 So, hello everybody, welcome back to yet another episode of our series Power People which profiles the hussos of the music business industry and today on this very special episode we have with us Mansoor Rahim Akhan, founder and CEO of Beat Oven AI, welcome you here today Mansoor. Thank you so much, I am very glad to be your part of this episode. Mansoor sir, can you share the inspiration behind the creation of Beat Oven AI and what motivated you to start this venture? Yeah, so basically a little bit about my background, I am a musician, I come from a family of sitar players, we are a 7th generation owner of Gharana and I am personally trained in music by my dad, I have been playing for the last 20 years and I have played 200 concerts globally, I have released music on various streaming platforms and millions of streams on my music, but I have also been in the technology space, so I have done my undergrad from an IT in electronics and communication engineering and then I went for my masters in music technology at Georgia Tech in Atlanta in the US. And I have worked across several startups, primarily in tech and product pools and I am really focused towards audio, so I have always been combining my music and technology skills throughout my career and also my co-founder, the business partner, who I started this company with, he has a similar background, who also got a master's in sound music computing from Europe, also a bachelor's in information technology. So both of us combined our skills and we found this very interesting problem in the market, which was the production music library requirement for content creators. So basically people who are making games, videos, podcasts, they require background music and the way how it was being solved today was by using stop music libraries and we saw an opportunity to build a solution where these content creators could compose their own soundtracks by giving very simple prompts and this is something that we thought about back in 2021, very early, now it's a very common thing of a generative AI, but the original thesis was on that basis that okay, there is a need for a tool that allows people to compose their own music in a very simplified manner, so that's the idea behind and Genesis behind V2Window.io. Okay, so what types of users or musicians do you envision, I mean benefiting the most from V2Window, what type you guys look after? So we have been around now for almost three years as a company, our product has been live, we have close to a million users who are using our product. Now mostly the biggest category that we see today is independent filmmakers, people who make like short films, documentaries, documentary series on YouTube, so these are the most common users that we see on a product. The second category is people who make like podcasts and games who require like in-pro music, background music while the game has been played and then there's a small subset of users who also use this for creating lo-fi playlists, so for calm ambient listening on YouTube. So yeah, YouTube is our biggest use case is what I would say, people who create content on YouTube who require royalty-free music, they are the ones who primarily use our product as a play. So on that royalty-free, I would like you to ask regarding that, could you elaborate on how beat-oven AI platform provides content creators with royalty-free, affordable and original music and how does the AI technology facilitates this process, I mean? Yeah, so just to break down on how beat-oven works, right? Like so we have one side of the business, which is the supply. So the supply is basically we work with a wide pool of artists, musicians, producers, so these would be like your music producers, session musicians, studio recording artists, and what we do is we do agreement with them and we also pay them and we do full licensed acquisition of data from them. So basically what that does is there's a transfer of copyright that happens between the artist and beat-oven and hence we own all of that data. That's a step one. Now the step two is we take only that data and we train models that are able to compose newer forms of music using those data sets at the building blocks. So that's the second part, so we build models for music composition, models for mixing, mastering, for production, all of the different aspects of music creation, right? So all of that is taken care of by the AI. So those are completely built in-house. And then that technology is served on our interface, which is beat-oven.ai, where the users can just come and sign in on a web application and start using it and they can give prompts and they can generate royalty-free music. So that's basically the whole end-to-end how it functions as of today. Okay, okay, okay. So recently Mansoor Sir, we are funding round-raised INR 11 crores with participation from several prominent investors. How do you plan to, you know, to allocate this funding to fuel beat-ovens AI, I mean growth across different aspects of the business? So one of the biggest challenges that we have discovered while building beat-oven is the expectation from the product is very high. In the sense, a lot of our users are global users. We are not, although we are an Indian company building out of India, but 96% of our revenues come from outside India. So then the expectation of the product quality, the music quality, music, all of those aspects are living very, very high. So we need to serve that need and the main purpose of raising this round was we want to invest a lot into the direction of product development. So we are hiring a couple of PhD folks, PhDs in AI music from Europe, who have already joined our team. So these guys are going to be working on the research side of things. So how can we improve the music quality, the music composition side of things? Then we are also going to be investing in our artist partnerships. So right now we have worked with close to 200 artists. Now we are going to expand it to 500 artists in the next two years. So more artists, so more data sets, so more content on which we can train our models on. So these are the two main aspects and a small percentage. We will also be experimenting on how we can distribute this product to more users, like in the marketing, not really marketing, but marketing experiments, just to figuring out a go-to-market strategy. So that's basically how we are planning to invest this much. So as far, your journey is concerned, so as an entrepreneur, what are some of the key or greatest challenges you have faced while building and scaling B2 on AI and how have you overcome them? So I think one of the biggest challenges being an entrepreneur is you are looking at things which the market is not really ready for. Like you are already kind of thinking about something, you are into the future and you don't know whether this is going to work or not. So you just have to keep telling yourself every day that okay, like you need to wake up and you need to keep doing this. You just need to keep building it in one direction. And