TalkingHeadz Bonus with Evan Macmillan at Gridspace

Dave and Evan meet with Evan MacMillan, CEO and Co-Founder of Gridspace.

We had to pull Evan out of the garage to record this podcast. Gridspace is a collaboration between SRI Speech Labs and some fellow Stanford engineers. It is pioneering real-time speech infrastructure for contact centers and developers. 

Gridspace was formed in 2012 as a collaboration between SRI Speech Lab, the lab behind Siri and Nuance, and a team of Stanford designers and engineers. Today, the company's software scans and automates calls for healthcare and financial services customers. The company is one of the largest Kubernetes cluster operators for streaming speech and has analyzed billions of call minutes with its cloud and telephony agnostic speech technology.

Transcript

Dave Michels: 0:12

Welcome to talking heads in what is our first bonus episode episode? So excited? Are you excited about a bonus episode?

Evan Kirstel: 0:18

I am. But I don't really understand what a bonus episode is exactly. So please clarify.

Dave Michels: 0:23

Oh, that's a great question because we just kind of made it up. But but a bonus episode is, as you know, we always do, and we're pretty good at this. I think we're right around Episode 80 or 90 is like that. But we always do two episodes, two interviews a month with the movers and shakers of enterprise communications. And we have good guests at the same time. But we always do two interviews a month, and this month, we're gonna have three, this is a bonus episode, we didn't have a plan that wasn't on the roster. It's a bonus episode, because we have an exciting story to tell about a company called grid space.

Evan Kirstel: 0:58

Well, before we get to grid space, let's chat about the top trending stories this week. Have you been following the apple news?

Dave Michels: 1:06

A lot of interesting stuff there. I would love to talk about it. This seems to be all private?

Evan Kirstel: 1:11

Well, it is about privacy. And Apple has basically now declared privacy as a human right.

Dave Michels: 1:17

You know, they're not the first to do that. But I think they're the first to actually mean it. I really like what they're doing. I think one of the more interesting ones, you know, we had the whole Facebook stuff and tracking. And that was before, but they're really expanding on that. And I think the most interesting thing that they're doing is this thing called private relay, which people in our space will recognize as a VPN, but should be a really great way to use a VPN, because money VPN is really are just harvesting information anyway,

Evan Kirstel: 1:42

yeah, it'll be baked in iOS 15 will make websites much more difficult to track you it'll, you know, block a lot of spam, and really provide additional layer of security and privacy. So I'm excited for that.

Dave Michels: 1:56

The other big news, and this one actually affects me because I'm not an apple user, generally. But the other big news that I'm excited about was that FaceTime is going to be supported on Windows and Android. Did you see that?

Evan Kirstel: 2:06

I did. And I think Apple look at what zoom is doing and others and said, We want a piece of that.

Dave Michels: 2:11

Give me some of that. It really positions Apple nicely because they've got already their Apple Pay. And so from a b2c perspective, you'll be able to buy a service and communicate all from one company. I think it's a really interesting play. And they said it'll work on Android and Windows and which was kind of disappointing because I use a Chromebook. But then they explained it's going to be through a website, so I think it's gonna work great on the Chromebook too.

Evan Kirstel: 2:37

Well, finally, now you can get an iPhone. But before that, let's get on to our amazing guest.

god: 2:44

TalkingHeadz is a semi monthly podcast with interviews with the top movers and shakers and enterprise communications and collaboration. Your hosts are Dave Michaels and Evan Kirkstall, both of which offer extraordinary services including research, analysis and social media marketing. You can find them on Twitter, LinkedIn, or at talking points calm that's points with the Z and Devin curse co.com. That's kr STL.

Dave Michels: 3:12

Today's Talking Heads is the first Talking Heads episode was to Evans,

Evan Kirstel: 3:17

did you know that Evan is stone in Hebrew?

Dave Michels: 3:21

So we're doing one podcast with two stones?

Evan Kirstel: 3:24

Exactly something like that. But seriously, our guest today is Evan mcmillon, the CEO and co founder of grid space

Dave Michels: 3:32

and before that he was the founder of Japanese Apple he dude, payments tech company. Welcome, Evan.

