Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Building & Scaling Enterprise AI Solutions | Bhavin Shah (Moveworks)
454 Plays2 years ago
"The order of your customers can oftentimes make or break your company."
This crucial insight from Bhavin Shah highlights a key lesson for founders: choosing your initial customers strategically is paramount. It's not just about finding anyone interested, but finding the right partners whose problems you can solve effectively now, setting the stage for sustainable growth.
Bhavin Shah is the CEO & Founder of Moveworks, the enterprise AI copilot platform valued at over $2.1 billion. A three-time entrepreneur with 20+ years in tech, Bhavin previously founded Refresh.io (acquired by LinkedIn ) and gained IPO experience at Leapfrog. He holds degrees from Stanford University and UC San Diego.
Key Insights from the Conversation:
- Custom AI is Crucial: Building powerful enterprise AI often requires developing proprietary models rather than relying solely on generic, off-the-shelf solutions.
- Leverage Existing Systems: Integrate with tools employees already use (like Slack, Teams, Okta) for faster adoption and impact.
- B2C vs. B2B Mindset: Transitioning between consumer and enterprise markets demands adapting skills and strategies, but mastering B2B can be incredibly rewarding.
- Listen to Your Users: Let employee needs and pain points guide your product roadmap for enterprise solutions.
- Power of Focus: Saying 'no' strategically is essential for entrepreneurs to maintain focus and achieve excellence in their core mission.
- AI for Productivity: AI can significantly reduce workplace friction by automating routine IT and HR tasks, boosting overall employee experience.
Chapters:
00:00:00 - Intro & Bhavin Shah's Unique Background
09:42:98 - Growing Up Amidst Silicon Valley's Birth
27:51.62 - Leapfrog: Toys, Tech & IPO Experience
38:33.34 - First Startup: Gaming with Gazillion & Marvel
56:37.94 - Refresh.io: The Digital Dossier Acquired by LinkedIn
01:20:23.16 - Genesis of Moveworks: Solving Enterprise Friction with AI
01:47:20.92 - Why Picking Early Customers Can Make or Break You
02:10:50.86 - Building Moveworks to a $2.1B Valuation
02:26:26.30 - Switching Mindsets: From B2C to Enterprise B2B
02:32:43.20 - The Future: Enterprise AI & The Power of Focus
Hashtags:
#Moveworks #BhavinShah #EnterpriseAI #SaaS #FounderThesis #Entrepreneurship #ArtificialIntelligence #ConversationalAI #FutureOfWork #SiliconValley #VentureCapital #StartupJourney #LinkedIn #TechFounder #ITSupport #HRTech #Productivity #AIPlatform #StartupAdvice #StartupFunding #FounderStories #B2B #EnterpriseTech
Recommended
Transcript
Introduction to Bhavan Shah and MoveWorks
00:00:00
Speaker
Hello, everybody. My name is Bhavan Shah, and I'm the CEO and founder of MoveWorks.
00:00:17
Speaker
I'm sure you would have heard of this new tool called chat GPT. It can answer questions like a knowledgeable human being. This is probably the future of how we search for information. But did you know that within the enterprise space, chat for solving routine business issues is already an established product category?
Life and Ventures in Silicon Valley
00:00:33
Speaker
In this episode of the Founder Thesis Podcast, your host Akshay Dutt is talking with Bhawan Shah, the co-founder and CEO of Moveworks, which helps companies become more productive through a chat interface for solving employees' IT-related issues. Bhawan has had an amazing journey growing up in the Bay Area and living and breathing the startup culture of Silicon Valley, surrounded by the giants on whose shoulders we stand today.
00:00:55
Speaker
Bhavan started his first venture in the edtech space way back in 2005 and his second venture was in the social networking space which was eventually acquired by LinkedIn. His current venture, MoveWorks, is worth $2.1 billion and is operating in the massive enterprise productivity space. Stay tuned and follow the Founder Thesis podcast on any audio streaming app to learn about founders from India who are changing the world.
Early Fascination with Computers
00:01:27
Speaker
Both my brother and I were very interested in computers. My dad would bring home PCs. He would bring home IBM PCs. He would bring home clones, ones that other third-party companies would make. So from a very early age, we spent a lot of time playing games on those computers, writing software and code. My first computer was a RadioShack TRS-80 that I programmed on. There was a store.
00:01:52
Speaker
that's now gone out of business back when I was growing up called Fry's Electronics. And they were very popular in the Bay Area and in some parts of California. And Fry's Electronics had all the computer gadgets, like they had circuit boards, they had chips, they had technology, software, computers, everything that you can imagine.
00:02:11
Speaker
Intel 386 was one of the first computers I built from scratch by buying all the parts from Fry's and then I built a 486 and so on and so forth. So there was a neighbor down the street. He lived about six houses down and I remember his house was always a little scary. He wasn't very friendly and our parents would tell us just don't ever stop there. Just keep walking past his house. One day he is in his garage.
