Making Sense Of Technology in Multimedia

Technology Series: Voice Technology

• Daniel Douglas • Episode 9

Share episode

Send us a text

About This Episode

Today, my guest is Carl Robinson, the co-founder, and CEO of Rumble Studio. This software company helps companies create podcasts for marketing using cutting-edge voice technology. Carl is also the host of the Voice Tech Podcast and has interviewed more than 100 experts in voice technology.

Carl's ambition is to democratize audio content creation so that businesses large and small can establish a presence on the audio channels of the future.

 What to Listen For

  • How living abroad has impacted his life and ambition in Artificial Intelligence 
  • Audio Marketing & Business Podcasting
  • Asynchronous Audio Creation
  • Voice Technology
  • Democratizing Audio
  • What is Rumble Studio

Check-in with Carl

Shoutout To The Show

TU96xZNNzxr4FGt764I6

Buzzsprout - Let's get your podcast laun
Start for FREE

Disclaimer: This post contains affiliate links. If you make a purchase, I may receive a commission at no extra cost to you.

Daniel Douglas  0:00  
Hey everybody. This is episode 9 of The Making Sense of multimedia podcast, the Technology Series. More than ever, technology has proven to be a lifesaver given the global pandemic, and other circumstances that we now face across the globe. We will talk with an artificial intelligence or AI data scientists to learn how voice technology shapes how we communicate, and get our messages out to consumers. Let's get into it.

Today, my guest is Carl Robinson, the co founder and CEO of rumble studio. This software company helps companies create podcasts for marketing, using cutting edge voice technology. Carl is also the host of voice tech podcasts, and has interviewed more than 100 experts in voice technology calls our ambition is to democratize audio content creation, so that businesses large and small, can establish a presence on the audio channels of the future. Thank you for agreeing to come on making sense of multimedia call.

Carl Robinson  1:24  
Daniel, it's great to be on the show. Thanks for having me.

Daniel Douglas  1:27  
Yeah, I really do appreciate it. We're in different time zones. And sometimes that can be a little trying, I guess, to coordinate things. Indeed. So I want to start backwards, Carl, and backwards by I mean, you did say in my research about living abroad had some challenges. Can you talk about that for a little bit?

Carl Robinson  1:55  
So living abroad, it does have a lot of challenges, I suppose it depends where you are. I've, I've spent more than the last 10 years out of my home country now. So I'm from the UK, as you can probably tell, in 2009, I moved to China, I lived in Beijing. And then after that, after seven years there, I moved to France. So I've been spending a lot of time abroad. Obviously language is is the biggest challenge when you first arrive at least. And certainly in China, I was heavily reliant on the friends that I knew out there who could speak speak Chinese, the culture as well to, to get used to the the different way of doing things the way people interact, how to get things done, in your day to day life. Both China and France have interesting bureaucracies I should say. It can be quite slow to get things done in both countries compared to the UK. And just establishing a life really just to avoiding avoiding loneliness. Moving your life forwards, meeting friends, relationships, these kinds of things is that there's a lot of challenges to living in, in a foreign country. But it's very rewarding. I definitely don't regret it for a second. It's brought me a great, great many benefits in my life. And I highly recommend it.

Daniel Douglas  3:12  
So you're in Paris now? I am indeed Yeah. And so what I guess what made you choose Paris over any other country or going back to the UK? I'm certainly no stranger of relocation. As we talked about Off mic as a second ago, I recently relocated, although I'm still in the US. But relocating to an area that you're not familiar with, can be some challenges. So that's why I want to start backwards. Why did you select Paris?