Evan MacMillian: 3:40

Thanks. Great to be here.

Dave Michels: 3:42

So Evan, as I understand it, grid space teaches machines to better understand human conversations. So the obvious question we have to start off with is why? Why would we want to teach machines understand human conversation.

Evan MacMillian: 3:57

Demand for conversation often just completely exceeds the supply of available conversationalists. And I think we all saw this in the last year with surges and calls to healthcare companies and brokerages on the most volatile days of the pandemic and most volatile days in the stock market, it was just out of control. And you know, with this surge in calls, we also saw a lot of agents out of the contact center. So you had high call volume, high age and absenteeism. And that was just a recipe for disaster.

Dave Michels: 4:33

So a question that I frequently get is Why call it all because the reasons you just described, people just keep on asking me, why don't we just use mobile apps and chat?

Evan MacMillian: 4:42

Yeah, I mean, you know, virtual agents are a really interesting solution to this capacity problem. I mean, they require no onboarding, they can carry on now really productive 10 minute structured conversations. For the customers that are calling these contact centers. There's no whole time there. quick answer to a complicated question. And it's useful in normal times to not just crises. I think this question of calls versus websites is really great as well. And I think that for considered purchases and these really emotional situations, people really crave conversation and voice conversation in particular, not just to chat. And when voice is really good, it's really good. It's amazing. I mean, I remember these conversations, usually after frustrating web chats or website experiences where that phone call just fixes everything. But for companies on the other side of those calls, it's a hard experience to deliver. And it's really hard to do it inexpensively and consistently. So I think that's kind of the challenge of 2021 is people are still craving voice. It's just really hard to do it well.

Evan Kirstel: 5:53

So conversational AI is all the rage. But when did you become interested in conversational AI? So what started the conversation for you?

Evan MacMillian: 6:01

Yeah, so I've always been interested in phone communications from a tech perspective. In middle school, I set up several websites that would communicate visitors and chats to me on my very first cell phone and then wait, that's Middle School. You said, Yeah, I had a an analog startac. And I got it working with my first website. So it was pretty, pretty cool. I was one of the only people with a cell phone at school.

Dave Michels: 6:27

I had a star tag too. But I was in the middle school. And I got

Evan Kirstel: 6:30

Dave is more of a hot guy. So fast forward from middle school. What happened next?

Evan MacMillian: 6:34

Well in college and right after college, I was really interested in programmable polyphony in at first with some of the open source projects that made it possible to make calls with a server. And then later Twilio, which was, of course built on Asterix and then free switch. So I got into that I thought it was magical to be able to call or text from computer. And then after my last startup that Groupon bought, I really got into Twilio and kind of understanding sort of the problem of understanding customers from the perspective of a business, you know, how do you actually understand all these calls that are coming in? How do you automate some of those calls? And that's when I really got interested in this big hairy problem of making sense of conversational speech specifically spoken conversational speech.

Evan Kirstel: 7:30

Wow, that's exciting. Well, Jeff, from Twilio listens to this podcast, I'm sure it'll be he's been delighted to hear that. And absolutely. And so fast forward to starting a company around conversational AI. How did that get started? And kind of where are you exactly today?

Evan MacMillian: 7:46

Yeah, like a lot of companies. It's a little bit of serendipity. And that was the case with grid space, I was coming off of the acquisition of my last company, and it was working on new things. And two of my really good friends and smartest friends from undergrad we're available. So that summer, we got to work and slept on couches and bought a domain name. And we were off to the races with grid space.

Dave Michels: 8:11

That sounds pretty technical. And so obviously, you're having a lot of fun, and your roommates are having a lot of fun. Sounds like a geek fest. But how do your customers use grid space?

Evan MacMillian: 8:22

So that's a great question. So we serve a lot of customers, customers like USA and square use grid space primarily for their phone calls, their voice calls, and some for web chat as well. And they use it to scan and automate all of the sort of conversations that they're having with customers. So our product provides a suite of observability and automation tools. And it really simplifies the contact center operation for agents for QA managers, for executives, folks, that before would have to go into an individual call and listen to it to understand what was going on. So we kind of make that understanding work at the call population level. And we do it in real time. You know, if you just zoom in on USA, before grid space, they were using an on prem recording appliance, and it was meeting their needs from a regulatory standpoint. But it was really tough to use it to quickly access a call and understand large collections of calls and real time. And we were originally the crew that came in and made real time call understanding at massive scale 10s of 1000s of concurrent calls possible. So it was pretty cool. And then we we built a lot of what we would call an AI powered workflows around that core conversational understanding component.

