is the bonus features. It has an entirely additional layer of content. And while that's not a precise example or framework by any stretch for what we're talking about with voice, it's a good starting point. What additional functionality or feature set does voice need to bring to the table for authors in fiction or non-fiction to justify either an existing price tier or a premium price tier above what that book is say in softcover prep or just the straight-up audio book version? If I can buy a version of the Harry Potter reboot that's released in 25 years. I can get the audio book version for five bucks, or for ten bucks I can get it with all sorts of behind the scenes interview audio that's in there, and there may be some interactive elements, where maybe there's a glossary and I can ask Alexa at any point what one of the words means, or to jump to a certain thing or tell me about this character....I forgot who that was. All of that. I'm very interested in that as well and I'm glad to hear that....it's nice to have somebody else who can have a conversation with Amy, because you're right, there's not that many of us out there thinking about this.
Kevin Old: [00:14:48] Yeah, absolutely. Amy, I'm fascinated in this space as well, because I think that we're all trying to find that next level of where we can take the resources we have, like books and stories and reference material, those kinds of things and then get that value add, where there's a piece of technology that can offer us more insight into whatever it is we're reading or working with. I wanted to know, from your perspective, do you think with the platforms that had been developed by Amazon and Google primarily so far....do you think that we have all the technology we need to develop what you envision, or do you think there are holes that are left, and what are those? What's left to develop, if anything?
Amy Stapleton: [00:15:40] Well, the biggest missing piece right now is just the maturity of the text-to-speech. Again, I don't know whether that's something that everybody envisions the way I envision it. I really want the interactive experience to be between the person and the personality, so that's why I'm relying on the text to speech. Amazon is really doing a lot of cool things with that. And Cory and I are going back to our Skills and we're trying to leverage some of the capabilities with the speech markup language and the "Speechcons" that Amazon has recently brought out that add a lot more dimension to her voice. So if we can do that, and then even find a way to like easily layer the dynamically generated text-to-speech with sound effects and background music, that would be perfect. Because what we want to do is create a real storytelling experience. But again, with Alexa or the Google Assistant talking in their own voice, I mean, what are the challenges with being a third party developer with a Google Assistant? Right now, they don't even let you utilize the real Google Assistant voice. You have to use a different voice, and those text-to-speech voices are not nearly as good as the one that you hear if you interact with the first party skills on the Google Assistant, so that's kind of frustrating. Whereas if you develop for the Alexa platform, you can leverage the Alexa voice as well as all the Speechcons and all that kind of thing. That makes it easier to create a better storytelling experience so that mundane type of thing right now is just text-to-speech limitations. That's the biggest hurdle that I think we're facing right now, because a listener that has the choice between listening to one of our stories told us that in a story narrated by an assistant voice versus a story that's human narrated, they're probably going to pick the human narrated story. But again that's just not the experience that we're going for.
Kevin Old: [00:17:46] Certainly. It's interesting that you mentioned layering. So the assistant voice, and then sound effects behind that, I can see that being an application for what you are envisioning. I agree with you in that I see the scalability as the biggest factor in wanting to go the route that you're going, versus using a human voice. And I think that most businesses from that perspective with print, if they're able to take their print offerings, and run some automated software to be able to provide a voice experience, versus a recording studio with a human and editors, etc. … rhe cost perspective is drastically different. Is that a motivator for you and your team?
Amy Stapleton: [00:18:43] Yes, I would not be telling the truth if I said that wasn't a motivating factor to me. We have a content database that I can just add story after story after story in, and that way, people that listen to our Skills have a lot more content at their disposal. But that wouldn't be possible if I had to have human voice-over actors record every single thing, so that definitely plays into it. But I still think the number one hunch that I'm working towards is that people are going to want to engage directly with the personality. So what we're doing is different than if you're trying to create a voice skill for a brand. I mean, I can understand that if you're a brand you would want to have your own specific voice, but if you want to have a fun engaging experience with Alexa, I think you want to hear her talk you through the story or work through the game or whatever and in her voice.
Kevin Old: [00:19:49] Yeah. I completely agree. I mentioned a couple of times on the show that I have a 6-year-old boy, and he loves to interact with Alexa, to have it play a song for him, or do something, or tell him a piece of information. And I can see the storytelling aspect as more of a supplement to not only us reading as a family, but for him to be able to read a book, which he's learning to read now, and he can do so quite accurately. But these interactive devices could help him with a word if we're not available for that moment, or if he wants to know who this character is, or some more information. I think being able to take stories and then chop them up into ways that we can get additional information out of them is incredibly valuable. So working with these platforms, with Amazon and Google, what challenges could you name me, one or two challenges for either platform that you've run into, and specifically around developing a voice user interface which in software, is a new thing? I've developed UIs for web and mobile, and when I started learning about voice user interfaces, I had a lot of learning to do and a lot of iterations.