I think consistency is like super important. For me, what has really helped is discipline. Like I have a routine which I follow, like I wake up in the morning, I do my workout exercise and then I come back, I get to work. And in that routine, what I figured out is that there's a pattern that emerges. Like I keep doing that every day, every little bit of progress. And that just helps me to consistently keep building small, small aspects of the business over a period of time. And that same thing also needs to be reflected in my team. And one of the most important things is what I believe is hiring. So you need to hire the right kind of people, people who are intrinsically motivated. That you don't need to manage. You know, like people who will do things by themselves. And fortunately at Vitoven, we have still our team, like all of the folks that are working are like people who are intrinsically motivated and very talented people. So these are very important aspects because I believe it's a people's business ultimately. You know, the people behind the product are the ones who actually drive the innovation. So it's very important for the people to be motivated and be really believe in what we are building in the long run. So the second part I would say is, you know, raising venture investments and funding, especially in this kind of a market, which is very challenging last couple of years, which you would have observed. And not just about that, also in the creative industries, right? Like India, like most of the creative things come usually from the U.S. You know, like you'll see like creative applications. But now slowly in India, we are starting to see some of that happening. So it's a new market. So you need a lot of convincing to do. You need to tell people, okay, you know, these are also markets that exist. There's a lot of potential for this. So yeah, I believe that that is a challenge that we have to face as the first generation entrepreneurs working in this direction. But yeah, I think we really enjoy doing it. Like, so that's basically how I took it. Yeah. So as you mentioned that you were a musician first and then you studied music tech and you wanted to build this company. But what made you think about this decision that I will not pursue singing in the music industry and in the music business, you wanted to do this? So how did this shift came? Right. That's a great question. So like my family has been very active in music. So I currently have 15 musicians in my family who are very big artists performing globally right now. But what I saw is music is a very, very challenging industry. You know, there are a lot of problems and through the whole process of seeing my father, seeing my own journey as an artist. I saw there are so many problems that exist in the music industry and I was very motivated to solve some of those problems. So basically, even before I was building Beto One, I didn't know that I was going to be building Beto One. My perspective was I want to build a startup in the music industry. So anything that has the potential to drive more value, more revenues into the music industry is something that I wanted to work towards. So the case in point, I mean, although Beto One is sounds like an AI that might replace musicians. That is something that people would look at it. But the way how we look at the business and the way we have built it is we are actually trying to build a business model where more revenues can come inside of the music industry. So you know, so AI is just like one new avenue of the revenue generation for the music industry. So my perspective was like, I want to build a business that will create more value for the music industry and solve a little bit of the problems in the music industry. Yeah. So you ensure that the music generated by Beto One maintains a high level of creativity and originality. Right. So there are two parts to it. One is we have models which are continuously learning and from user patterns, user behaviors, and also from the data that we accumulate. So we keep improving those systems to make them more and more creative over a period of time. So that is one way of doing it. So the second is, and that is also very feedback driven. Like we listen to users, customers, they keep telling us, hey, you know what this is sounding very loopy, this is sounding too monotonous. Can this be more dynamic? So all of that feedback we take into account and then we drive the research in that direction. That is one way of doing it. The second is we also give the user a lot of creativity. So it's not just about us generating music and giving them, but we also give them tools to edit their soundtrack. So they can customize, they can move things around. They can, like how you paint on a canvas and then you have the flexibility to paint whatever you have in your mind on a canvas. So similarly on Beto One, we are trying to replicate that in the audio domain. So when you're creating a music track, can you move things around and you know, like just pluck and play and the music practice and you shape rather than you wanting to know a lot about music? Like our users don't know about music a lot. Like they are, they're non-musicians. So how do you build a tool that helps non-musicians to compose music in a way that they understand? So that's basically how we give that creative freedom to them as well. Yeah, definitely. So as you know, AI is growing and in this time and it is generally a new market to explore. So how do you see the role of AI evolving in the music industry in the coming years for music business and where does Beto One AI fit into this evolution? So right now it's very early to comment because a lot of exploration is happening. Like not even the traditional players like the labels and all of the big, big guys have also not really hardcore ventured into this technology yet. You know, they're still very exploring. They're seeing like, okay, what can we do with this technology? But there are some patterns that are definitely emerging. So one is around creation. Like, you know, when you're creating music as an artist, the tools that can assist you to better create music in a faster and more efficient manner than second would be search-related problems. Like let's say there is a sound that you are looking for that you have in your mind and instead of searching through sounds, can you generate that sound? Right? Like so those are some other opportunities. And Beto One would fit in the second category actually. So we are actually trying to solve the search problem. So all our users already know like, okay, this is the kind of music I want. It should sound like this. This is how it should be. Can I just give a prompt? Or can I just give some input and get that output file directly from Beto One rather than going on a library and searching through hundreds and thousands of tracks and listening to each one of them, right? So the search problem is basically where we would fit in. Okay, that's okay. Yeah. Okay. So giving your expectations, I