Transcript

Craig: The question for us was how do we. pull out a tiger team of innovators without disrupting everyone else's roadmap and the important heavy stuff that they've been working on. And we called it the A team. there's a PM there's a designer, and six AI engineers and we committed to launching a new AI product every single month for 12 months the first one we did was AI recaps in May and in June we did AI scorecards. Next month we're gonna do AI playbooks. So if you're a sales rep, or you're a recruiter, it'll tell you what to ask. And we just have all these things. It's almost like a startup inside of a thousand-person company.

Ted: Hi, I'm Ted Blosser, CEO and co-founder of WorkRamp, where we're redefining the corporate learning space with the world's first all-in-one learning cloud for employee and customer learning. Welcome to the LEARN Podcast, where we learn from the biggest leaders in SaaS and hear what makes them successful. Hope you enjoy the show.

Ted: Hey everyone. Welcome back to the LEARN with podcast. We have an amazing guest with us today, Craig Walker, the CEO of Dialpad. I've been following Craig for more than a decade now, so I was so excited that our venture capital firm Be Capital can introduce us to Craig for this session. So Craig, why don't you get us kicked off.

Give us a little bit of an elevator pitch on yourself, and also Dialpad, and then we'll jump into the thick of things today.

Craig: Yeah. Hey, Ted, thanks for having me. It's great to be here. elevator pitch on myself. I've been a bit of voiceover IP communications startup, entrepreneur for, over two decades now. and Dialpad is my third startup. we started 13 years ago, came outta Google. Used to run the Google Voice product there, and really what we wanted to do was make this one unified platform of totally modern communications and apply it.

To enterprise and business communications. we started that in 2011, but really took a turn around five years ago when we acquired a real-time AI company called Talk iq. And we built that real-time AI into the core of everything we do. So now we call Dialpad a AI-powered customer intelligence platform, but it basically means your sales team, your support team, your recruiting team, and basically everyone in the company using.

The phone or using messaging is gonna be better at their job because they're gonna get real-time coaching and sentiment and scoring and things like that, based on the AI we provide just as part of the service. AI is a pretty hot topic these days. We've been at it since 2018.

We've done over 4 billion minutes through our platform of real-time ai. So we have really large. Data training set to be able to do some unique stuff for that. we sell to everything from the largest enterprises to the S M B. So anyone who wants to have their business run better, that's a great communication solution for that.

Ted: Well, you could, say that the market finally caught up five years later, but you were on the forefront. Thank you AI for just releasing that little beta for everyone get psyched on and now the market is up to where you want it to be. Craig, so we'll dive into AI for the majority of the time today 'cause you, have a lot of really cool views on AI.

Before we jump into that, I don't know if everyone knows this about you, but you found two other companies before Dialpad, before you, went off, from Google and started Dialpad. What were the biggest learnings from those two companies that you've applied at Dialpad?

Craig: Yeah, so I think, working at another startup or prior startups, I think you get more and more confident in your product vision and what you wanna do. the first company we got acquired by Yahoo in 2005. Then we left and started this company, Grand Central.

We got bought by Google in 2008. And the team that you've had with you, they were effectively the same people throughout. So

My chief product officer and I, this is our fifth company together. We sold to Yahoo. We started Grand Central, we sold to Google, we started Dialpad, and we've been working together for 20-something years.

So I think one of the things is you just know, you have this core group of people that you can start a company with and you know, What they're gonna be excellent at. And you don't have to go through this process of like, do we share the same values and, work visions and do we work hard and do things like that?

It just made it a lot easier. So didn't learn a whole lot of lessons other than got a lot of confidence in. how to run a company. we think we know where the technology's going. Like Dollpad was gonna be all about work from anywhere and free you from being, a slave to a phone on your desk.

And then, the pandemic makes everyone a remote worker. And it kind of validated that vision. And then we did that 2018 acquisition of an AI company. And then, like you said, G P T. Puts out a large language model of, generative AI and everyone's like, oh my God, this is amazing.

It was super great validation for the things we saw earlier. So I think it just gives you more confidence and gives you real good starting team to go after it with.

[00:04:33] Ted: I've seen repeat entrepreneurs do that. What's a critical mass of a team to start a new company? Are you talking two or three people, or was it 10, or 20 people each time?

Craig: I'd say it's the core five to 10. And the good thing is as you grow, probably your first 50 people are one degree of separation from that core five to 10. So you really do end up with this shared value system and really easy way to communicate and, execute in the early days when you're, 50 people or less and there's only one degree of separation.