00:02:31
Speaker
And he's like, hey, kids, come in here. We're like, oh, maybe we shouldn't. But we did anyways. And he's like, take this chip and go plug it into your Atari. And it turned out this guy worked at Atari, one of the first video game companies right in the world that obviously changed everything. And he was giving us a beta version of these early games. And these were microchips that we had to plug into the main Atari chip board. And then last fun story, I guess I would say is in seventh grade, when I was about 12 years old, my school got a call.
00:02:59
Speaker
saying to pick a student to go have breakfast with the founder of Apple, Steve Wozniak, Steve Jobs co-founder. And I had known about Apple. Growing up, Apple was not too far away. I could ride my bike there in about 25 minutes. And so I had breakfast with Steve Wozniak. And again, that felt very normal because the Bay Area was all about technology. And that's what I thought was
00:03:23
Speaker
much of the world. So these are the types of influences that clearly got me very interested in this space, but it obviously still took much time before founding a company like MoveWorks and getting to this point. Growing up in the 80s in Silicon Valley in San Jose, the Indian community was actually very tight because the community was very small. And so there were a lot of Indian friends of my parents who had started companies.
00:03:48
Speaker
Some very important companies in the history of Silicon Valley, companies like LSI Logic and others, were all being founded by friends of my dad, friends of my mom. So that was definitely the backdrop. I never got pressure to do engineering for my parents. It was natural. I have an older brother who also did computer engineering as well.
00:04:06
Speaker
And then after graduating, so I went to UC San Diego for undergraduate. They have a fantastic engineering program. Just really, it's very practical, very hands-on. You learn a ton about what it really means to build using code. And then, however, what happened there is I started to diverge. I liked
00:04:24
Speaker
the computer science. I like the technology, but I also got involved with the project at UC San Diego with one of the professors there to educate children throughout the United States about Earth science. So what was happening was Dr. Sally Ride. She was the first US astronaut in space. She went into space in the
00:04:44
Speaker
early 80s, very well known, very famous worldwide and very inspiring. And she had a project to educate kids using a camera on board the space shuttle that would then take pictures of the Earth commanded by the kids.
Education and Early Career at LeapFrog
00:04:56
Speaker
So I actually started to code for that. That was they paid her program and I got a grant through by working there to pay for my undergraduate. So I worked closely with her and I just fell in love with this idea of technology and education, not just technology for technology sake, but technology and education.
00:05:12
Speaker
So then what I did is I decided to get a master's degree, but not in pure computer science, but a hybrid kind of new emerging degree in what was called education technology. So I went to Stanford university. There's only a few universities that had a program with this kind of mindset. Stanford had a program. Harvard had a program. University of Michigan had a program, but there weren't too many of them out there. So being from the Bay area, obviously Stanford was my.
00:05:35
Speaker
first choice, went there and it was a combination of computer science classes, education classes, and then I threw in some business classes. So I went to took some of the electives at the business school at Stanford and I made my degree what I wanted it to be. I wanted it to be those things and so graduated from Stanford. So I didn't take any time. I went straight from undergrad to graduate school.
00:05:58
Speaker
And then, obviously, having graduated with a slant towards education technology, I met two founders of a company called LeapFrog, which was an educational toy company. And at the time, they were a small company, about 60, 70 employees, and they were like, wait a minute.
00:06:14
Speaker
You've been studying education technology. We built a company to do this. This is fantastic. Would you like to join us? So I joined LeapFrog in 2000 when I graduated from grad school. And over the next five years, the company grew tremendously. The 1200 employees became the third largest toy company by market cap on the United States after Mattel and Hasbro. And we went public on the New York Stock Exchange in 2002.
00:06:39
Speaker
So I learned a ton there, Akshay, about what it looks like from the inside to see a company go from something very nascent to something that's publicly traded. I learned a lot about product. I learned a lot about engineering, but it was a really fulfilling and fun first job, but also a place where I could soak up and learn just about how everything worked in a company.
00:07:02
Speaker
So LeapFrog was selling like stamp kits and DIY stuff for kids to learn like that was what they were doing. LeapFrog was well known for its several products but its focus as a company generally speaking was to educate kids on how to read phonics.
00:07:19
Speaker
So basic reading. And so they had this product called ElitePad, which was a plastic kind of book that you opened up and had technology inside. And it had a pen. And when you move the pen on any point, it can detect where the pen was. So you put these books down and they had book story books and science books and technology, education books and reading books. And kids could actually touch the word and it would say the word. They could sound out the word. They could learn how to spell. They could go through these little quizzes and little games on the pages.
00:07:49
Speaker
This became a $700 million business for LeapFrog annually. So it was a very large business for the company and became one of the most well-recognized brands in America, in particular for educational toys. The culture at LeapFrog was interesting because LeapFrog was headquartered in a town called Emeryville. And Emeryville is just across the Bay Bridge from San Francisco. There was another company that was also forming at that time
00:08:18
Speaker
and was growing fast in Emeryville as well, which you'll know, which was Pixar. So Pixar animation and all the movies that have come out of Pixar from Toy Story to Finding Nemo to all of the great hits that we've seen that Disney eventually bought was happening just literally down the street. So a lot of creative power, a lot of talented technologists, authors, storytellers, it was just
00:08:46
Speaker
It was an incredible time to be in that industry, in that town, at that place, going back to my previous context. So imagine growing up in the Bay Area and imagine every uncle telling you about their startup and the whole, the whole dot com boom had happened and bust while I was at LeapFrog.