mean, in deep learning and audio applications, what are some of the emerging trends or advancements in technology that excite you the most? And how might they influence in the future of music tech? So from the way I look at it, like there are a lot of new models that have come recently, right? Like diffusion models, then transformers, large language models. A lot of that work initially was happening in text and images. Like you would have seen products like ChaiGPT, mid-journey, which caught a lot of attention. But that didn't impact much in the music market because of a lot of restrictions around copyrights, lots of data restrictions and all of those things. But I believe that now the technology is already kind of there to even disrupt the music industry. But my perspective is the disruption has to happen in the right manner. It's not just about technology and deep learning. It has to be in a manner that is also inclusive of artists. It shouldn't be like, okay, like, you know, you have built an AI that is taking away all the revenues and shares of the artists. Rather, it should be like, can it generate new revenues for the artists? So building business models like that is very important. So it's going to be more about business models and less about tech because tech is definitely going to get good. And we are already kind of there and probably by next one year or two years, you will not be able to differentiate between what is AI-generated music and human-generated music. You know, it is going to be that good. But then what do we do with that technology? Do we want to use that technology and put it to use directly and then just replace the artist? I don't think that's the answer, right? Because if you look at how these models are trained, they are being trained on artists' data. So the supplier of the data is the artist. So make sure that you compensate the artist rightfully. So that's basically the most important part for a sensible future and a more inclusive future which combines artists and creativity and technology together. Yeah. So on that only enhancing, you know, user experience seems to be a key focus, I feel, for B2Bin. So could you detail some of the improvements users can expect, particularly in terms of music accuracy, quality, and the design of editing features? What we can expect? Yeah, so if you see historically how we have evolved user experiences, like first we used to be a selection tool. So people used to select, okay, I want a one minute track in a cinematic genre, happy mood, two minute duration. So there used to be directly these buttons that you could select and you could put inputs and you could enter it. Then we evolved that to a more text-to-music chargeability style kind of interface where you can just type anything you have in your mind. You can also give us like big, big paragraphs and then we'll understand what you're trying to say and then we will generate music which is closely aligned to that input. But what we have realized is now the next phase is something called a multimodal. So we are going in that direction. What multimodal means is you're not limited by what you can give as an input. You can give either text, you can give us a video, you can give us a song, you can give us like any sort of input that you feel is relevant for us to understand what you want to tell the system. And on the basis of that, we will generate a soundtrack and give it to you. That will significantly improve the music accuracy. So basically people will get the output which is closely aligned to what they're looking for. And the second part is music quality. So we are investing in more partnerships with artists and also in research so that the quality of music is very high so that it sounds like music that we listen to and not just like some computer-generated robotic music. Yeah. So what are the collaborations and partnerships we can look for as you mentioned? So we are already like doing independent. So we are not like partnering with any big companies, but rather what we are doing is we are partnering with smaller independent artists. So a lot of these music producers from different networks. So initially it used to be from our network, but now we are also exploring international markets. So we are working with artists in Europe, we are working with artists in the US. So we have already started doing that. So these are the kind of partnerships. So it's independent artist partnerships. So our vision is we want to create a new ecosystem for independent artists. So in the long run, when Beethoven does become a big company, there will be a rod of the revenues that will be passed on to the independent artists that are becoming a part of the Beethoven ecosystem. So that's the way how we are going about it. So what is the, as you mentioned, the revenue? So how the revenue is generated through this? So we generate revenues using subscriptions. So basically we have two models actually. So one is subscription and one is a paper use. So let's say you're a user who's creating maybe let's say a podcast. So you will not require recurring music, right? You will just require once intro music for that podcast episode. So for you, paper use is a better model. You can just come and you can download one track and you can pay like $3, $6 depending on how much. So we charge on duration. So per minute or per second is how we basically charge you. So it's a usage-based pricing model. And subscription is more suited for somebody like a YouTube creator who's creating three, four videos a month who has a recurring need of creating more and more content. So they could just take a subscription plan and they can start using the product. So we only charge the user for downloading. We don't charge for using the tool. So using the tool is free of charge. Okay, okay, okay. That's great. So finally looking ahead, what are your, you know, long-term aspirations for Beethovenian? How do you envision the company's impact on the music industry and business globally? So our vision the next three to four years is we want to build the best tool in the world for music and sound effects generation. So anybody who's creating content and requires original music and sound effects and wants to be able to flexibility and also customizability, they should be able to do it on Beethoven in a very simplified manner. So that's the product vision that we have. The broader vision of how it will impact the music industry is we want to build an AI tool that is inclusive of the music industry. So all artists on one side that is on supply. So any artist who wants to work should be able to freely come and supply their data to Beethoven. And on the basis of how their data is being used in our models and how it is generating revenues, a certain percentage of that we want to give it back to the artist community. So that's the broader business vision that we have. And that is how it will significantly impact the music industry globally. Thank you so much Mansoor for your time. It was our pleasure to have you. Thank you. Thank you.