S1E1 – Rasa – Alex Weidauer | Journey of Rasa NLU, Rasa Core to Series B funding, Community building

Just another WordPress site

S1E1 – Rasa – Alex Weidauer | Journey of Rasa NLU, Rasa Core to Series B funding, Community building

Alex: When we were working on RASA NLU, or what became RASA NLU, the internal name for this was Parsa, so parsing texts And when we thought about like, Oh, we should really open sources, we felt like Parsa is kind of a weird name And so we were looking for alternatives and then ended up seeing that Rasa.ai was available as a domain And that just sounded like a cool name Kunaal: Welcome to the Co-Learning Lounge Podcast This is the first episode of the Rasa Podcast series Rasa supplies, the standard infrastructure for conversational AI, providing the tools required to build better, more resilient, contextual assistants With more than 3 million downloads since launch Rasa open-source is loved by developers worldwide With a friendly, fast growing community, learning from each other and working together to make a better text and voice based assistant Hey, if you’re new to this channel, please don’t forget to subscribe to our channel And if you like some part of the podcast, please, don’t forget to mention in the comments section below Today, we have with us, the co founder and CEO of Rasa, Alex Today’s podcast, we will go through the Origin Story of Rasa And look at some of the decisions Alex and his team made to build a strong community driven product company We also get an overview of how rasa evolved from Alex’s kitchen until today, along with the first hand experience from Alex about the recent series B funding Just so that, you know, the same investor also invested in technology, such as Facebook, Pinterest, Slack, Skype, and a long list of other companies And we’ll be touching upon many topics on the view of the founder of Rasa, Alex So let’s explore the origins of Rasa Let’s welcome, Alex, who is the co founder and CEO of rasa Alex likes to float on the beach, eat salad and loves coffee that goes along with communities Namaste Alex, welcome to the CLL podcast Alex: Namaste Kunal Kunaal: How have you been? Alex: Hey, I’m super well, thanks Kunaal I’m really excited to be here with you and chat about Rasa, opensource and the community Kunaal: So Alex, I’ve always been fascinated with how a day looks like for a super cool tech founder like you Can you tell us something about your daily routines? What do you follow and how do you give yourself breaks between hectic days? Alex: So for me, it really depends where I am, because I spent my time between San Francisco and Berlin And when I’m in San Francisco, my day usually starts very early Because you’re like nine hours behind Europe And so that usually means like getting up early and then going for a run to just kind of free my mind a little bit And then I usually started working very early, like 6:00 AM or 7:00 AM Pacific time And when I’m in Berlin, it’s a little different And so there, of course it’s not that I’m nine hours behind, which is actually a good way to start a day I usually also go for a run and then I meditate and then I spend some time reading So like non-work-related things And then I usually start my day later at like 10 or 11:00 AM Berlin time And so I have more overlap over Pacific time team So, that’s normally how its during the week, and then I try to on weekends, get out of the cities and go hiking or Do something in nature to really just switch off my brain Kunaal: Absolutely So even I, I do a run and a meditation schedule everyday myself You see the difference between the days that you don’t run your productivity dips or increases, you know, when you’re running Alex: Absolutely Yeah That’s definitely something that I noticed I mean, I will say that I’ve done that like running in the morning routine now for maybe almost like five years or so And I mean, some days I’ve, I’ve not done it And it’s definitely something that I feel, I think harder to focus for me if I don’t exercise in the morning And even like, if I go like travel for work, I always take my running shoes and go for a run It’s also something you can ever do do in every part of the world right, which is something that I really love Kunaal: Yeah, no, no barriers actually You got to it You got to it So let’s talk about RASA’s origin story We know that you and Alan met in UK at a startup weekend and out of boredom from educational degrees, started a company called Treev Just for the audience, Treev was launched in April, 2015 Treev is such platform for all your online drives such as Google docs, Dropbox, Trello, and more Also, it was well received fifth trending product of the day on Product So Alex, can you tell us about the parameters you used to decide to close the company and how you learn how not to make a product and how it led to the formation of rasa? Alex: Sure So, I mean, the, the facts that you mentioned about Treev, really the like highlights, I guess the story of three fast includes like just a lot of things that haven’t worked And I think for us, like Alan and me as the founding team, the biggest insight was that, although it sounded like a good idea in theory, in the end, the usage wasn’t really as good and like as often, specifically and regularly as we wish it would have been