Entrepreneurial Spirit and Gazillion Gaming
00:09:06
Speaker
So companies had been created, all this value had been realized in some cases and then destroyed very quickly. But the entrepreneurial spirit was still very strong in the Bay Area and very much, I couldn't go to a party or an event or somewhere that people would say, hey, you're going to start a company.
00:09:21
Speaker
Now, I never felt for that. I never used that as my guiding principle. I was always of the mindset that if I have a good idea, if that idea is something that I can really put my focus behind and I can believe in it for a long time, then I'll start a company.
00:09:37
Speaker
Now, it turned out that my brother, who's a little bit older, had also started a company about three years prior. So he had been doing it and had been seeing his journey. So at the right point, what happened was a couple of folks and I started to get to know each other. And we decided to start this gaming company called Gazillion.
00:09:56
Speaker
which was initially focused on an educational idea of using video games to help with education. So again, LeapProg, if you remember, was an educational toy company. So physical toys, you bought stuff at the store and you played with it. Here we were looking at like online PC gaming for kids, specifically the massively multiplayer online game space like World of Warcraft and all the stuff coming out of China at that time really inspired us to start the company.
00:10:23
Speaker
So I was becoming an expert at LeapFrog, thinking about this. I'd done the degree in grad school, and now there was an opportunity to start a company focused on helping kids, helping them learn, but in the new modern way of using very graphical multiplayer games. What are the targets you had to save audiences LeapFrog? LeapFrog would be, I'm guessing, like maybe a two or three year old till about a seven year old.
00:10:48
Speaker
Yeah. Yeah. So LeapFrog was about four years old to kind of 10 years old was their range. The game company was a little bit focused. There's some overlap, probably about seven, eight years old, all the way up to 12 or 14 years old. The initial focus that we started.
00:11:04
Speaker
And how did that go? Did you get a breakout hit in terms of a game which became wider? What happened was several years into the company, we found that the journey of building an educational video game, especially a massively multiplayer online one, was very difficult. It's a very hard process to get right. And to your point just now, which is, did we have a hit? It is a hit-driven business. Gaming can be very unpredictable until you find a franchise that then you can build year after year. So we didn't find that hit.
00:11:34
Speaker
And we didn't get there with our idea. So we ended up partnering with Marvel, the superhero comics company, before they were acquired by Disney. So this is when they were selling independent company. And they worked with us to produce a series of Marvel games. The first one was actually a kids game focused on the Marvel hero superheroes called Marvel Superhero Squad. And it was a cute, diminutive version of all the superheroes that fought and bash. Nothing to do with education.
00:12:00
Speaker
Nothing. It was just straight up gaming. And then we did other games, like we partnered with LEGO and we actually did some M&A and we did a published a game called LEGO Universe as part of that. And that was another online game. And then we did another game called Marvel Universe, which had more of an adult focus. So that's where the company ended up focusing on rather than education. And how did you do Friend?
00:12:23
Speaker
Did you raise funds or was it based on your savings? No, it was based on venture funding. So we went out in your typical Silicon Valley and venture investors. My co-founders had experience raising capital. This was my first go. So I was learning. I wasn't the CEO. I was an operational role COO and managing internal aspects of the business, but in the room and all this stuff was happening. So that gave me a lot of perspective on when I was going to do it as a CEO, what I needed to understand.
Founding Refresh and Acquisition by LinkedIn
00:12:53
Speaker
And it also allowed me to build relationships and get to know the network that exists out there, which I ended up, after about five years of doing Gazillion, I departed to go and just try my hand at something else. Because again, remember, if you think about the origin of the company, it was about education. We were no longer doing that. So my necessarily passion for it wasn't the same as
00:13:15
Speaker
When I first started and there were a lot of people at that point, we had about 250 employees that were very focused on gaming and core gaming. And that was wonderful, but not my, not where I wanted to spend a lot of my time. How was gazillion distributing gifts? Was it like selling online or online? So these were all.
00:13:33
Speaker
what they call massively multiplayer online games so you go you create an account you're playing at the same time as other people are playing you can see their avatar you're in these common rooms you go on these adventures etc so very similar to that so this is now two thousand and twelve that i'm thinking about what i want to do and i had
00:13:50
Speaker
really started to having had experience with consumer at LeapFrog, having had experience in the gaming world, working with consumers. I definitely was familiar with that world to an extent. And I had been fortunate earlier in my career when I was at LeapFrog to go on a very kind of unique trip to Afghanistan. And when I was there, it's a longer story that involved LeapFrog developing some products that
00:14:20
Speaker
were being used by Afghani women to help educate them about prenatal health. So we did some educational stuff for them. And I went out there to go present some of this. And in meeting the president of Afghanistan and these other dignitaries, I noticed that they all carried the briefing dock.