Dave Michels: 9:45

The grid space really isn't that well known. I think I learned about it from some folks over at Google. But you mentioned customers like USA and you just in square. I see on your website, you've got Bloomberg. These are not small customers, small companies. How do customers find How did they find you

Evan MacMillian: 10:01

mostly through podcasts that I just didn't exist. So this community is pretty small. And is only a handful of companies right now that are building full stack products for speech automation. I mean, most folks are leveraging third party API's to kind of begin to get at this problem of understanding conversational speech. So the folks that are I would call them power contact center users, users, the really innovative folks are finding us and in the last year, that group has grown, so we have more people finding us organically. And we've teamed with a handful of amazing channel partners, Google's among them.

Evan Kirstel: 10:41

So I like to think of myself as a member of the speech community sounds like a very exclusive club, although I don't have a membership card. So what about you? Do you know everyone in the speech community?

Evan MacMillian: 10:53

I don't know everybody in this speech community. Yeah, I guess everybody loves a good speech. But the speech community, I was really trying to get added to the speech and language understanding community. And they're the folks that are sort of working on some of the most interesting components for making sense of large collections of conversations or otherwise kind of automating them.

Dave Michels: 11:15

I also see on your website, several familiar brands, you got Avaya, Cisco, Tullio Amazon, are these are the companies that you want to acquire a grid space.

Evan MacMillian: 11:28

Now our customers use these companies products to make calls one of the benefits of grid spaces, we're really telephony, and cloud agnostic, meaning that, you know, whatever your telephone system is, and wherever you're doing most of your cloud work, we know we're going to work with it. Right now, a lot of the voice minutes are still switched on prem. And that makes grid space especially interesting to many customers, because they're not quite ready to move their entire contact center operation to the cloud. But they really want to know what's going on. And again, to figure out what parts will really benefit from Cloud infrastructure. So we kind of design the product in such a way that we can work with all those folks that are still switching their calls on prem and work with the cloud that they feel most comfortable with. Great model.

Evan Kirstel: 12:19

So we've covered customers and partners, and some interesting use cases. But I'm still pretty unclear on what grid space actually does. Can you take us walk us through the core offering?

Evan MacMillian: 12:29

Yeah, absolutely. So grid space sift is our flagship product. And if you run a website and run a bunch of servers that power your website today, you might use a cloud observability tool to see what's going on and troubleshoot issues with your website. Now, if you run a contact center, that observability tool that makes sense of server logs doesn't work so well for conversations. And that's one of the big reasons why we built grid space was we want to give that benefit another sort of workflow benefits to contact center operators. So you can actually peer into those unstructured conversations, see patterns, check for procedural compliance, whatnot. And that's really how grid space works. So at a high level, we scan every single word of every single call in real time, we provide these AI power workflows, where the understanding components are actually sort of making context generator, shift operations, like QA and dispute management go faster. And then for executives and managers of these agents, we provide the dashboards that give them that holistic view of what's going on. So that's what we got started with, that's what's processing the most minutes. And then recently, we announced grid space grace. And we're really, really excited about grid space grace, and that's our virtual agent offering and it's built on the same foundation is grid space.

Evan Kirstel: 13:52

So we've heard of Alexa. And I've heard of Siri. So who or what is grid space grace, tell us more about her. It,

Dave Michels: 13:59

Evan, you're engaged. He's always trying to get phone numbers. You know, he must Grace's phone number. But

Evan MacMillian: 14:06

sometimes grace calls you if he really likes you. So Grace is exciting, because it takes on the biggest burden of running a contact center, which is taking and making calls. And that's right now something that we get asked for all the time, we were providing this observability software, and we sat down with these executives, and they're like, Well, can you just answer these calls for us? And it's a really difficult thing to do. But we were working on the problem of understanding these calls before and we were quite good at it and grace actually sort of benefited from that core understanding capability and we gave grace a mouth and we gave grace a dialogue system capability, tiny brain and that was enough to make grace be able to handle some 1015 sometimes 25 minute conversations with callers about specific health care, onboarding issues and financial services. Questions.