I think that was like a major insight fast in the end around, although something can sound like a good idea that might not mean that people use it and really solves, like, as some people say, like, yeah, hair on fire problem The way how we kind of got to that conclusion was really fruit Just a lot of iterations and like being really strict about like what means product market fit for us and like how we would define a successful product and, and one key piece of what we’ve been building back in the days And it was like a freemium product And so one key piece for us was retention So we wanted to make sure that people not only like, initially excited about the product, but they actually keep on using that search bar that we build throughout their work days And I really just like constantly using it because otherwise you will have a hard time later on to monetize something like that And so that never really happened So we had, I think, 5,000 users or so in the end and maybe 10 people that were using it, like every day A lot, right And so when we looked at those people, we realized that it’s not really a big segment, so it’s not that they’re like 10 million of these people and we just need to find them It just happened to be people that had just like very specific workflows, I guess, and for them Treev work really well, but for most other people, it didn’t, and that’s also why we decided in the end to shut down Treev as a product And then we can talk a bit more about how we got into conversationally Kunaal: Awesome So, obviously, you know, when you’re building a product, tracking the usages is one of the biggest part of it with any product that you’re building now, you have a stronger analytics background that you track the usage very actively Is that how you’re doing it? Alex: It’s really a good point actually So I think that’s one of the challenges of building an open source product because in like Rasa opens us specifically, there’s no tracking whatsoever, right? And I think that’s something that also the community appreciates a lot, but you are somehow stepping in dark in many ways Right But then there are of course, other ways to look at product market fit and so fast, the way we think about product market fit nowadays is we call it like magical despite moments And so the idea is pretty much that you think about something happening despite your company being really small or like despite your company, having only two people and like, despite it, nobody has ever heard about you, but people started adopting your product And when we look at Rasa opensource, specifically, those despite moments have been there from the earliest days It’s just to give you one example, Rasa NLU was the first product that we launched in December 2016 And even before we announced it and launched it, we already had somebody contributing to the open source and making a pull request on the GitHub repo And so that was fast Like, Holy shit, how did this person even find us? And so somehow this person found us on GitHub and they were looking for like an NLU engine open source And I think just the fact that this happened is just I mean, and of course you can’t build a company only on this one moment Right So there were more moments like that, but it’s just an example of what we mean with a despite moment And I think we’ve just seen a lot of those for Rasa open source, and we haven’t seen any of those for Treev Kunaal: Interesting to hear I think so, they got a lot of eyeballs to that product immediately I also do do note to a follow up questions on like, once you launch Rasa right like these things that you did was hiring a first member of company right, so what does that look like? What was the value that you’re supposed to built add to the company when you’re first starting a journey? Alex: For us, it was really all, very organically Rasa And so when we launched Rasa open source, maybe a bit of back story, how we then even got into conversationally I so like Treev, this like search engine we just talked about, the idea for that for us was always to turn that into some form of like a system that you can chat to And it’s really more about like, I don’t know, asking questions, like where did Kunaal put the pitch take off Rasa or something like that Right So that type of questions we thought would be cool to build, but then when we shut down Treev We, of course didn’t follow on and didn’t work on this anymore But we got really excited by Slack opening up their platform for developers to build chatbots And that was really like the starting point for our journey and conversationally AI, and then the like very condensed short version of this is that we tried to build our own chat bot We use like tons of the tools that were in the market back then, like with.ai and like dialogue flow, et cetera And all of them didn’t really work for us And they all had like limitations And then we started building our own tooling, which then let us to build Rasa opensource and like the framework that we are still entered today pushing forward And I think for us, what we found really quickly was it, a lot of companies were interested in Conversational AI, but they didn’t really know how to do it Um, and so in the early days of Rasa, we actually did a lot of consulting projects with big companies And so the first hires to answer your question were really people that could help us like implement these projects So like back in the days, we had Tom, who started helping us out building projects for big companies in the early days And he was like back then still working for his own company that he had like a, he cofounded an agency and he started working part time for us And then overtime started working full time for us Now, Thomas is our director of engineering