00:14:36
Speaker
And I read these briefing docs and I was like, these are fantastic. They gave you all the notes, exactly what you needed to say. You last met so-and-so in April of 2010. You spoke about these three topics with her. Wow. Life is quite nice when you have all this work prepared. And I saw this. I was with Secretary Tommy Thompson from the United States and several other folks. And so I got this idea of like, wow, wouldn't that be cool if everyone had a tool like that?
00:15:02
Speaker
So fast forward to 2012. So that was back when I was at LeapFrog around 2004 timeframe. 2012, I was like, wonder if we could do that digitally. Wonder if we could build something using Facebook data, Twitter data, LinkedIn data, open graph data types of kind of information. So that before you meet someone like, Hey, actually, did you know that you and I went to the same college or did, Oh, I didn't know that based on your Twitter post that you were in Thailand last month and Hey, I've been to Thailand. We can talk about that. So a small talk.
00:15:31
Speaker
type of an idea. So we ended up developing a tool called Refresh that was inspired by that. And you can put that in the mobile productivity category. It's mobile using mobile phones. It's leveraging the internet to these databases like Facebook and so forth and pulling it all in. So that was the inspiration and then eventual journey of building a product that essentially provided a digital dossier is how to think about it.
00:16:00
Speaker
Was it a premium monetization model or was it pure subscription to use? Once we got it launched, we really just focused on distribution and was fortunate that a lot of key people starting in Silicon Valley were early users. It was a free product. Anyone could use it. Anyone could get access to it.
00:16:19
Speaker
We started to see a lot of professionals use it in a setting like sales. And so we started to build something that we could sell to a larger enterprise. And that's what we were going to monetize. However, it was at that point that we got a lot of inbound inquiries to acquire the company. And we talked amongst ourselves and realized that this was a fantastic idea, but it was still generally a small idea.
00:16:44
Speaker
It wasn't going to be a massive thing because it required cooperation with all these different third-party services. We weren't in full control. You were at Denmark, so they could revoke access at any point of time.
00:16:55
Speaker
Yeah, and we were seeing more. It's less so that was a concern. It was more just that the world was shifting too. People wanted more control of their data. They wanted their right to be forgotten. And all of these things started to emerge. So this idea of people data got quite messy. And we also knew that our product would be a great fit inside of one of these larger applications and bringing it together. So the team at LinkedIn, who literally was down the street, we were off a street in Mountain View called Castro.
00:17:24
Speaker
And if you go down Castro, you take one left turn and a right turn, you're on shoreline. And shoreline was where LinkedIn was located. So a very close neighbor, but eventually we decided to sell to them. And this was when Reid Offman was still learning things. After Reid, Jeff Weiner was the CEO at that time. Yeah, okay. Okay.
00:17:43
Speaker
And did this eventually become the sales navigator product of LinkedIn? No. A lot of it has been pared down. You know when you open up a LinkedIn profile and it says, hey, here's something in common that you have with this person? Connections with common or you both studied it? Yes. So it's that latter part. The connection common they had, but it's more of that. You can't do as much if you're just doing LinkedIn data. We were obviously pulling in lots of other data sources, which would be like, hey, did you know you actually went to the same restaurant?
00:18:09
Speaker
You could find that out from a Yelp API, but LinkedIn doesn't have that data. So less interesting, but more professionally focused around universities, maybe around certifications, maybe around types of commonality. Pre LinkedIn acquisition. What was the way in which a user would use it? You would like type in a name and it would show you various people whose names match that.
00:18:32
Speaker
No, so it was tied into your calendar and what it would do. A lot of the smarts of that system was to figure out when in my calendar said, hey, meeting actually that which one is that? Who is that? And sometimes it would be an email address and maybe that would be good. If I invited you, we could start there and try and find who you are in these social networks.
00:18:51
Speaker
Not that, but a lot of times all we had was a name. So then we had to think about your graph and your secondary through tertiary. And that was a lot of the smarts is that we were more accurate than many of these other tools that were out there, which were trying to aggregate identities of people.
00:19:07
Speaker
Sales tools did this. Salesforce would try and bring in social data around an account, an opportunity that you've loaded up. We would do that, but we did it quite accurately. 15 minutes before your calendar meeting, refresh would pop up saying, hey, I have some insights about Akshay. Did you know there's, you guys have three friends in common on Facebook?
00:19:25
Speaker
Oh, I didn't bother to check that. Who are those folks? Oh, they're cousins of mine. That's a small world. That kind of thing.
The Birth of MoveWorks and AI in IT Support
00:19:32
Speaker
There would obviously be like a one time setup needed from the user where he would sign into all his social accounts so that you can map his graph and all of that. Exactly right. Okay. Okay. Quite a full app to be giving out for free. This is even today, there's nothing like this available for free. These are all like really premium services.