Evan Kirstel: 15:00

So you mentioned a tiny brain is that a brain bigger or smaller than Dave Michaels just to set the expectation here? No comment, okay, that just sort of my personal interest but better than a virtual brain? Has grace been helpful to customers what's been the result today.

Evan MacMillian: 15:17

So Grace is really helpful at gathering information and triage and caller requests. And there are all sorts of procedures that take people away from their contact center jobs, or in the in the healthcare arena, sometimes actual point of care responsibilities, nurses have to jump between the clinical setting and the phone bank. And that's not so great. So Grace is really, really good at gathering information and routing callers deeper into a contact center. So that's where we shine today. And I think we're going to add more and more of these base bots that can handle more and more complicated inquiries and requests.

Dave Michels: 15:59

So these are kind of unusual names. You've got grid space sift, which is your main product, you've got grid space, Grace, which is your IVR solution. I'm not seeing the family there. I would like George and grace are George and Gracie, or you sift and stir those kinds of words. But so I'm almost afraid to ask you. What about the name grid space? What

Evan MacMillian: 16:21

does that mean? Well, we gave ourselves one day to come up with a name that wasn't too embarrassing and had an available.com name, and reference something cool. Like space, my co founder, Anthony actually worked on curiosity, the Mars rover. And that was a nice collage to space.

Dave Michels: 16:43

I actually, I thought that might be a space, I was thinking of the holodeck rooms got that kind of a grid pattern. But so we were close. We were close. But let me ask you about the virtual agent business. Because isn't that a really tough business? Everybody claims they have a virtual agent. And customers have a tough time evaluating whether one is better than the other?

Evan MacMillian: 17:01

Hmm. Yeah. So it'll take the first part of it hardest, good, hard, valuable, fun projects, bring together the best teams and have the most impact on the market. So we've never been afraid of hard and that's a Max Levchin, quote, hard, valuable fun is the sort of sweet spot for a technical company. And speech understanding is not just hard, it's very valuable and fun, especially in the contact center domain. So we like hard. I think when it comes to differentiation, customers are getting a lot smarter. They're no longer just trusting the marketing material. They're calling into these things. And they're giving the virtual agent companies, real challenges, and real business objectives to hit. And that's where we really shine. You know, we love it that customers love talking to our bots. And we love that they get the job done. And

Dave Michels: 17:57

you're telling me this wonderful, complete answer that didn't answer my question is what makes grid Space Race different

Evan MacMillian: 18:06

was different. Because, you know, Grace came out of an organic product development with customers and customers, were already using our flagship product that provided speech understanding as transcribing and analyzing billions of billions of calls minutes. So you know, we really came at it from the perspective that, you know, great understanding on conversational contact center speech was the first step. And it just so happened, we were already doing that first step, when the demand for these kind of capacity reducing bots really surged.

Dave Michels: 18:42

Okay, so you've provided the brochure now all except that, so what does a customer do to actually evaluate your IVR? Check to know that it's better?

Evan MacMillian: 18:52

Oh, have a conversation with her. And that is starts with a conversation and then collections of conversations to folks inside at your company, and then eventually, reference and users and customers. So starts with a conversation, I got to hold your feet to the fire on one more point here. You've said earlier, that you build your own stack. Who does that? There's so many good options out there for your stock? Why don't you just choose one of the other setting options? Well, we try to bench of parts. And the challenge was twofold. One is we completely figure out how to put them together in a clear way. And there were also things that we wanted to change and we couldn't because we didn't build them. So as you're assembling a product like grid space grace, you find out really, really quickly that the performance of the system is extremely sensitive to not only the performance of individual parts, but the combination of those parts. And you know, any disturbance and a component can like literally reverberate across the entire system. So You know, we took the harder approach that by controlling and innovating on the left at layer, the conversational understanding layer and the dialogue system layer, and even the sort of UI layer for the operators, we could deliver much better experience.