So he’s been one of the first people that joined and then we also needed somebody helping us out with like project management And so we had actually Philip, who back in the days, he just finished his master’s degree And I still remember like he called me because we were, we still friends at that back of the time And he was saying like, you know, what should I do after my master’s degree? And then I was like, Hm, I have an idea for you And then I said, Oh, well, you know, we were looking for somebody helping us, all of these projects And so Phillip started joining part time and now he runs our business development function And then, you know, just other people that we kind of knew Yeah And over time, I think when we were 10 people or so, we also realize that we really want to have somebody dedicating their time, full time on the community and like all the work that we’re doing with contributors And so helping those people contribute more, but at the same time also to just creating more educational content for developers And it’s really when we started looking for like a developer advocate and then found Justina who has been part of our communities, she’s been a contributor before she did like this amazing, like two hour video on how to build an assistant with Rasa, and we’ve just been really blown away by how well it was done and like other uses like a great fit for this role And there was like two years ago and then she joined and we’ve been just like blown away by the work that she’s been doing And she now started building out a team as well of developer advocates and leading that effort So, those are like three examples of people that joined early Kunaal: Awesome to hear, you know, some of the new folks joining in and then growing with the company, even I’m a educational content maker, and there’s a lot of importance to bring the content in different ways so that you have a larger adoption by communities And they know, Oh, you know, we could use this product in this way and that we have And I think since you are a community focused person, you see that, you know, this was adding a lot of value And hence bet on that early on When you first bot, that is the data bot in 2016, you used Slack and Facebook, the APIs and obviously With.ai Then you switched to building a wraparound spaCy And just so that audience knows ,spaCy is an open source software leverage for advanced natural language processing written in programming language bytes, and ensighten Now, what are the biggest challenges conversational AI is facing currently? What do you expect from the open source NLP AI community spaCy having space to uplift the overall chat bot ecosystem Alex: Yeah Great question And I think a lot of stuff has happened ever since, right? Since like didn’t have years ago, since we first open sourced Rasa NLU Of course, I, especially like NLP research has been going really strong and I mean as you guys might know, we also doing a lot on our end on that side with this whole idea of like, bridging really the gap between like the latest and greatest research and production and like what developers can actually use and make sense I mean, I think the earliest days of Rasa were all about this realization that as much as like NLP is important for conversational AI and it’s like one of the key pieces, it’s really not just all about NLP And so when we think about conversationally, I they’re a bunch of like, I think components for us in all of that But in addition to like understanding what somebody said, which is all about like NLP, it’s of course also all about handling more like complex and like complicated dialogues And that’s just something that we’ve early on as developers realized it’s extremely hard to build So it’s not just about understanding what somebody said, but it’s also like, what do you respond to that? Right And I think that, what do you respond to that part can also be like split up in like really the dialogue logic and like what you decide happens next And then What exactly you’re going to say And I think all of those parts, like constantly developing and evolving Yeah And our goal is really to make sure that Rasa open source is always cutting edge and allows developers to use most state-of-art components And I think like in this whole space of NLP, a big topic, of course, in the last couple of weeks have been all about like GPD free and all the great work that open AI has done there around like specifically language generation And that’s of course also like an interesting component, right? So I guess I will say like lots of stuff happening, especially on the library level, on like spaCy, I mean, you mentioned hugging phase with transformers There’s also flare, which is a really amazing open source NLP project So there’s so many things happening and we see those folks really as like innovating on like one part of the stack, right and fast, we really innovate on other parts, right Which are, for example, around dialogue management And then we of course have all like NLU components I’m not sure if you guys follow the developments around diet and the new classifier that we built there And so I guess it’s constantly moving field and it’s just really exciting to be part of it at the moment Kunaal: Absolutely Even now I was just going to some of these GPT-3 articles Interestingly, they are even generating SQL codes and generating codes overall altogether Interesting to see that, you know, we’ve come up with that classifier, which is again, a great addition to whole ecosystem So, Rasa got immediate traction on tech crunch, post launch in December, 2016,