00:19:49
Speaker
Yeah, it turns out it's a pretty hard problem and now even more so since it's not the same as it was. So did you stick around at LinkedIn? Or did you? The engineering team for the most part did to help create some of the technology, but everyone kind of went on their different paths. I in particular wanted to use that opportunity to reset and
00:20:12
Speaker
think about where I wanted to take my life from that point. Refresh was based on some experience I had before. MoveWorks was actually a little different. It was based on us just doing a lot of conversations and research with IT leaders, with CIOs. There's four of us that founded the company. Vaibhav, myself, Varun, who was at
00:20:34
Speaker
Bhai Bhav was another serial entrepreneur, and Verun, Facebook, he'd also been part of early stage companies before that. I think who was an expert at machine learning and had spent many years at Yahoo and in many years at Google doing some very core machine learning work that ended up inspiring what we do today. Imagine us each looking at this problem from a different lens. I'm looking at it from a business or a product standpoint or market standpoint. Verun
00:21:00
Speaker
product genius looking at this. He'd been doing chatbots at Facebook and building them for internal use cases. So he understood what worked, what didn't work, what was vaporware, what was real, what was possible, and the art that exists in making something effective.
00:21:18
Speaker
And then John was doing a lot of machine learning work in the NOP space. When you ask that question to Google, why is this guy blue? Or how do helicopters fly? Or what food do I give this kind of a plant? You know, there's a paragraph, the answer that tells you what
00:21:34
Speaker
know about it. He was the one that created that with the team at Google. It's called WebAnt, and they had to do NLP on the fly. They had to do a lot of deep key learning to get that right. Combine that with Varun's background in chatbots, Vibod's background in building large data engineering systems, and in my background,
00:21:53
Speaker
After about 35 conversations with CIOs, we kind of looked at each other and we said, I think there's something really big here. I think we can go and start a company that uses an LP to answer and solve your very common and very routine IT issues. It wasn't like some moment, like you would imagine, where all of a sudden we said, that's it. It was more that as we started to get into it, talking to customers,
00:22:18
Speaker
We also saw some big killer shifts. So Slack has come out and Slack was everywhere. People were talking about how Slack was going to change the enterprise and we're using Slack more than they were email to talk to colleagues. You still use email external or internal, it was Slack. And we said, you know what? This looks like the future. Then how are you going to get support? You're going to file a ticket?
00:22:42
Speaker
You're going to go through just a standard portal and make a request? No, it should just be right there in Slack. It should be conversational. And we started to see that, plus we got validation from CIOs as to how important this problem was and the value of solving it. We then convinced ourselves to found the company.
00:22:59
Speaker
How big is this market? You must have done some survey of what is the total addressable market of the help desks. You were essentially disrupting service businesses. I'm sure Infosys, with all these Indian companies would be providing an outsourced IT help desk service to companies in the US. So those were the kind of companies that you were looking to disrupt and replace.
00:23:21
Speaker
It's interesting, you know, we never really thought about displacing those companies. In fact, what we ended up doing was really looking at the customer need and the customers were showing us through their data that on average, it would take three days to resolve a typical IT issue. And we said, this is crazy. We're talking about this, we're in this new world of instant with Uber and Ordash and everything's starting to come up. Why does it take three days to get help? And we'd recognize that.
00:23:48
Speaker
Rune was at a large company, Facebook. Jung was at Google. I had a quick experience at LinkedIn and it was all very much the case. No matter how innovative your company was, it still took a long time. So what we wanted to do was solve that problem. Now, there is still a need for people. There's still issues that you and I will have as a professional that someone in IT needs to come over and help with or needs to go reconfigure a server to give us the right to privilege access.
00:24:14
Speaker
But it turns out there's this very common set of issues that come up every day that those things were taking days to solve. I want to get access to a new application. I'd like to use Tableau. It takes three days, four days sometimes to get access. How do we accelerate that? So we were directly working actually with customers.
00:24:31
Speaker
had their own help desk initially. They weren't outsourced with pros, TCS, etc. Not every company in the world uses those organizations. I think it's like maybe 50% of customer companies at that time were using those and other 50% had its
00:24:46
Speaker
house. So we were focused on evolving their needs, how to get something to scale, how to get something to work, and especially in the world of conversational AI. Because if you think about the world of chatbots, many of these things were over promised. I like to say companies deployed them and employees ignored them. They just didn't work. You would say something and it would... You probably had that experience yourself with a consumer product like Siri or Alexa. If you don't get the words right, just start over.
00:25:13
Speaker
Right? But with MOOC looks, we're like, no, we can't have that be the interface. It's got the requirement. It has to be aware of what people mean when they say, I just ran over my laptop. What do I do? Am I in trouble? Okay. The bot comes back and says, Hey, I need you to fill out this loader laptop form so we can get you a new machine right away.
00:25:32
Speaker
Oh, sounds good. And they fill it out and it gets done. These are small things, but they're very powerful because the companies, many companies in the world have built these automations. They built a form to get a new app. They built a widget so that you can go and get access to something. But how often do you need to use those? Not very often. So it becomes a discovery problem.