Carl Robinson  3:45  
While it was the it was the obvious choice, really, because when we're in when we're in China, were deciding where to go next. It was time to leave Asia and come back to Europe. And really the two choices with the UK or France because my my girlfriend at the time from French, from France, sorry, she is French. She's now now my wife. And so we're we're looking for somewhere where I could actually continue my studies. I wanted to get back into technology. I did a computer science bachelor's degree way back when I graduated in 2004. I think it was I wanted to get back into the tech, certainly the AI stuff. So I was looking for where I could do a master's degree. And France has fantastic universities, as I'm sure everyone knows. And it's also a lot cheaper than UK. So that was one factor. The other was that obviously, my wife was from Paris. So she had all her family here and we were looking to start a family. So there was a practical element on that side of things as well. And lastly, for us is just a nice place to live. You know, I've visited many times, you know, my life and I definitely wasn't against moving here. And it was a way for me to continue the adventure. Having done that experience in China. I still wanted to live abroad and experience different cultures. I wasn't quite ready to go back to the UK Maybe never will be. So France was the good choice and happy with that choice.

Daniel Douglas  5:05  
I I've never been to France, the only place in Europe that I've been was Spain. And I really love Spain when I was nice. spent a few days there.

Carl Robinson  5:17  
Yeah, it's been it's lovely. Actually, my my brother's just moved to Spain. So I visited for the first time, just a couple of months ago, actually. And it's beautiful. You're picking the the oranges off the the orange trees, which they have a lot of in Spain. And it was great.

Daniel Douglas  5:33  
You mentioned you took some courses in AI? And certainly, you did mention the education would be a lot less expensive in in France than the UK certainly a lot less expensive than here in the US for sure. Oh, yeah, for sure. Why AI?

Carl Robinson  5:52  
Yeah, that's a good question. I mean, I could see that it was the next big phase in computing. I've always had a, you know, a career or at least my more recent career has always been in startups and technology. And, you know, always looking to the next big thing where the innovation is happening, how can we create the next generation of products, and for sure, AI was the and still is really like the the driving force in computer science today. I was working at a startup in Beijing that was using, like a chatbot style interface. Essentially, it was an app that we were building that helped doctors and health coaches speak to people with diabetes, and convince them to, or coach them to eat better and lead more healthy lifestyles. And actually, the conclusion that we that we reached, having built that product, which was a kind of a human to human kind of chat experience, was that it would be fantastic and much more effective if we could actually automate it with some kind of voice technology. So I got turned on to voice technology and the whole idea of chatbots through that startup, and there was a lot happening in conversational AI. Around we know, we saw Alexa at the time as an Alexa and, and these type of conversational assistants really, really taking off and getting a lot of attention and investment. And so I really wanted to work in a field that involved that involve conversational AI to some degree. And I thought the best way to do that was to really understand it at its at its roots. And studying computer science. And an AI specifically seemed like a good, you know, good way to do that, as well as integrate into a new country, learn the language, etc. So, there are many reasons why I wanted to go back to school. But that's primarily the reason I wanted to learn AI.

Daniel Douglas  7:42  
My background is in technology, I was a technology director for a number of years, okay. Now, since I've started doing photography and videography, where I've talked to a lot of folks, especially photographers that tell me, AI stinks, I hate it, it's gonna take over the world. And I just for the life of me, I don't understand it. I think they should embrace technology. It's certainly there. If you look at the software, especially with Adobe, because a lot of photographers use the Adobe suites, I use Photoshop, and Lightroom. And their AI technology is called sensei. And I have to tell you that embracing that technology, me personally, not only is it saving me time, it has saved me a lot of money because all of these different apps that I had to use to accomplish the same thing, I don't have to do that anymore. So when you embraced artificial intelligence, how long ago was that? When did you first get into that? What year?

Carl Robinson  9:03  
Oh, I think I left China in 2016. So that was when I was really thinking a lot about it.

Daniel Douglas  9:10  
Okay, so that brings us up now to talking about some other topics that I know you're excited talking about. Let's start with audio marketing. What's the story on that?