and I’m sure you you’re monitoring the first few users, 400 users closely What are the biggest learnings that you got from these first early adopters, now that is a big part of Rasa? Alex: One of the big things that we saw early on was that Rasa NLU was used in very large companies, actually So one of the first emails we received, that was from one of the largest banks in the world saying that they have deployed Rasa NLU in production and I think that’s something that we’ve continued to see a lot Right We’ve continued to see that large companies are very interested in the topic But at the same time, I think struggle with how to operationalize it And so with that, I mean like what tools do you need to like, not just like explore something, but also actually like create a whole process around, like, in our case, like building an AI assistant with a team and that’s something that’s informed a lot of our product work and product decisions, and specifically around like Rasa X, which is it’s like tooling layer on top of Rasa open source where we see that it’s been used by Teams And I think those teams have like all workflows around that topic And so from the earliest days, there were lots of open source projects as well, and provided like some of those workflows, for example, like labeling training data, there were a lot of, still are lots of open source projects build on top of Rasa open source, which is really awesome And I think what we’ve been seeing is that, especially in big companies, you want to make sure that these like workflows are all sync up in a nice way And ideally don’t involve like an Excel spreadsheet, which I think a lot of people still use for it for labeling training data And that’s something that I think in addition to what I just said, that conversational AI, it’s not just about NLP and there are other components, like dialogue management, as well as like other like language generation parts It’s also all about like workflows, right and so I think that’s something that we’ve been seeing a lot is that every company does it a little differently And I think something that eventually will find a way into one method of doing things And so our current thinking on that is actually something that we talked a lot about during developer conference Alfred AI, but also Alan has been talking a lot about that and this is idea of like conversation driven development And the idea is really that you start with like an early prototype of an assistant and then through real conversations, improve it and learn from that And I think that’s something that’s still, a lot of teams are struggling with like imagining how this all could work And so it’s really early of that Kunaal: Awesome Not only are you working on some of the later stuff with it, but you’re also looking at pushing the boundaries in some of the latest barriers that the industry is currently facing in terms of chatbots So, was the initial launch we know, you’re funding yourself by providing chatbot services to companies But later you stopped and focused on building the product company with a strong focus on community driven contribution How do you think the company and the community will benefit from this model? Alex: It’s a really good question I think in the end, the way how the community benefits or has been benefited so far was that we’ve been able to just pull off a lot of things that we wouldn’t have been able to do if we would have to focus on services revenue, and working with customers, or like with services customers And so if we look at that, right, if you think about that, the Rasa master class, for example, which is a huge project that we started last year in December, now it’s almost like a hundred thousand views on YouTube And that’s something that I think we wouldn’t have been able to pull off if we would have had to like focus on creating constant services revenue to finance as right And so that’s one, I think the other big piece is all about like research and specifically everything that we’re doing around tech machine learning research I think that’s something that is also harder to pull off if you have to focus on like day to day services projects So in the end, and that’s really also the reason why we decided to go the venture capital route was that when we looked at what we are doing and when we looked at our mission of like pretty much enabling all makers in the world to build the best assistance We felt that this all needs much, much better product, much, much better research, much better education and community really get there And we just didn’t feel like we can do that as fast and like fast as it should be with financing as like this kind of bootstrapping idea around services project, which is why we decided to raise a seed round in 2017 and since then have done follow on financing rounds And with that, I think the community really benefits from a better product and more education around it which is something that I think at that scale wouldn’t have been able to pull off for us Kunaal: Awesome So, Alex, while you’re mentioning that community driven and the education part of it, we have good listeners here today with us, so what contribution do you think RASA needs currently? Not only in terms of the product development, but also in terms of educational video So where do you see some gaps that you think the community can fill in both of these areas? Alex: Yeah And so actually we, we like last year introduced this contributor project on top, because what we noticed is that there were a lot of people really interested in like contributing back to the community, which is awesome and