Because now you know, we're. Almost 1200 people and you're, hiring people all around the world. They're not always, one degree of separation there. A lot of times there's zero degrees of separation, so it's a whole different ball game at a different scale.

Ted: I was telling you in the pre-call I worked at Cisco, I thought of this category as unified communications. Actually, when I was doing my research, reading through materials on your website, doing research on you, you literally are changing the name of the category to the customer intelligence platform.

Tell us about that shift. Give us some background on what that even means, does that mean unified communications is gonna be a thing of the past? I mean, you are probably one of the top, I would say, two or three people in the world in terms of leading the charge on unified communications originally.

So how do you see the category now? What have you evolved it to?

Craig: I always thought Unified Communications was this total misnomer because it was always a silo. we're unified communications, but we only do the phone system. they didn't do meetings and they didn't do a contact center, but it's unified. It never made any sense to me.

We wanted to have one platform that could do everything if you had a really kind like. Google-type microservices architecture that can scale infinitely and you can add, features very easily and everything's little microservices. You can innovate much quicker than, these legacy players that weren't built that way.

And so we called it. True cast instead of UCAS and CCAS, we just said true cast and it's one truly unified thing. that is true and that is still, compelling. And if you wanna have one solution for everyone in your organization, one thing to learn and they can do everything with it, that's awesome.

but. I fell in love with AI in 2018 when I saw the first demo from this company called Talk iq, and we were gonna partner with 'em and we were gonna do this in a p i connection to our platform where when you're on a real-time, phone call, maybe you're making a sales call or talking to a customer, in real time would give you coaching and it would say like, on a sales prospecting call, it would be like, ask these four questions, then would check it off if you actually did it.

And when I saw that in 2018, I was like, this is the next level. we're connecting everyone and it sounds good and it's easy to use and all that type of stuff. you can actually make decisions to provide better service or sell more.

And just by using the platform, that's going to coach you as you do it. And it'll also give you a report card of how you did. It'll also tell your boss and then give analytics so they can make you better. We realize that the platform is allowing us to get tons of customer intelligence.

My strong belief in the future is there's gonna be two types of companies. Ones that learn all their conversations from when they talk to their customers and prospects and recruits and things like that. And the ones that just do it the old-fashioned way and just hope to wing it and keep doing it the old way.

And I think AI gives such an advantage to the ones that are learning. Based on the conversations they're having. And those conversations can be digital now, they can be over social, they could be in meetings, they could be in video calls. It really doesn't matter because it's this true cast platform, but it gives you the ability to go sell against any competitor and be like, look, not only is it a better solution just on the communication side, but it's going to help you run a better business and be more successful and have happier customers and close more deals more often.

And that's a pretty compelling pitch.

Ted: I'm almost visualizing what I call the unified communications world was really the pipes into any type of device, whether it's telepresence, whether it's, zoom, whatever it is. but what I love about the vision you're painting is, about the data and maybe even applications on top of that.

Yes, we will provide. The plumbing, but the real value is what you're gonna do with that plumbing. and even in real-time.

Craig: That's exactly right. It was kind of magic in the early days, like 20 years ago, if you're like, my phone call went over the internet. that was magical, right? But fast forward 20 years, and now it just works. So the really interesting thing is what you get out of it.

Ted: I was still explaining that to my dad two nights ago, and he was like, wait, what? you could call over the web. so some people haven't caught up. Some people haven't caught up. So, let me deep dive into your doubling down on AI. And so I want to bring up a stat.

I was reading Thomas Tongues, the famous venture capitalist who said, About 90% of the companies that he researched said, Hey, they're doing something around ai, even including companies like Work Ramp, but only 30% of 'em have actually put their money where their mouth is, where they're actually hiring new engineers with new, fresh, open headcount that's focused on AI. It seems like Dialpad has gone all in. On this, walk us through how you went all in and when you did that?

Craig: Yeah, we went all in back in 2018 when we acquired Talk iq and they were, backed by Salesforce Ventures. They were backed by scale ventures. they were a serious, legitimate Silicon Valley venture-backed company. And they had, probably about 35 employees with a handful of PhDs, but the thing that was so impressive is they own the entire stack and our, view of how we can be we wanna be like the apple of business communications. We want every single aspect of it to feel natural and to be excellent and to be delightful. And you can only do that if you own the stack.