Carl Robinson  9:23  
Are you marketing Yeah, so it really stems from the the conversational interfaces and, and seeing all the different ways that people are using voice enabled devices now. But then it bleeds over into into podcasting and audio content. And my story essentially was, as I was saying, I was working first with chatbots. Then I studied AI. And I actually did a an internship at a research center in France called outcome where I was building machine learning models to change the emotion in the voice and I was working, you know, pretty much alone most of the time. That's what, that's what interns do in the lab. And at the same time, I started a podcast called The Voice tech podcast, which is all about smart speakers, conversational AI. And really, it gave me an opportunity to speak to lots of interesting people, as I'm sure you find yourself, it's just an amazing tool to, to get to have great conversations and and also boost your profile. And it made me realize just how many companies in the conversational AI space wanted to be on a podcast, right, and the fact that they'd be prepared to speak to me who, you know, I've just, I'd gone back to school, I wasn't young, but I was, you know, I didn't have any authority in the domain at all. And yet, these people were willing to spend an hour or more on the on the, on the call with me and talk at length about the technology and their products. So I saw there was a real demand from companies to want to promote themselves through audio, I could see podcasts were really booming, like just the number of podcasts being produced, the amount of podcast listening, especially in the US, was really going through the roof. And I was talking all about conversational interfaces, and how you know, these things are going to transform the way we we interact with computers. And of course, you know, the thought occurred to me at some way, somewhere along the line that you know, maybe I could put these two things together.

Daniel Douglas  11:19  
Yeah, let me jump in here for a second. So you mentioned a couple of times conversational. I guess conversational AI or conversational? Explained. Explain that a little bit.

Carl Robinson  11:33  
What is conversational AI?

Daniel Douglas  11:35  
Yeah. What is that? Yeah, sure.

Carl Robinson  11:39  
So I suppose it's a, it's a collection of different technologies that enable machines to listen to what people say. So to actually capture the words that people are saying, and convert the audio that they record, into into text. That's, that's one, one technology, often called speech to text. And then another technology would be natural language understanding. So for the the machine to actually understand, at some level, what the human actually meant, that the intent behind what they sent, maybe they wanted to do something or get something from the machine, maybe they're expressing a certain thoughts, certain set of concepts or ideas or even an emotion in their voice. So the ability to extract that and represent that in a structured way. And then also the ability to generate responses in text. First of all, and I'm sure your listeners have heard of models like GPT, three, which can generate these realistic phrases. And then finally, to be able to take those written phrases that it's generated, and convert them into audio speech. So text to speech, or synthetic speech, is another technology. And so if you put all of those four technologies plus many other technologies exist as well, like emotion detection, and voice transformation, all of these other things, you put all those together, I would group all of that under conversational AI, because they all use AI models or, you know, machine learning models, at their core to, to operate. So, put together they form something greater than the sum of their parts, like you can create these systems like Amazon Alexa, that can actually listen to what you're saying, understand it, come up with a response, and then speak it back to you. And that's what I see as a as an AI system.

Daniel Douglas  13:28  
We can't see each other, but I have this big grin on my face. Because when, when I moved here into the new house, I told myself, I want to make this house smart. So I'm, I'm invested into the Apple ecosystem. Okay. And no apple. It's not paying me to say that. But I had a few friends that came over to see the new house and they said, Well, your house is smarter than you are. So that all ties in. And that's why when you were mentioning Alexa and all of that, of course, since I'm, you know, only Apple platform. It's with Siri. And sometimes she can be a little unruly, but yeah, I just thought that was funny.

Carl Robinson  14:20  
Yeah, they're not perfect. None of the platforms are perfect. Rarely. I've got the Google set of my home listening to everything I say, and it gets things wrong on a daily basis. It can be quite infuriating at times, right?

Daniel Douglas  14:33  
You also talked about how AI and interactive voice technology will just disrupt audio content creation, because I'm really curious about what that disruption will be.