a lot of people have done it already I mean, we’ve over 450 contributors now, globally, which is super cool But what we noticed is that like, either people tend to not know where to get started and because there’s question just inside, right? So it’s like so many things to do, where do I start and what is really needed? So that’s one And then on the other hand, we also have seen like, people just duplicating content, which is not that they like copied content or so, it’s more like they had the same problem, like somebody else, but they couldn’t find the solution to it So they just wrote a blog post about it, for example right And so both of those problems we tried to solve for this contributor board, which is something that, I mean, if you, if you can, somehow we would love for you to link to that in the podcast, somehow, maybe in the notes And this is really where, like we have up to date information about that those are the things that are needed So maybe just to give a quick overview on what are the things we think the community needs So there are a bunch of different things Like one is really all about like organizing local meetup groups So that’s something that We’ve seen working really well in some countries and some cities and that’s of course, super exciting We’ve seen a lot of amazing contributions around that in India So that’s really, really awesome And that’s like one kind of area And depending on where you live, there might be either already Rasa meet up or not And so that’s one Then more around like specific educational content I think what the community always appreciates, it’s like more information on like more like the machine learning layer on the underlying layer on that So that’s something that is always really appreciated And then of course more on the code contribution side, we also have a bunch of like open github issues there, but in general, I think anything that you face as a challenge while developing assistant, we would love for you to contribute back in some form And I think the best way to figure that out and see if that fits into the product roadmap is always through opening a PR on the gateway Kunaal: Absolutely We definitely provide the links in the description section for anybody who’s listening to this, can go there and dive right into whatever Alex has mentioned in terms of contribution, and obviously with the community We have seen Yogesh show his energy in terms of contribution in lot many ways, in fact that’s how me and Yogesh actually contribute back to the community And we’ve done some of the video programs on Rasa and it was a good success for us working on that particular model So obviously now October, 2017, at that point of time, you had 30 K downloads, over 300 members, as you mentioned, over 40 contributors and tons of thousands of active developers You open source Rasa Core And just within seven months, you grew to over a 100K downloads, over thousand members on Github and over hundred contributors How did you handle the pace at which community was growing? Did it overwhelm you? Was this the time you hired more employees and grew significantly? Alex: It was definitely, really wild times in many ways And, uh, I think like just on a lot of different levels, but it was really just like so much pull from the community And so we ended up, um, in, in that time, uh, raising the seed round and with that money we raised like around a million dollars And we were able to hire more people and that helped us on, on the product side But also, as I mentioned earlier on the community building side, you see now, for example, and I think that time fast was all about just like keeping up I’m trying to make sure that people that are using Rasa, I happy and we can fix backs really quickly and all of that And at the same time also really like figuring out a way for us to like help teams even more, especially with like more tooling Because as I mentioned earlier, like we saw a lot of different tools popping up and I think its like super cool to see so many tools being developed in the community But then we also see a sided, some patterns, like, for example, if you think about like, Labeling training data It’s just something that everyone has to do And we felt that it probably makes sense for us as a company to provide that in some form So in 2018 or so we started experimenting like tooling, which ended up like being Rasa X and in 2019 Um, but yeah, lots of things happens, but mostly around like hiring and making sure that we can keep up with the pace Kunaal: Absolutely Alex and I remember 10 years back when I started doing some amount of NLP back in the days with accent, labeling was a big thing and it was definitely paid for, I wish there was some tools that I have to label it properly So you mentioned about your Series A funding, which is roughly around April, 2019 You got a funding of $30 million And just for the audience by this time, Rasa has over 500K downloads and over 3,500 members on forum and over 300 contributors Now I wanted to bring out an interesting aspect, which I’ve always wanted to know is what goes on behind procuring this funding and how much of your bandwidth is occupied to bring all the aspects of whatever is needed to ensure you’re procuring the funding So what are the major learnings there just to ramp up the time between the funding you procure to just, let’s say a few months before that? Alex: Yeah, really good question I think it’s probably like something where we can spend like a whole