And if you look at Apple, they own everything from the silicon all the way up to, the hardware and the applications on it. so that's what we wanted to do. So AI, we saw the vision 2018 we're like, we have to own this and build it into the core of what we do. we don't wanna be dependent on a third party's AI, just like we'd never be dependent on like Twilio for voice.

We have to own that in order to do it. And the problem is everyone's just jumping on this open AI bandwagon.

They're like, look, we have AI too. we can turn things into text and ask it to summarize it and it will, but that's only scratching the surface of what you can actually do with AI if you don't have a real AI team, great. You could have a bunch of prompt writers who are gonna be able to write prompts to some third-party thing.

But similarly, we feel like we wanna own that. We wanna be able to control the output. We wanna be able to customize it. We wanna be able to have the speed, right? We wanna make sure it's private and secure so your data doesn't go out to a third party to get analyzed. We can do it internally. The beauty is, we have 11 AI patents, two of them pending, but nine of them issued. We have a team of over 50 people just on the pure AI core engineering side. We published over 12, research papers in, respected journals. and that allows us to do real custom things.

Given we now have this training data set of 4 billion minutes and growing every month. Since we own it, we don't have to pay someone per minute basis to get it. That allows us to turn it on for more and more of our users. That means we get more and more training data. That means it gets more and more accurate and it's this total virtuous cycle.

Instead of having to say, stick around and answer a survey, or we're gonna text you some questions, it can just listen to the call and give you a score. And so now instead of basing your customer satisfaction on, a 5% response rate, you can now analyze 100% of your calls and get a score.

And it's so much better. And so we can tune 'em for different industries. We can tune 'em for different roles. And that's the beauty of owning your own team that can do that type of data or machine learning type stuff. So

We think it's a major differentiator. even though everyone goes out and raises their hand and says, we have AI, if you read any of these releases, Mostly they're running summaries through GPT or Google Vertex, and they're really, limited on how much they can put through. 'cause the APIs have these capacity constraints that you can't just flood 'em with, a silly in minutes.

Like you can only do a certain amount. So they can show that it would be cool for one out of every thousand calls. it's not super scalable or compelling.

Ted: Hey everyone. We're having a great conversation here with Craig Walker, CEO of Dialpad, and he's talking about how to build an AI-first company.

Also want to tell you about WorkRamp and the investments we're making in AI. With WorkRamp's Learning Cloud Assist product, you can now create content with only a few clicks, leveraging our new AI models. If you want to check it out, visit us at www. workramp. com to learn more. Now back to the show.

Let's turn back the hands and time of the 2018 time period. How many employees was Dialpad before the acquisition? Do you remember?

Craig: We were probably about 300, maybe 350 right around then.

Ted: And so you acquire a company that's 35 people. I've seen it happen. I think Five Tran did something similar if I read that correctly, where you go and acquire a company and almost infuse or take over your roadmap of a much larger company.

How did that all go down? I love to see how you parlayed that into team structures today.

Craig: I started my career in the nineties in Palo Alto as an attorney.

And Cisco was our client. And Cisco at the time was, massively doing m and a in the late nineties, and they were the world's best at it, The founders would stick around the products would be great.

And then you see other ones who do terrible jobs of it. and so I always took from that of like, when we're gonna go acquire a team, number one, it has to be something that, Is accretive to us. Like when we bought Talk iq, we had no AI talent in the building, so there was never gonna be this jealousy of which team's gonna own it, who's gonna do it?

We don't have it. You guys do welcome aboard. You're now in charge of it. so number one, it had to be complimentary technology. Number two, just culturally they had to be. Similar work styles built on modern architecture. Talk. IQ their, CEO was an ex-Googler who was like really good friends with my co-founder.

So we had this deep relationship, to begin with. And they were built on Google Cloud platform. We were built on Google Cloud. Just made it easy. and then third, gotta have this shared vision. You put this incredible communications platform together and you, merge it with this real real-time AI.

Platform, you put those two things together it's gonna be one plus one equals three. And so that's what we were really looking for. And then we've done the same thing. We've, done a tuck-in acquisition of a video conferencing company when video obviously became radically important during Covid.

We did an acquisition of an omnichannel company, so now whether you're talking through WhatsApp or Twitter or what have you, you can come through the same interface. And then we did one for a semantic search, AI-powered chatbot too. So all these things, and they all fit that same model of complimentary technology.

Great founders, a great cultural fit, and a great technical fit as well.