Amy Stapleton: [00:21:19] Yes, I think....it kind of goes back to what I was talking to Bradley about the beginning. How do you invent an interaction model that is going to be interesting? If you have the character tell a story, then what's a way to get the listener involved in the story? And in such a way that you know you're dealing with all the constraints of the platform? I can't start telling a story and then ask....let's say it's a child-focused story, I ask the child how do you feel about this particular character, or what do you think is going to happen next? There are certain prompts that I can provide, but then the prompt has to be constrained to a certain extent, so that I can program the pathway. So that's the challenge, how do you come up with these interactive story games that are engaging, they're not boring, but they still follow a constrained path? And I think with the Tricky Genie Skill we came up with, that sort of works. But that's just one interaction model. The challenge there too is what happens if the person goes off script? With Tricky Genie, they have to pick between three sacks that the genie is offering to find the best solution. But then again, they might not say the right number, or whatever. So you have to program in all the things to kind of keep them back on track to where it is you want them to go within that story experience.
Kevin Old: [00:22:50] I want to talk a little bit about the technology. What are you currently deployed with? For Amazon, are you using all of their infrastructure offerings? With Google, what are you using there?
Amy Stapleton: [00:23:04] Cory is the technical guy, and I would say that we're not taking advantage of some of the latest tools that Amazon has developed. I know I've been on several of their podcasts where they have done some programming using … I think it's called the Skill Builder tool, which is still in beta … but it's basically almost a visual type of Skill Builder, and Cory has developed everything in Java, just kind of on the back end it's in Java, because he has all his experience on that. He does all the business rules and everything and Java. And we've done the same thing for the Google Assistant deployment. That's the way he's worked it, and he's built a separate content management system specifically for me to keep the BASIC rules-based programming separate from the content, so I'm able to just log into my CMS, and put all the stories in there, and then there's no requirement to resubmit the skill every time I create a new story.
Kevin Old: [00:24:12] Awesome. So you've developed a bit of your own rule engine to kind of help with the story flow. Is that correct?
Amy Stapleton: [00:24:21] Yes. Amazon actually did a case study on us a few weeks back, and we published a diagram in that case study, which is available on the developer's blog, the Alexa developer blog, and it shows sort of like this content management system, and there's also like a state machine that Cory has created, and he can describe that a lot better than I can. But I think what he's trying to achieve is that we were creating multiple Skills or voice apps, they can all leverage the state machine without him having to code all the stuff separately for every single Skill, and that way, it speeds up the deployment for other skills that we're going to use, because they can leverage some of the existing common infrastructure.
Kevin Old: [00:25:11] Absolutely. Yeah, that's a great use case of a state machine, now that I can think of it. And we'll link to that blog post in the show notes if anyone wants to follow up with the specifics. One last question: Storytelling is how information has been passed down through history. And we're at this crossroads, right? We've had printed materials for hundreds of years. We now can read information across the world in a minute. Right? With these voice technologies, it seems that games and reference, storytelling, bot-type operations that do this for me. In your perspective, outside of just the storytelling, is there any other area or application that you would be interested in, that you see these platforms going toward?
Amy Stapleton: [00:26:11] Oh gosh. Yes, there's tons of things that these voice assistants can help us with in daily life. I think the therapeutic aspect is really interesting. There's a Skill I just tried out the other day. I think it's called Bravo Tango. That was developed by I think Zap Media, who created it with some folks from the military. I don't remember which branch. And it's supposed to help soldiers that return from the battlefield that are maybe having anxiety, or whatever. So you can actually interact with them. With this case I think it's on the Google assistant platform, and it will walk them through some calming breathing exercises, and things like that. And I tried it out and I thought it was really effective. I think that's a really cool application of kind of a conversational technology. And you could combine that with some storytelling as well. I just think there's a lot of different places where this could go on and I'm just really intrigued about trying to work with authors and other kind of creative people to come up with some interaction models that really resonate.
Bradley Metrock: [00:27:30] Amy, you're off to such a great start with Tellables. We will certainly link the Amazon developer blog, which is great, in the show notes, so that people can read that about Tellables. We're coming up on the end of the year. We've got The Alexa Conference in the first part of next year. But my question for you is ... and this is my last question … in one year's time ... as you're preparing for the holidays this time, next year, where do you want to Tellables to be at that point in time?
Amy Stapleton: [00:28:06] One year from now? Well, I would really like it if we had one or two additional Skills and Actions, voice apps in general, that we're really resonating with a large audience, that people enjoyed, and they were returning to, and they were engaging in kind of a fun story experience directly with a voice assistant. If we could find some more interaction models that people like, and they come back, and they listen to the stories again and again. And they tell their friends, and we're successful with that, then I'd be very happy.
Bradley Metrock: [00:28:47] Awesome. So for people who have listened to this show, who are intrigued about Tellables, who want to reach out to you - what is the best way for them to do that?
Amy Stapleton: [00:28:57] The best way is to shoot me an email at firstname.lastname@example.org and you can also check out our Website at Tellables.com. I try to write a blog post once a month or so, to talk about the latest things we're doing, so they can stay informed that way.
Bradley Metrock: [00:29:15] Amy, thank you so much for joining us today, and sharing your insight and expertise with us.
Amy Stapleton: [00:29:20] You're welcome. Thanks a lot. I appreciate the conversation.
Bradley Metrock: [00:29:24] Thrilled to have you. Kevin, thank you as well. I appreciate you joining me once again.
Kevin Old: [00:29:31] Thanks Bradley.
Bradley Metrock: [00:29:32] Yeah, it's always a pleasure. For the Alexa Podcast, thank you for listening. And until next time.