00:25:51
Speaker
oh, wait, I'm supposed to send an email for this? Oh, I didn't know that. Oh, you want me to fill out this form for that? Well, I didn't know that either. And so what Moverups is doing is bringing all of this together in a very simple interface and allow us to deliver this kind of modern experience on these new platforms. And by then, a few years into this, Teams came out, Microsoft Teams, and they became a very big player in providing chat support within the enterprise, modernizing a lot of companies.
00:26:20
Speaker
So a lot of this came together, and as we started to see it play out, we realized, okay, this was going to be a very interesting company to build.
Innovations and Trends in Enterprise Support
00:26:28
Speaker
One of the things we, what is such a, was that independent of the company, the questions that people were asking, the help that they needed was very similar. And it was our belief that employees should get help when they ask for it. There shouldn't be a delay. But to do this, and the reason to build an AI company,
00:26:46
Speaker
is not to build something that requires a lot of professional services and a lot of building from scratch. In fact, we'd seen many companies try this. They use a generic machine learning model or a set of sort of language models, and they would go to each company, try and build a chatbot that would understand our use cases.
00:27:03
Speaker
That doesn't work. In fact, it wasn't working for anyone. Customers were telling us we deployed these things and no one uses it. It doesn't work. So what we found was that there was a homogeneity, a very common set of issues that we could train and retrain and learn from. And so that gave us strength and the kind of wherewithal to be able to come in and have impact in the very first minute, the very first hour that a customer turned our product on.
00:27:30
Speaker
So it was interesting. We looked at some of the stuff that was off the shelf. Microsoft Lewis or Facebook with that AI library and systems out there. And they were good for building something to get started. But if you wanted to build something at scale that works for all these large enterprises.
00:27:45
Speaker
Boy, you didn't really understand this at the basic level in systems yourself. So all our adult systems are built internally. They've been built up over time. There's not just one model. There's hundreds of models. There's different ensembling of models. There's techniques that we figured out that when someone says, hey, can you add Akshay and Bhavan to the finance distro? That's a complicated sentence. You had to figure out who Akshay is, who Bhavan is.
00:28:09
Speaker
finance, which distribution list, distro, alias, what have you. So there's a lot of technologies we had to build that in 300 milliseconds or whatever the turnaround is, we understand you mean this Akshay, you mean this Bhavan. Here's five different finance distribution lists we could find inside of Office 365. Which one would you like me to add them to?
00:28:30
Speaker
Oh, and now that you want me to add it to, I got to get the approval from the list owner before I can add you. All of these things required us to really understand the enterprise, integrate with these key systems. There's several dozen key systems that most enterprises use for this kind of stuff that got us to where we could then come in and immediately have impact. And that collective learning that
00:28:52
Speaker
we're doing every minute, every day, is benefiting every new customer, every existing customer. There isn't this situation where we have to trade it for your company and then trade it for this other company. We do a little bit of that, so don't get me wrong. Your company may have something called Jupyter, or you may call it Tableau Atlas, so you have your different acronym. We can learn that. Those are customer-specific. They're spell correctors that are customer-specific.
00:29:20
Speaker
misspellings and things of that nature. But a lot of it can be very much operationalized and to end on our backend. So we're monitoring this. There's millions of interactions happening now on our platform. And that allows us to really understand how models are performing, how they're drifting. We have labeled data, some of the largest collection of IT label ticket data that we use for machine training. And that continues to be something that gets better as we write operations.
00:29:49
Speaker
Tell me the journey of building this up. You recently worked with some beta customers where you would have done this free, of course, to learn. How did you build it to a stage where you can show results from the first minute? So we never gave our product away for free. Not once. Not to a customer. We always charge for it. You want to make sure that there's attention being paid by both sides. If I don't pay you anything,
00:30:12
Speaker
Then I also am not going to necessarily pay much attention when you have a question or a concern, or you need my help, or you need my assistance. So we understand that we find customers that had the right, that had the need for this product, that had the vision for this product.
Importance of Early Customers and Global Expansion
00:30:28
Speaker
It was willing to pay a little bit, not as much as you would pay for a fully focused product, but something that gets started. So we always had paying customers, but you asked a really good question, which many people miss in the journey of building a great company. And that is you have to pick your customers well. Just because someone is interested, it doesn't also make them a good potential customer. You have to be someone with the right kinds of problems that you can solve today, not six years from now. And it's about...
00:30:53
Speaker
pressing those customers. So the order of your customers can oftentimes make or break your company. If you take on too big of a customer, you're going to sit there and work on all these really strange kind of arcade problems that maybe that big customer has, and you won't be building something for the average customer who doesn't have those kinds of issues.
00:31:13
Speaker
If you build something for a customer, you're going to be dealing with immature systems that don't show you the real complexity that you need to get familiar with when you go to large customers. So we were very fortunate that when we started, we started to pick the right customers and they started to find us.