Ted: Thanks for that blueprint there. and you're totally right. I remember when I was at Cisco, we probably acquired, and this is in 2006, we acquired like four companies, a month almost. It felt and hopefully you're, making good money as a lawyer back then. But when you look at 2018 and beyond and how you've structured your teams all the way to now to support the AI vision, We were even debating this internally at Work Ramp is, Hey, how do we go do this?

Do we spin off a separate team? Do we go do guilds? Do we kind of embed everybody as a partial AI engineer? Walk us through how you've even structured the teams. I'd love to start maybe on the engineering side or slash product side all the way out to, now what you have even on the go-to-market side.

Craig: On the engineering side. So we have separate teams, we have a team that's responsible for the infrastructure and the platform. Obviously making sure this worldwide communication platform is up 24/7 and optimized and is, excellent enterprise-grade quality everywhere in the world. so that's one team.

Then you have the product team that works on the applications of the mobile apps and the desktop apps, and what are the features and how's that gonna work? and then we acquired the AI company and their, AI team.

So that's kind of how that is structured. But then When ChatGPT came out and the generative AI models came out, we said, look, we have all these advantages in AI. If we could take everything we have and tie it in with a large language model that can do a lot of really interesting stuff and make it seem very natural, we can be super innovative.

But the question for us was how do we. pull out a tiger team of innovators without disrupting everyone else's roadmap and the important heavy stuff that they've been working on. And we called it the A team. There's a PM there's a designer, and Six AI engineers and we committed to launching a new AI product every single month for 12 months for like, the 12 months of AI. the first one we did was AI recaps in May and in June we did AI scorecards. Next month we're gonna do AI playbooks. So if you're a sales rep, or you're a recruiter, it'll tell you what to ask.

And we just have all these things. it's almost like a startup inside of a thousand-person company. it's a small little startup that has the support of the platform and everything else we've invested in over the years, but they're just free to like, go make it happen.

And it's been. Exciting for everyone in the company. It shows what you can do when you have kinda this background and this base, to build upon. this is still a startup, but we're 13 years into it. It's, great to be working with just like the a team that's just cranking on new things.

As soon as they launch one, it's been a ton of fun. But I do think you have to almost pull 'em out their day jobs and create a separate silo so they can be free to go experiment and do stuff.

Ted: And congrats on hitting the, 200 million mark and you're still innovating every day. Anything unique on the Go-To-Market side overlays, or is it mostly embedded on the go-to-market side?

Craig: It's interesting because,take ACIO at a big company, had the luxury of taking his time to decide, can I get a couple more years out of the Cisco on-prem solution?

I know ultimately I have to move to the cloud, but when will it be, and things like that. And so they've dragged their feet. The pandemic accelerated that a lot. So all the people who thought I'm just gonna live OnPrem had to like support remote workers. So that was step one in accelerating them.

But ever since March when ChatGPT came out and shown what the future holds, you better be thinking about it right now, this is going to transform how your business operates. So you don't have the luxury of sitting back and waiting three to five years to decide whether it's a real thing because

your competitors who adopt this are gonna be so far ahead of you. So it's been really interesting to see how larger enterprises have really taken this AI opportunity and embraced it. it's really just starting to happen because again, a lot of people have been showing off kinda like, prototype stuff, but no one's been able to bring it out GA yet.

And so that's gonna be the exciting thing, bringing it all out in ga.

Ted: Let's talk about the future. That's gonna be a million dollar question 'cause you're probably one of the, best thought, leadership experts around it. How do you see the. Market playing out between the big players, OpenAI, Google, Microsoft, I'm sure they're all talking to you as well too, in terms of, as a either, potential user of their larger platforms.

I know you do a lot of it in-house as well, but how do you see the market playing out, let's call it on the next two to three-year time horizon?

Craig: It's interesting 'cause OpenAI grabbed all the headlines, right? And Microsoft had given them big investment. Google has come on so strong lately, and Google obviously has such. Prowess in the AI space they're literally cranking it up and we're working really closely with the Google Vertex team on this as well.

So the way we built our architecture It's this very clean interface. So we can use ChatGPT where makes sense. We can use it. Google Vertex, We can use our own large language model, which makes sense, and we can use other things that may come along in the future. It's very extensible and so how's it gonna play out?

Those are two massive trillion-plus dollar value companies that have great, insights and research and a ton of, data. They're gonna battle it out for supremacy in these, Uber large language models and where we see the, opportunity for us with our own AI team and our own stuff is I don't need to build a large language model that's going to, write you a Shakespearean play or write you a program or, paint you a picture, do a bunch of crazy things that.

