Welcome to the upside podcast powered by upside global and hosted by Julian Blinn, founder and CEO of upside global. The upside podcast is listened to weakly by over 6000 sports and tech executives from all sports leagues and teams in the United States and around the world. Julian has been developing technologies for professional sports teams for over 10 years, and has worked for major tech companies, along with sports tech startups. In each episode, Julian interviews, global leaders in sports to share knowledge on emerging technology in the sports industry, and how these technologies can help improve the performance of individuals and organizations, both on and off the playing field. And now here's your host, Julian Berlin.
So today, we have the honor to interview Sam Liang, CEO of Otter.ai the leading transcription and summarization platform for enterprise collaboration. So Sam, welcome to the show.
Thank you for having me.
Great. So what thank you for being on the show. So hey, Sam, what I want to talk to you about today is first we'll talk about your background, or your company and your product. And then we'll discuss the benefits for specialization to use your platform. And then we'll talk about your plan for the next 12 months. And then lastly, we'll touch on chatgpt and generative AI. How does it sound?
Great. So maybe, just to get started, could you tell us about your background.
Of course, I was born in Beijing, China, I grew up there, studied computer science in college, then I came to America in 1991. So it's more than 3032 years ago. Later, I did my PhD at Stanford, focusing on large scale distributed systems. After that work in few high tech companies, I joined Google in 2006. At the Google map location platform, which powered the Google Mobile map app, which is one of the most popular apps in a world google in 2010 to build a mobile startup in Palo Alto. It was later successfully acquired then in 2016 with my co founder Yun Fu, we started Otter.ai in Mountain View.
Okay, and how do you get the idea of your company? What was kind of the the epiphany
there couple of motivations. One is building a startup, I have to talk to a lot of people, I have a lot of meetings ever, I found that it's hard for me to remember so much information, it's hard to remember my action items. Hard to share the information with our team, I talked a lot for capitalists, I talk to customers, when they need to do recruiting to build the team, so tons of conversations. So there's a big pain point for me, and arch and our team. On the other hand, technically, I'm fascinated by voice and the tremendous amount of voice data in the world. If you look at human history, before audio recorder was invented by Thomas Edison in the 18 Watt, or APNS is seven, I think, no audio data was even captured even after the recorder was invented. Still, most voice data was not captured. And for those that was recorded, most of them didn't have transcript. So it's very hard to search for information. And not, not to mention analyzing the data. So we see that there's a tremendous amount of value if we capture voice, transcribe it and analyze
it. Yeah. And that makes sense. So that's great. That's how you get the idea. Now, how big is your team today? And I know you raised quite a big amount of funding, right? So can you talk about that as well.
We raised $63 million. So far. It's not the biggest in the startup world, but we are actually super efficient. We have about 100 people that we deal with the best speech recognition engine and the summarization engine in a world. We have transcribed and more than a billion meetings already probably one of the top three in the In the US, the work with Zoom platform Microsoft Teams Google Meet. So we are platform agnostic, our mobile app is also popular on Apple App Store and Android so that if you meet someone in person, you can use Otter as well. So everything is saved in one place, no matter where you are meeting.
Yeah, that makes sense. And I have seen some sports teams actually using Otter AI before so but how would you describe your product? You know, what are the use cases? What's up to users? In the context maybe of teams, right? I can tell you in sports teams, as in researchers, head of performance, using your platform. So how would you describe in a very simple way, what your platform does your product, how it works, and also the benefits for a sports team to use it?
Yeah, the, the UI is super simple. If you synchronize your calendar with Otter. And whenever you have a Zoom meeting, Google Meet meeting or Microsoft team meeting order will automatically join your meeting at the beginning. It provides a live transcription, it captured any slides used in the meeting automatically, you don't have to do anything Otter use of lesion algorithm to detect the slides and put them into the transcript. So your transcript is rich. Note, you can highlight anything important highlight the action items, we're working on new algorithms and all packet detect action items. And at the end of the meeting, Otter will provide a summary for your meeting. It also recognize each speaker and how much does each speaker talk and what are the key words each person discussed? The lot of enterprises are using their for their team meetings. Yeah, as you mentioned, you know, a lot of researchers whether they're in the infantry or in the sport worlds probably no coaching can be captured Otter as well. We have a couple of investors actually from the Super Bowl
I remember that. Yeah, I think there were some top athletes right as well. Within a quarterback or football player, remember,
Harris Barton who won three Super Bowl Super Bowls, with 49ers. Also, James Patola is a raptor group. You know, he is a co owner and Celtics with the team.
I know. I mean, I don't know of him personally, but he gets our newsletter is pretty. You know, he's an avid reader. So that's great.