Carl Robinson  14:48  
Yeah, sure. So yeah, like I was saying, when I was doing the podcast talking all about conversational AI, I, it occurred to me that, you know, a lot of a lot of the questions I was asking on the podcast, I would ask Ask the same, you know, I was asking them multiple times. And there were certain things in the podcast I was just doing again and again with every guest. Often, you know, I was tackling the same topics, you know, there was a similar structure to the show. And because I was talking about conversational AI, I just thought, well, maybe a machine could do some of what I'm doing right. And maybe I was a little bit over optimistic or a little bit naive. Because you know, what we do as podcast host is actually very complicated for machine to do. But that inspired me anyway, to be able to the idea of getting a machine to be able to do what I was doing, like on the mic, interviewing these people, and then maybe perhaps turning that into a product that can actually help businesses market themselves through audio. And so and so yeah, this is, this is the root of the idea that that led me to start my business, which is, what is today rumble studio,

Daniel Douglas  15:54  
right. And we're gonna talk about rumble studio in a minute. Because when I looked at your links that you sent to me and looked at all the information, I said, this, this is kind of interesting. He, I want to know more about that, especially now, like I said, at the top of the show, with the pandemic that hopefully is easing now across the cross the globe. But there's still a lot of folks that still doing remote work. And there's a lot of folks that just don't want to go back into the office now. Hmm. So I can see this technology, and especially how you have it set or how it's how it works, where you don't actually have to be on site. And I stumble a little bit because a couple of years ago, before the pandemic, I was doing some video work. And I wanted to interview a president of a college in Boston, and she just did not want to be on the telephone. She didn't want to do anything with that. So I know I'm jumping around. But I'm gonna jump right into rumble studio. I think rumble studio would have been perfect for her because I could she could you explain it. You can explain it better than I can say it.

Carl Robinson  17:20  
Yeah, sure. So I'm hearing is that this person is guest didn't didn't feel comfortable on the mic. Is that right? Like too shy to subconscious? So

Daniel Douglas  17:28  
is this yes, she she was her assistant said she really does not like to be on the microphone. She was happy to do a telephone interview. But at the time, this was maybe two and a half years ago. And at the time I was using Skype. And we all know Skypes audio is not the best in the world. Zoom wasn't a thing at that time, at least not that I know knew of. So I think something like rumble studio would have made her more I ended up not getting that interview for that little short documentary. Yeah, I think rumble studio would have been something perfect for that. So instead of me following my script here, let's talk about rumble studio now.

Carl Robinson  18:15  
Sure. So in a nutshell, Rumble Studio is a SaaS platform or software as a service that helps companies, agencies, individual podcasters grow their brand or their online presence using audio. And it's an online tool that you can use to conduct remote guest interviews, like we're doing now on we're using Riverside today. Right? There are many tools that offer remote guest interviews. But the The unique part of Roma studio is our interviews are asynchronous. So they're not live interviews, like we're doing now. But rather, you would add some questions into our system, you send an invitation to your guest or guests, it can be as many people as you like, and then the guest, in their own time, answers those questions, recorded their voice, and is taken through a structured interview by the machine itself by Rumble studio. And then after they've recorded, the Creator can go back in, retrieve that audio, and manipulate it, add some of their own audio jingles, reorder, export, and produce a podcast episode or any other kind of audio in order to publish that.

Daniel Douglas  19:27  
What are some of the use cases for that? Because mine was a perfect one. I thought

Carl Robinson  19:33  
Yeah, for sure. So there's there's so many advantages to doing things that asynchronously and some disadvantages, I'm not not going to pretend that there's it solves everything without causing some other problems as well. So it's a it's one of the tools that audio creators can can have in their their toolbox. You mentioned one which is that it's it reduces the pressure for the guest because often, you know when you're a podcast guest you're really put on the spot. The host will say You know, ask you a question you're like, well, that's a big question. I'm really sure. Not really sure how to answer that one, but your life, right? So you got to say something. And that that really stresses people out and puts a lot of people off, even agreeing to be on a podcast. So I think that's a, that's a shame, because obviously, there's a lot of people out there with something interesting to say, and you don't want the pressure of the environment, putting them off. And also, it's not just, it's not just the pressure, but it's also the fact that you haven't really had time to think about what you're going to say. So often, you blurt something out without really thinking about how you're going to structure it or all the points you want to make. So one of the advantages of with asynchronous interviews, and is the guest can think about what they're going to say beforehand, and therefore the answers are better informed. Now, this could be for sorry, go ahead.