These other ones do. I need to have a large language model that's gonna make your business communications better and more excellent and help you be better at your job. And so it's a very small subset of the training data of the entire internet, circa 2021. it's really gonna be interesting to see how those two battle for supremacy who's gonna have the ultimate large language model and then how do we use what we do particularly well and leverage those where we wanna leverage those, but then not be dependent on them when we need scale and need to be able to do the things we wanna do.

Ted: Yeah. It almost seems like you could supplement maybe on the edges. what you said reminds me of what, Chamath said on the All In podcast where, he was saying, Hey, the gold Of the AI market is who has the best proprietary data.

Everything else is going to get commoditized,in terms of the N L P and, the generous stuff, but it's all about, hey, who has the best data? And for you, the business communication data is your wedge long-term in the market. That's such a good strategy.

Craig: Think about it, like the models of what a conversation on a support call. are entirely different from what it would look like on a sales call. Support calls. Someone's already familiar with you. They have some sort of a problem, they're coming in a little frustrated and you're trying to solve their problem, get 'em as quickly as possible and make them happy.

So there's this like, sentiment turnaround that happens on a sales call. Like it's you calling someone who is not prepared, you're interrupting someone in their day and you're trying to convince 'em to spend some time to learn more about your thing. And then recruiting is more similar to sales and different from support. But the nice thing is if 4 billion minutes of training data, that's literally almost seven thousand years worth of training data and you can now start. Breaking it up, okay, these are sales calls, these are support calls.

These are recruiting calls. Now we can start getting into it, and then in the future, you could potentially even go industry-specific. this is a law firm, this's Healthcare. This is, insurance. And you can start even making it better and better.

That's why that private data set is so important to be able to give better, advice and sentiment scoring and next step type suggestions to all these different industries.

Ted: Well, Craig, I'm sold If anyone else listening is not sold, I don't know what you're going by for a, communications platform or customer intelligence platform. So, Craig, thanks for giving us the background there. we're wrapped up on the AI portion here. I wanna actually close with what we call the learn rapid fire round, where we're actually gonna just ask you a few questions.

give us one or two line answers to each. The first one. I'm gonna ask you is what is one podcast book blog that you have learned the most from, or even that you're listening to now?

Craig: As far as like a business book, more than anything, Frank Sluman, I think is the greatest CEO of all time. So CEO of Snowflake, data domain and, ServiceNow. I'm also reading a Team of Rivals, which is all about Abraham Lincoln's cabinet and how he, put all the people he ran against.

For a president in his cabinet there was this tension-filled cabinet, but he got the best out of all these different folks. So I thought that was cool on leadership. But if it's a business thing, it's, Amp It Up by Frank Slootman.

Ted: Great playbook. I've not read that Abraham Lincoln one. I will have to check that out.

Next one. Maybe this is Abraham Lincoln, but if you could learn from one person alive or dead, who would you learn from?

Craig: Abraham Lincoln went through so much hardship and civil war and, turmoil and all this stuff, and handled it with class and dignity and brought all these people together and was just this humble, hardworking, public servant.

the more I learn about 'em, the more impressed I am. So that would be it.

Ted: Awesome. Alright, I'm gonna close you off with one last question. career-wise, what's the one piece of career advice you would give you've had an awesome career and you're still going, what would you give the audience?

Craig: If you had told me, when I was going to law school in the early nineties that I'd end up as a voiceover IP, entrepreneur, I, would've never seen how I thought ever could have worked out, but it came because little opportunities along the way.

Like one of my clients was starting a venture fund and he asked me to join him and I left the law firm to go to this upstart venture fund happened to be at the end of the nineties and he was investing in telecom and internet stuff. Perfect timing. Right? And then that led to another venture fund that led to going in and running one of their portfolio companies that led to getting acquired by Yahoo that led to go starting Grand Central.

That led to getting. Bought by Google and that led to starting this. it's really hard to give career advice 'cause there's no straight path opportunities will present themselves, when they do, don't be afraid to take 'em.

They may not always work out, but I've, Pretty much found if you work hard and you surround yourself with good people, you got a chance to figure things out. And it's always worked for me. So look for opportunities and don't be afraid to take when they come.

Ted: Well, Craig, thanks for closing us up with that great advice. I appreciate you joining us today, sharing your thoughts on ai, a little bit more on Dialpad, and again, appreciate you taking the time to, hang out with

us.

Craig: Thanks so much Ted. That was awesome.

Ted: All right. See you soon.