Yeah, they are investors in Otter. That's great. So
how would you what would you say is your competitive advantage compared to maybe other platforms in your space?
Number one, we have a very advanced AI system. You know, we started building this since 2016. We focus on speaker conversations and this actually more than just system used in Siri or Amazon Alexa? Because those products can only handle one speaker. We actually handle multiple speakers and a long form conversations. This is important because in the meeting, which has an hour and can involve, you know, five or 10 speakers, we need to understand who talk about what and then the summarization is complicated to understand all the key topics people discuss. Then, yeah, with you, were going to ask about chatgpt We, we've been working on this conversational AI since 2016. So definitely a lot more to do. As I mentioned whenever you viewed a AI data is critical you know, because we transcribed You know, we're in meetings already. Every day we're generating a lot more meetings, that can definitely contribute a lot to the AI training system, because without enough data, you cannot build good.
Yeah. And, you know, we'll get back into chatgpt, hot topic, but what would you say are the the things that your customer liked the most about your platform.
It's the user friendliness it's convenience. As I mentioned, once you set it up, you know, you don't need to remember doing anything, auditors join your meeting, on your behalf, even for the meetings, you cannot join personally. So people actually use this to actually skip some meetings that don't require their live contribution. A lot of you know, we have done research and study, that show that they need 30% of the meetings in enterprises are not absolutely necessary for every person to join. So you know, everybody is suffering from mental fatigue these days. You know, skipping some meetings, but catching up using a meeting summary provided by Otter can reduce people's fatigue and actually improve their productivity.
I agree, I think it's a great, great data point, I never come across that 30%. But I'm not surprised. You know, I even myself get tired sometimes of getting input into some meetings where I'm like, you know, what's my contribution? You know? So, could we maybe talk about a bit about your business model? Right. So how would you describe your business model? Is it a licensing fee? I know that a free version, right, of your product? So can you talk about that as well,
Or Business model is a freemium SaaS model, anyone can sign up for free, or using it for free? Once they like it, and they need to use Otter for more meetings and use more advanced features, they can buy a subscription. So our business model is actually quite similar to many other successful SaaS companies like Slack, Asana, Figma. Once people start using it, that they share it with their colleagues, because the meeting notes are usually more useful when it's shared with all the meeting participants, then it starts the viral loop. The product grows by itself when people use it. And then enterprise is in a what they see they have hundreds or possibly 1000s of employees who are already using Otter. Many of them decide to buy an enterprise license,
and what what's the range of the enterprise license. The price range?
Oh, it depends on the feature set. You know, it could range from, you know, 20 to 30 to $40 per month per user.
Okay. That makes sense. That, could you tell me about your big your plans for the next 12 months? Are you looking to add new features, introduced new roadmap items? You looking into maybe raised some more funding or expanding to new geographical areas? What are your plans for next 12 months?
Um, several areas, we're investing a lot more into AI to build more intelligent AI systems to better understand even better understand human conversation because obviously, conversations can be complicated. Especially when you have multiple people involved it used if you look at chatgpt is a one on one chat between a human and a robot. It's very simple, interactive. Once you have multiple people talking to each other, it's way more complicated. How do you know who agree with whom who disagree? I mean, to what degree and what's the sentiment? Well, so our call, compared to chatgpt is all kinds of speaker chat system. And also, we see that Otter can participate in the future in Live Meeting, instead of just passively taking notes. Otter can answer questions in the live meeting, so those are the technologies that we're developing. Secondly, if you look at chatgpt, is trained on public data, it doesn't proprietary information in any enterprises. So if you ask chatgpt, anything related to your own enterprise, it doesn't have any knowledge.
is we're building a system that can incorporate propriety proprietary information into the system. Of course, all this data is encrypted, you know, we, in in Otter people in Otter can not see their data, but the machine can respond with their internal answers when they ask Otter any questions. So this will just further improve people's productivity and provide a chatgpt service type of service in enterprises, which have their customized knowledge.
So can you give me an example of how because since your system is using proprietary data for the enterprise, customer, right, so if I'm, let's say I have a live conversation and Otter the AI is listening to it and doing a transcript? What types of example? Or scenarios? Could we imagine with the Otter.AI using those types of generative AI? Can you give me a sense of how it will be different?
Yeah, let me give you an example. For our own company, all my, all our meetings in the last seven years are in the Otter system. Okay. If I asked Otter in doing a meeting, we're discussing a product release, we said, hey, what's the status of that bot? chatgpt doesn't have that information. But our own system because it hooked up with our own bug tracking system. Otter can look up the status of that bug can can provide the answer by the way. On the In addition, we, in our own system, we actually have knowledge of all the employees, you know, the names of each employee, the role of every employee. So if I mentioned a person that said, Michael will do that. So Otter actually knows which Michael, we're talking about, and it could just assign an action item to that person in the system.