Daniel Douglas  20:46  
No, no, I agree with you. Because in this same video segment, and I interviewed another CEO. At the ACC, guess I can say it doesn't matter. She's not the CEO of Campbell's Soup Company in the US. And her assistant wanted to see the questions that we're going to be asked ahead of time, which is not unusual. Well,

Carl Robinson  21:17  
if the bigger the company, the more I think that happens, we've certainly come across that. And there are some areas, there are some industries that are very sensitive, like the medical industry, for example, where there's a lot of liability. And they're very scared of saying anything that they could be, you know, taken to court for, for example, that might mislead people. So, right, but in those environments, reviewing things in multiple stages, and getting all the stakeholders to see what's going to happen, and then also review the content after it's been produced, as is something we've seen a number of times.

Daniel Douglas  21:48  
Right? Again, in that my use case, that would have been perfect, because she would have had an opportunity to formulate her responses. You know, instead of me standing there, even though she had the questions, and I'm asking those questions. A couple of times, we had to do a couple of takes, because she just wasn't comfortable with what she said at the time. So talk about some other use cases for this because this to me, this is exciting.

Carl Robinson  22:18  
Yeah, absolutely. So these cases were focused on. For companies, it's really creating a branded podcast. So they can, you know, represent themselves represent their brand on the audio channels of tomorrow. And that can be traditional podcast, but it could also be voice interfaces. It can also be audio on their websites as an audio FAQ, for example. So there's lots of ways companies can use audio for marketing can be internal communications, as well. So bigger companies, big groups, often have 10s of 1000s employees and producing a podcast for consumption internally can be a great way to keep everyone up to date. It can also be really good for sales enablement. So if big companies are releasing new products, they need to get everyone in the company up to speed on what the product is, you know how to use it, why you should sell it, you know, in terms of you know, the salespeople, how to sell it, what the pricing is going to be like. So there's lots of reasons to create an internal podcast, but not everyone in the company as an audio production expert. And so an end to end solution like rumble Studio can be used for that. It's designed to be used by by people who don't have any experience producing audio. And it also because of the asynchronous nature of the interviews, allows for that step by step process of construction, with all the reviews in between as well that the managers need to have a look at things before it moves forward. So there's that. There's also things like, Client Testimonial gathering, for example, or any use case where you would want to ask the same questions to multiple people. And this is something that's that's great about doing things on on a satellite Rumble is that you can write a set of questions, for example, client testimonials, you might, you might want to ask, What did you like about the product? Why did you want you know, why did you like those things? What were you not so keen on? What would you improve? And maybe you want to send those questions to 100 customers and capture their audio so you can really hear the emotion in their voice so you can really learn you know, what it is that they were happy about and not so happy about your product. You can just put those into an interview, you got a single public link, what I say public, you get a single, shareable link that you can share with people privately or you could even make it public, put it on your website, put it even on social media. I say client testimonials, you could do the same thing for social media engagement, for example. So you could you know, as an influencer, can create a single question on rumble and say, you know, what do you think about my my new initiative that you know, I love Last week, you know, the the videos that I put out last week, what do you guys think of that? You know, I'd love to hear your comments and click this link that I've put on social media and then your fans can click it, they can record their answers. And that influencer can go in, listen to the audio, read the transcripts as well, because we have that, and then pick the best bits from their fan base, and republish that out not only as a podcast, but back out as videos onto social media, and really excited about that, that loop, that that created all that community engagement that it can, it can create, and then the repurposing of that content to put it back out to the community. So you're showing them that you are listening, and that you do care.

Daniel Douglas  25:41  
You answered a question that I didn't ask you. All right. And I'm glad he went on. In the use cases, a lot of photographers listen to this podcast, and videographers and hopefully other folks. And so I was sitting here as you were talking, I'm saying, Well, what would be a use case for photographer? Testimonials? I mean, you couldn't have said it better testimonials? Because in the photography, business, word of mouth is huge. Yes. And so I can see a great use case for photographers, for rumble studio with just testimonials and social media. Let me ask you this. Are you familiar with a service called Speak? Pipe?