podcast episode on as well Yeah But I’ll, I’ll try to keep it short I mean, the key to all of this is to have a product that people want and if you, and if you have that, it’s like, I wouldn’t say easy, but it’s definitely, it definitely makes the whole thing much, much smoother And I think in our case, We just had like really strong traction on the community side So, I mean, the numbers that you just mentioned, but also in general, just I think the amount of like adoption that we’ve been seeing for Rasa has just been really phenomenal And so that was something that we could just show and see, like, you know, this, this is something we build something new people want it now let’s just get more fuel, kind of to grow faster It was kind of the overall storyline But also, we’ve only been seeing like some financial traction as well around revenue So we in like 2018 launched a super early version of an enterprise product and we started selling that a bit And with that, got like some good revenue as well And I think those two factors combined We’re like good arguments, I guess, for, for investors to consider giving us money And I think if you don’t have that, it’s just, it’s really hard right I think it’s just hard to convince anyone to put money in, in our case what happened then a bit more specifically, I was then still living in Berlin and I was planning on going to the Bay area to meet a bunch of investors And so I had like in early January or mid January or something like that, I had this one week where I had 20 meetings lined up also with like investors And so it was kind of interesting because I didn’t really plan to raise So we, it was more like we wanted to get to know people Right And, and I think that’s something that I would always recommend is like building relationships early with investors And so we were really just planning on like, just meeting people and building relationships and then maybe raising in like three months or something like that But then what happened is that people got so interested in what we are doing I think it was just a good time It was like early 2019 AI was like the cool thing, right Everyone wanted to do an AI investment and we looked like a good one, I think, around that time And so we just got like what is called a term sheet very, very quickly in the process within a few days And what that means is pretty much take a term sheet it’s like here, I give you money that amount of money for the valuation And those are the terms And still with that, I think we just created so much interest in Rasa that we got many, many more afterwards and within a few weeks, and then ended up finishing this process really fast and like two or three weeks in total And it’s probably not normal and in terms of how fast, but it’s generally the process, right? I think you want to talk to like a lot of different people Also not so much to like, I mean, of course, to increase the chances of getting funding, but I think what’s really always overseen Is that in the end? Yeah It’s like your decision who you work with, as much as you want to appeal to investors, I think it’s much more important that they appeal to you and yeah, you have a good fit with them and they are aligned with what you want to build And I think in our case still in 2019, I think building a horizontal AI company, like horizontal, as in all use cases, all like, you know, industries, all geographies, what’s something that not many people believe is possible So the bet at the time was all about verticalized AI solutions because people felt like, well, companies are not really able to hire all the talent to build their own AI And I think that really changed dramatically And I mean, we already saw changing back in the days But I think more and more people realize now that this is something possible to do Kunaal: Absolutely And I’ve worked with some of the software engineering companies, and I’ve seen a lot of teams try and fail And on this critical thing, you know, some of the things that they do is build up a stack of questions or questions and answers that that can be used in the chatbot That was the primary method But you know, when you actually saw that technology, it’s just a stack of questions You know, moving forward, you also, you know, secure our Series B funding, 26 million alert by again and Racine and before the Series A it was by Accel, just so that the audience knows here Accel is same company that funded Facebook, Flipkart, Slack, Spotify, etc And all of that, even in Series B, when we have Andreessen Horowitz They’ve also invested in Facebook, Pinterest, Slack, Skype, Github, Flutter Instagram, and the list is going to be a long list for these people So firstly these people on this end, the Series B process was much more faster as you know, now when the Series B funding came in, what these new investments and investors bring on the table for Rasa? So, what does it mean for Rasa when these funding come? Alex: Yeah I mean, so in addition to the, the money, right? So in addition to the dollars, which we will pretty much mostly deploy on anything around like research and development So we just opened a research happened in Edinburgh and hired Adam there from the university who already trained the team July 1st and like just investing even more in develop education So like, in addition to the dollars, the reason why we decided to work with Andreessen Horowitz was really two fold I think one was that there’s a strong belief in the mission that we have