00:31:31
Speaker
Wanna be able to visualize the product journey a little better? What was that primitive version of the product that first started? What could it do? And with time, what all functionality did you add in it? And what can it do today? Yeah, the product journey was actually quite interesting because if you think about what happens every day, is always submit tickets.
00:31:51
Speaker
to their IT team. And those IT teams work on those tickets, they get you the access, troubleshoot your issue, they give you an article to read, hey, go read this article, it'll tell you how to set this up. And so we could actually look at those tickets coming in and decide which ones start solving using AI.
00:32:07
Speaker
Which ones could we actually get good understanding of what to do and then start to get a base of understanding of what people say when they need help for that particular issue? So it was a very incremental journey in so much as we started to look at this and we'll say, hey, we see a lot of interest and need. And we also looked at what would delight the user first and foremost to go ahead and understand provisioning requests.
00:32:31
Speaker
Like people asking questions around, I need access to this application. I need access to this folder. And so we started there. And then we started to answer basic questions like, what's the status of my last issue in this ticket? And when I requested last week, all of that started to emerge. And so if you think about this journey, it was less of a.
00:32:52
Speaker
deciding what to pick and then hoping that it would work was more like we could see the data. Hey, 20% of these questions every day are of this kind. Let's start addressing that. Then we would make that available inside the chatbot and we would drive users to say, hey, next time you have to ask for an application, go to the bot because we're really robust when it comes to you asking us.
00:33:13
Speaker
about different applications. We know all the applications available in your enterprise. We know all the ways you can ask for something like Salesforce and Access, and we got more and more sophisticated. So that would then lead to the next capability and the next area of support. And that sort of slowly built up. So sure, today we can solve a lot of issues right away, whereas five years ago we couldn't. But there was still always value, if that makes sense.
00:33:40
Speaker
Even if we were only doing a few percent of their overall ticket volume, that was still a start, okay? And it got better and more and we can in many environments do over half of their ticket volume completely autonomously by just by reading the texts that we understand now very well. So the bot would like, for example, log into the Salesforce master user account and then create a new account if someone or change permissions or whatever the request was like.
00:34:06
Speaker
That would be the hard way to do it. The simpler way to do it is to actually use the governance and controls that IT has already built. So if you think about another product like Okta, which people use for single sign-on, but also for different access to different applications, they've really built some of those logic into Okta to provision a new account. We try and minimize
00:34:30
Speaker
the change that customers have to do to support, to leverage our product. If you've already got all that built out in Octo, let's talk to Octo. Let's ask, hey, what roles do you have available at Salesforce? Okay, these two roles, great. We'll present that back to the user, we'll confirm which one, and then get us that one. And then Octo will do the work to do that, to give you access to Lucidchart.
00:34:50
Speaker
or smart sheets or documents or whatever, we would rather, instead of talking to each application separately, in this day and age, you can just talk to these central hubs. In some cases, if you don't have Optum, you might be using a sale point or you might be using Microsoft AD natively because Microsoft has a lot of these tools as well.
00:35:09
Speaker
Again, hope can self-serve. In theory, actually, they could go and... But what happens is you don't ask for applications every day. You forget that's how you do it. So using our system, just whatever comes to your mind, type in, see what happens. And if the bot takes care of it, you're like, great, this is done.
00:35:27
Speaker
No, okay, okay. So in terms of learning about, you gave the example that I ran over my laptop and so the solution is full of this form. So that is something which would happen on the fly. Mobile would see that this is how these tickets are responded to by a human being and then it would start to mimic the human response.
00:35:44
Speaker
It's important to understand this. We never pretend to be a human, ever. It's important to not ever be something you're, be honest what you are, you are, you know, in your product and in like, I guess they would say. So what we do is we are looking for the best way to resolve your issue. Okay. When you type in a Google long sentence, what's the battery range for this model Tesla model year, et cetera. Sure.
00:36:09
Speaker
want the answer, don't you? You don't want to have a long conversation about it. You just want to know what the answer is. Google has just tell us what you need and we'll come back to the best answer. Very similarly, our product works on that premise is that you say, what's your problem? Hey, I just got a new docking station for my Toshiba laptop, but the adapter is not working with the docking station. It works with my Macbook. What do I do? It's the Mootworks intelligence says, oh, sounds like you need to order a new accessory. Let me help.
00:36:39
Speaker
We don't have to have a conversation around your Toshiba laptop. In fact, you prefer us not to. You just want to know, hey, what do I do next? Or, hey, I just got a new iPhone and I don't have the two-factor setup, so I can't log into the work system. So as a result, our system can take that information, go find you the right answer, which might be, here's the five steps to how to set up two-factor again on your new device. But then they're like, oh, that's great. That's easy. That's simple.
00:37:07
Speaker
Amazing, okay. Tell me about some numbers with timeline, like what kind of top line have you been doing? What kind of fundraising have you done until now for more books? Yeah, so we've raised about $315 million to date from some of the biggest of these folks like Bain Capital, Kleiner Perkins, Iconic, Sapphire Ventures, Tiger Global, Alkeon. It's a great set of investors, and we did that last year, which valued the company at about $2.1 billion a year.