Carl Robinson  26:36  
I am Yes. And some of the people I speak to in the podcasting world do you speak by and speak highly of it? To pardon the pun. I think that the difference between us and speak pipe and I think maybe there's one or two others that do something similar is that we focus on the like a multi turn conversation, essentially. I think maybe they've updated their platforms. I don't I don't want to say anything wrong. But from what I understand SpeakPipe is a is an audio drop box where you, you put one question and then the guests can record as much as they like in a single in a single go.

Daniel Douglas  27:11  
Right? They haven't changed. They haven't changed.

Carl Robinson  27:13  
Okay, no. So Ron will take that a step further, where you can break that interaction with the guest into multiple questions, essentially. So the first few questions can really be warm up questions. It can be about, you know, who are you and these kinds of things, get them comfortable with recording their voice. And then you can go into more depth. The vision for Rommel studio and this is actually what we were building right now with our data scientists and that the AI side of rumble, which is not released yet, but I can tell you a bit about it, is to actually listen to what the guest says, Sorry, listen to what the guest says in their first answer. And then dynamically generate in real time a follow up question, based on what they said, to actually bring it closer to a live conversational interaction. This is the bit that I'm most excited about. And that comes from the you know, the voice interface world.

Daniel Douglas  28:06  
That is because when I do these podcasts, and I usually tell people up front. I like to make these light hearted as much as possible and have fun with it. But also, as I'm asking questions, and I'm getting the responses back from my interviewee I'll spin off and ask a question based upon what he or she may have said that I just didn't think about totally, totally. And so now you with rumble studio, you're introducing that as a as an AI feature. I think that's fantastic.

Carl Robinson  28:46  
That's it. Yeah. And to be honest, once we've built that technology that can be used in a completely automated setting in an asynchronous setting. We could, in theory, bring that technology to live interviews as well to support podcasters running live interviews, because there's no reason why that technology couldn't listen to what your guest says, Live, and then just propose new follow up questions on the screen that you could choose to ask or not right? It could be a kind of decision support tool for a live podcaster. I

Daniel Douglas  29:16  
definitely see some use cases for that. Carl. He talked about democratizing audio. What does that mean?

Carl Robinson  29:26  
Democratizing audio? Well, there's a still a big barrier to entry I feel in producing audio content for a lot of people, individuals and companies, not necessarily producing your first podcast episode. If quality isn't really an issue, now, it's easy just to pick up a mic and record your voice. Right. But one of the problems with podcasting is that you don't realize the benefits of podcasting, certainly for businesses until you've released you know, a significant amount of episodes you can't expect to release You know, three episodes, and then for your sales to increase 100% or anything, you need to do it consistently, regularly and over the long term. And audio production these days is still a time consuming activity. And that either puts people off from starting, or it makes people quit. Soon after they start. And there's a term that I'm sure you've heard of called pod fade. Right? Which basically means, you know, people who start with all the enthusiasm of a new podcaster, after about seven to 10 episodes, I think is the average, realize this is actually quite a lot of work. And I'm not seeing any immediate benefits from it. Therefore, I'm just gonna put it on the backburner and come back to it later. And then that podcasts, you know, basically dies. Sometimes that happens gradually, which is why it's called pod fade, you know, they're releasing once a week, then they're released once every two weeks, and it's once a month, and then one day, they just start releasing, and it's a shame, you know, this is a, it's a real problem. And so democratizing audio content creation basically means to make it 10 times faster, easier, and cheaper to produce audio content, so that more people can get involved. And also people who do get involved continue to do it over the long term. I think those are two important challenges that we should tackle in the audio world. And that's what Rama studio has really been built to, to achieve.

Podcasts we love

Check out these other fine podcasts recommended by us, not an algorithm.

This Week in Photo (TWiP) Artwork

This Week in Photo (TWiP)

This Week in Photo