and like making conversational AI technology accessible to all developers and just like really big focus on community and educating developers around the world and kind of agreement that this is right way to go There was something that was the most important thing for us when we picked an investor and the other run is really more about this specific person that we work at Andreessen Horowitz with, which is Martine Cosato and Martine himself like has been a founder, he actually build an enterprise software company and sold that to VM-ware at some point in time And then joined Andreessen Horowitz as a general partner And Martine is just like great human being, which is of course awesome But also he has just seen a lot of, yeah, the open source companies being built over time and he’s on the board of, I think, six open source companies or something like that And I think with that, that will help us also just build an even better company Kunaal: Absolutely all that experience definitely on the table So obviously speaking of community, when you have all of these things aligned, the strong community focus, I wanted to bring our attention to how you work, to bring the whole community together? They develop a community together with events like at L3-AI So how did it quickly adapt to going from a non-virtual event which was almost going to get canceled and then move on to virtual event because of the Covid situation I’m sure you face a lot of challenges to bring the community together there Alex: Definitely And I mean, internally our marketing team has done a phenomenal job at managing this shift and doing things even better than I think we all could have ever hoped for So a little bit of backstory I mean, we, last year in September had our first real developer conference in San Francisco with around like 250 people or something like that was really packed in a huge room in SF was amazing And I think the energy there was just really incredible And I also think that for the attendees, it was just a great In many ways, right? I mean, it was just meeting other community members and hearing their stories, like learning more about conversationally So I think that’s been just like a huge success and we’ve been thinking about how we can scale that even more And so our original idea for this year was more to have like a bunch of those developer summits in different locations in the world And then of course COVID-19 started and realized that this won’t be possible this year And then ended up thinking about how an online conference could look like And what we really realized quickly was that it will be possible to, first of all, get even more amazing speakers because people don’t have to travel But also second of all, it will probably like attract more people all over the world And so we were thinking that it might make just more sense to talk more about conversationally AI from a developer angle in general and not so much only just about Rasa and that’s what ended up being, or becoming the L3-AI conference that we had in June which has been a huge success for us We almost had 5000 registered people and had speakers from all over the world, like the founder Um, I mean, you mentioned spaCy earlier, right? So, you know, this was there, the co founder of Explosion AI company behind spaCy Yeah, that was awesome Then we had Adam Chen, the founder of with, and Siri actually giving a keynote and tons of other amazing people And I think for us, the big lesson learned there was also that These on any events really fit better into our mission over all, all that stuff, accessible to more people, because it’s just almost impossible for folks to travel to SF if they are based in other continents Right It’s just, it’s just really expensive and it’s really far And I think if anything, we’ll double down on that online content on that like online event content, because I mean, as you probably know, we’ve invested a lot of time and effort into more like developer content on YouTube specifically Kunaal: Absolutely Alex I agree with you And the result of this, as we speak today for Rasa, Rasa has been adopted by global communities Then, you know, as we speak, RASA has over 3 million downloads over 10 K forum members and, over 450 active contributors So, you know, if you want to do something fun with you, Alex, we have something It is called the rapid fire round So let’s start the Rapid Fire Round for this particularpodcast Here’s your first question What is the best message you got from a community member? It can be personal or from a forum Alex: I just got my master’s degree where I used Rasa in the master thesis I think that’s been really cool And we got a lot of those messages Kunaal: What are the best message you got from clients who use rasa? Alex: The latest diet architecture just improved our accuracy by 10%, just switching one parameter in the conflict of the ammo file Kunaal: What all contributors, you know, by their name and their country? Alex: Yogesh, India It’s an easy one Then Julian, Germany and Nathan, Canada, and many more, but so many of them Kunaal: So here’s another one Micro management vs macro management What do you prefer? Alex: Macro management Kunaal: Early morning coder or Late night coder? Alex: It depends where I live

If I am in Berlin, late night coder And if I’m in SF early morning Kunaal: Bootstraped vs Funding, what do you prefer? Alex: Oh, that depends I mean, both can be great, but coz right now funding Kunaal: Define RASA in 3 or 5 words? Alex: Making conversational AI accessible to everyone Kunaal: What is the best bot which you have seen so far, and what’s so special about it? Alex: Lemonade, actually the insurance company from the US and I think what’s so special about it that said they integrated it extremely well into their existing mobile app, and just thought a lot about UX and how that all works for the user Kunaal: So with this we come to the end of the Rapid fire round So Alex, I had another question like also with the chat bots and all of that thing You think chat bots can evolve to a place where you ask it some business insights and it produces some of those insights Does this analytics and provides you those answers? I’ve always wanted to hear if that can be possible to a scale Alex: It’s a really good question I mean, in the end, I think, yeah, we have actually like talked