00:37:35
Speaker
And today we're solving about 40% of employee requests autonomously for some of the biggest brands in the world. Companies like Broadcom, Hearst, Call to Networks, Autodesk, AppDynamics, AppDynamics, and more. We have a lot of
00:37:51
Speaker
different partnerships as well. We have a global partner now with TCS and we've continued to really build out this kind of vision. We announced our newest kind of capability with HR so now we can help you answer your questions and people have payroll questions.
00:38:09
Speaker
What's my commission for this month and do I we can remind you got security training that you have to finish by tomorrow night all of these things are real friction points inside the enterprise that then folks and I think that much like we are.
00:38:24
Speaker
trying to help companies become a great place to work. We also want to be that as well. We were named the number one best place to work in the San Francisco Bay Area by built-in. Recently, we did the 2022 one that came out for the Bay Area. We were number two for that. We were in the top companies, places to work in America.
00:38:43
Speaker
A lot of different things that we've been doing. We've been fortunate in terms of the company that we have, in terms of team members. About 500 D plus employees based in different geos. We have obviously the Bay Area. We have Austin, Toronto. We have a fraction of folks in New York that's growing in Southern California. We have our Bangalore office.
00:39:03
Speaker
Really excited to continue bringing more folks onto this platform and help us build out this vision. I'll address a little market action, every enterprise in the world, because every enterprise in the world has an IT help desk, has an HR team that answers questions, right? That has the employees that need support. And when you think about that, we're focused on companies with a thousand or more employees.
00:39:26
Speaker
And our product is something that's typically in the six figures to seven figures in terms of US dollars because it does the work for IT in a way that allows them and their teams to focus on bigger projects, migrations, change management. These are the big topics today where there's so much happening now. There's so much that needs to get advertised.
00:39:48
Speaker
quickly that we're able to do a lot of this mundane work end to end for them. So that drives a lot of meaning. We price it based on the kind of employee base that they have. How many employees are they supporting and how can we be a part of that mix? If you've got a thousand employees, see, you can pay something very different than if you have a hundred thousand employees. That's how we structure the business.
00:40:12
Speaker
For an employee, what is the way in which they access the chatbot? Do they download an app or is it on the internet? Yeah, so it's built into the tool that they use every day, all day long. That's Microsoft T, that's Slack, their Cisco Webex chat, et cetera, that people are already collaborating in.
00:40:32
Speaker
So if I'm going to, I just send a message just a moment ago to my assistant. What did I, I did it on Slack. That's what MoveWorks uses internal. But right there in one of the starred channels is the MoveWorks bot. If I need IT support, help, I just go there. Same thing occurs in Teams. It's on the left hand panel. We want to be there because that's where everyone is spending their time. That's where their day is focused and also gives us presence on their mobile device.
00:40:55
Speaker
on their iPad, on their tablets, on their laptops, so we can be everywhere for them. We also can be accessed through the web. So if you have an intranet, you have a portal, you have a SharePoint site, you have a typical place where employees are looking for help, we're right there. You pop up, that's the same AI as you have inside of Teams, inside of Slack. The idea is that employees shouldn't have to go find us. We should be where they are. And number two, we can also practically reach out to employees.
00:41:23
Speaker
Hey, by the way, here's an update on your issue. I can help you with this. I noticed your account just got locked out. Would you like me to unlock? So there's a lot of proactive engagement as well as we continue to build these use cases over time. No, good. You've been mostly in like a B2C kind of set up pre-move works.
00:41:41
Speaker
How did you navigate that change? Building an enterprise-focused business needs slightly different skills, and the conversion would probably take months to happen. So tell me what your learnings from that switch from B2C to B2B. Sure, it takes time, and it didn't get comfortable and familiar for a while. But it was also an area that I found a lot of energy for. Sometimes in life, you do things, and you're pretty good at them.
00:42:11
Speaker
But you don't realize that you're actually much better at something else. And once you earn that something else, you fall in love with it. Maybe actually, I should have been doing this since I was 22 years old, but unfortunately, I discovered a little bit later. Having the passion for it, we had a store advisory board yesterday with 12 of our customers, Fortune 500, Fortune 2000, Fortune 100 companies.
00:42:34
Speaker
I just love this space. And our team is so good at what they do that our customers are so engaged with our product. So it is different. It is different. And selling to enterprises is different. And the emotion is different. And you have to learn a lot about this. But isn't that why you start a company? You get good at something new and to become a world expert at it. It's quite a rewarding journey. And that brings us to the end of this conversation.
00:42:59
Speaker
I want to ask you for a favor now. Did you like listening to this show? I'd love to hear your feedback about it. Do you have your own startup ideas? I'd love to hear them. Do you have questions for any of the guests that you heard about in this show? I'd love to get your questions and pass them on to the guests. Write to me at ad at the podium dot in. That's ad at t h e p o d i u m dot in.