Understand? Yeah, that's great. Well, maybe you could say something like, Hey, can you check Michael's availability? Maybe to the next meeting anyway, we will check the scheduling system and see when maybe that when that person is available, correct? Correct. Okay. Yeah. I mean, I think it gets very powerful. Right, once you have that level of intelligence, like you said, having the proprietary data, right, because you guys have had that right from some of your key customers. Right. So I think it's powerful.
Yeah, yeah. And as I mentioned, everything is encrypted. It's confidential. You know, the machine can do the lookup that, you know, we were not able to see there any proprietary data.
Yeah, that makes sense. I mentioned also, are you guys looking to raise maybe some additional funding or focus on new verticals or geographical area?
We don't have immediate need, because the company is super efficient. We we still have a lot of money in the bank. So we're not actively fundraising. The you know, focusing on enterprise adoption now. The we have more than 10 million users and it's rapidly growing. Yeah. That's great.
That's great. So, you know, going back to chatgpt. Right? So why do you think there's so much hype? About chatgpt? Because it's not really it's not anything new. It's been out there for a while, right. But now there's such a hype about it.
It is super impressive in the sense that it is a, it's able to understand your question, provide and generate a coherent answer, which, although it's not always correct, sometimes it sounds authoritative, that the information may be hallucination, but it's still the way it is and understand your question.
And generally answer and then when you give follow up instructions, it actually remembers all the previous information, and then refined the answer with some follow up questions. It's all really impressive. Of course, GPT for people saw the, you know, their tack report that demonstrate that GPT you know, pass bar exams, and, you know, score in the top 10% For GRE, which is, you know, impressive. How do we use this in a real real world? Use Case? It's another question is what? We're building more customized models? For enterprise meetings?
Yeah. In fact, that brings my next question, right. So what do you think are the most exciting use cases that you might have come across for chatgpt? In general, and maybe in sports? What are the most interesting ones?
No, for us, you know, we are using all this new technologies to build an AI enhanced collaboration system. Now, when people talk to each other, they try to exchange information they try to they try to communicate and
accomplish some tasks. How can AI improve process? And for sports? I think there's many ways you mentioned the research. And, you know, a lot of the sports team have their sales team as well, you know, we actually seen a lot of salespeople use in Otter to improve their, the quality of their sales calls, they track the customer's requirements, they track the next steps. You know, who is the decision maker, you know, all this can be summarized by AI, even sometimes, a human sales rep may miss certain things, AI can be much more reliable detecting those questions.
That makes sense. Yeah. So maybe you can integrate that. The, I guess, the data from the transcript for like a CRM system to right.
Marketers, I'm sure a lot of sports team have a marketing department. We have a lot. We've seen a lot of marketers actually use Otter, because these people have so many cause with you know, that other agencies are involved, and enticement, companies, they, you know, people who need to communicate a lot can benefit from Otter tremendously. Yeah.
And, you know, my last question, I know, you touched on some of the future use cases and capabilities of the Otter.AI platform, when you guys will be using things like chatgpt or generative AI, like you mentioned, right? Having a system being able to recognize people, right? For you know, to document action items or recognize the enterprise data right. So how do you guys intend to play in that space? Generally speaking, I mean, you know, you do use the use cases, but what would be your your vision there?
We see Otter will be a essential part in people's daily workflow. It's already integrated with with calendar can join your Zoom meeting, Google Meet meetings, Microsoft team meetings. It will be integrated with all other products like Salesforce CRM, HubSpot, or project management system like Jira, at at in Asana, or, you know, chat apps, like Slack or Microsoft team chat. Otter will be a essential part of people's daily workflow. So, you know, we're track or capture all conversations, you can always find things
in your own meeting, and after, of course, a lot of meetings in enterprises are open to people in the company. You know, how, you know, when you ask a question, maybe that question or was already answered by
another person in another meeting? So ordered the answer, by the way, rather than having you schedule another meeting with that person.
Interesting. We'd Yeah, it's fun. Yeah, it sounds like you guys have some interesting use cases. But that
that makes sense. Well, look we at the end of the interview, but I want to thank you for for your time today. Very interesting, and good luck for with everything.
Thank you for having me. It's a great pleasure talking with you.
Right, thank you.
Thank you. Thank you.
Thank you for listening. To access past episodes and other research articles and analysis of sports technology. Please visit our website, the upside.us Subscribe to the upside newsletter and receive full access to our sports tech business letter and website. Royalty Free Music is provided by ib audio.com. The upside podcast provides timely insights and interviews with global leaders and sports technology. Until next time, keep looking to the upside