a lot about this in the last couple of months about like the five levels of AI assistance and they, we just recently also updated And the way we look at it is said over time, pretty much the higher you get in those levels, the less you have to adapt to the assistant as a user Right And so to give you an example, if I think about like a level two assistant in FAQ assistant, you still have to know like what FAQs you can ask But also in most cases you still have to know like what information you need to provide in order to get an answer And so if you get all the way to like level five, This is not needed anymore Right And back to your idea of, can I get these like businesses insights through just talking to a conversational AI? I think that will be possible, but it will be only possible We believe if you connect the assistant to like other systems And so in this case, if you think about like Looker as like a BI tool, for example, probably like a really experienced local user is able to get the insights and the questions that you just had about like, how’s my business going or something like that They probably able to get that information, but the reason why they’re able to get it is because they understand what important metrics are for your business But at the same time, how to use the UI of Looker to get those pieces of information, I can do it to really allow any user to become an expert for any piece of software without knowing how to use the tool And yeah, I think that’s really exciting, but it also would mean that it has to work with other tools and systems Kunaal: Absolutely, Alex And also obviously when you’re doing a lot of things will evolve as you go forward Okay I just wanted to ask some more questions around RASA, how did you come up with the name RASA and what is the current focus in terms of future for RASA? Alex: Sure So the way we came up with aim was really that when we were working on Rasa NLU or would became Rasa NLU, the internal name for this was Parsa, so parsing texts And when we thought about like, Oh, who is really open sources, we felt like Parsa is kind of a weird name And so we were looking for alternatives and then ended up seeing that Rasa.ai was available as a domain And that just sounded like a cool name to us But also I think in the context of like machine learning, it kind of makes sense because tabula rasa and Latin means like a blank slate And so yeah, of course, in machine learning, it all depends also on your training data And so that was something that we just felt like makes a little sense and I think people appreciated it so far I think what turns out is that like a bunch of different Indian languages It always means like something really good and like something I think even King in one of them And I think that’s been funny, right? I mean, we, we did not know that at the beginning, in terms of like what’s focused fast now, I would say it hasn’t really changed much over the last one or two years fast And so our focus is really like investing in spending a lot of time on product development and research So that’s something that we will be doing even more now with the funding that we just secured And then the other big piece is really educating developers on building conversational AI And it’s a long journey and we’re really excited to see that community is helping a lot with that And it’s just super cool to see like all the amazing efforts that have been going on I mean, just to name one, of course, with the Co-learning launch on YouTube, which I think is also something that you guys probably post here in the comments It’s just super cool to see that you guys have been doing such a phenomenal job there and yeah, it’s a lot of fun and a long way to go Kunaal: Absolutely, Alex We are excited to grow with RASA So for the audience here, I like some of them want to know personally certain things about you So here’s a light question for you Do you watch any Indian movies? And if yes, you have any favorite actor Alex: I don’t, but I would love to start actually, if you can send me any recommendations, I’d love to love to check it out I generally don’t really watch a lot of movies and shows I’m always really behind on like the latest Netflix shows And can’t really talk much about that, but I’d love to check out some movies from India Kunaal: Awesome, so I guess Yogesh will send you some of the links Okay

Okay So last one, we know you’re going on a vacation and we just wanted to know, what’s your ideal vacation looking like? Alex: I mean one big piece is really just entirely like switch off and not work, which is hard if you run a company like Rasa But that’s really important to me And then the other big piece is really just being outside in nature and just, yeah, I think some, some like hiking or just in general being around nature Kunaal: Awesome I’ve gone to a lot of hikes with my brothers here in India, and I cannot relate to the feeling It just so serene and clean Awesome So with this we come to the end of Episode 1 of RASA Podcast Series with Alex Thank you for sharing your thoughts and bringing the RASA community together and making it a better place for a lot of contributors Alex: Yeah Thanks a lot for calling me But, I mean, it’s been a lot of fun and I’m really excited that you guys are starting this and let us know how we can support further just for everyone listening, really make sure to subscribe to the Co-learning Lounge on YouTube that I just mentioned and join the community there, as I said they’re doing a fantastic job and its really exciting to see our community grow across the globe and make sure to check that out Auf Wiedersehen Kunaal: Auf Wiedersehen