Evan Kirstel: 20:17

So there's a perception that voice is kind of dead. It's last year's news. But that's certainly not the case. Really, we've seen popularity of Siri and Google voice assistant, and really a doubling down of investment in those technologies. nuans just got acquired by Microsoft for a huge valuation. So tell us about your perspective on the voice technology market in general?

Evan MacMillian: 20:42

Yeah, I mean, nuance and Siri are Sri spinouts. Just like grid space. I mean, voices just getting started. And what's possible today is, it's pretty incredible. When you think about where this technology started out. And

Dave Michels: 20:58

Evan, telling you that people are saying voices dead, I reiterate that I'm getting this all the time voices dead. You're saying it's just getting started? I think you need to clarify this discrepancy?

Evan MacMillian: 21:09

Well, I mean, let's look at problems like where we can even get out voice. So you know, there's something like 500 million Apple devices out there. And they all have really high quality microphones, and people are talking to Siri more than they ever have, in the past, people are spending hours and hours and hours and audio experiences like clubhouse and on Spotify, they're listening to an audio podcast right now. So I think people are getting more and more comfortable with audio. And audio is getting a lot better because of that usage from a networking perspective, and from a understanding perspective, as well as a lot more going on. So I don't think it's that at all.

Evan Kirstel: 21:53

So how did the pandemic affect you at grid space? And more importantly, your customers? What was that? You know, last 18 months? What did that look like for you?

Evan MacMillian: 22:01

Yeah, so you were working with a lot of digital native financial service companies already, what it really did was accelerate the adoption of our voice observability and our virtual agent offering with healthcare companies. I mean, that's where we saw the really big change. with everybody working from home, these operators needed a way to virtually walk the halls and then also add capacity to their workforces with offerings like grid space. So I think it was a really, really interesting moment where folks had to do something that they might not have done as quickly if it wasn't for the pandemic.

Dave Michels: 22:40

So is this conversational intelligence that you're offering or enabling? Is that really a feature? Or is it a product? Is there a market or demand for a standalone conversational intelligence solution?

Evan MacMillian: 22:53

What a great question. I think the answer is absolutely yes, there is absolutely a market for AI first companies and enterprise communication, specifically companies that are helping contact centers, what we've seen the market for voice observability in automation is going to be actually quite big, probably 10 times bigger than the market for call recording an IVR, simply because these products are doing so much more work. And they can be sort of adopted and deployed without a lot of work without the kind of effort that was needed to buy these technologies in the past. So I think it's absolutely going to spawn really important products, and probably ecosystems as well. I mean, the incumbent products take call recording or legacy IVR, they just weren't architected for AI. And for a legacy Call Recording later, to kind of get to the point where they can do all the stuff that people are talking about in Wired Magazine, it's gonna be a big reach. So I think it's an exciting time for the field and looking at the

Evan Kirstel: 24:04

field in the marketplace, what is nuances acquisition by Microsoft signal to you about where the markets going?

Evan MacMillian: 24:11

Well, in new offices founded 29 years ago, and back then professional services were not so much a go to market strategy, but it was really a necessity for speech companies. And mean, this is 10 years before you could run anything on GCP AWS Azure, let alone ginormous state of the art ml model for streaming speech. So you know, this company got started when the assumptions of what you could do and how you could go to market were just very different, and are now at grid space. We can click some buttons and we can run beat pre train models on 50,000 plus concurrent calls in the cloud. It's really pretty incredible what was possible back then almost not And they still found a market and what's possible today. So we started grid space because these contact centers, were getting frustrated with the professional services model. And they want to deploy faster, upgrade faster and recover from issues and adapt to changes in their business faster. And that's why they wanted more of a software solution.

Dave Michels: 25:23

I'm glad you mentioned that 50,000 concurrent calls. Because I think when people think of small businesses, they inherently think that they can't scale. You obviously you can scale. Well, can you clarify how you do that?

Evan MacMillian: 25:36

It's an interesting time to run a really technical company. I mean, uniswap is not a lot of people. A lot of people don't know that WhatsApp was a pretty small company as well. And it was handling more communications and most major telcos combined. So when you get the system design, right, scale was no problem. What's exciting right now is just seeing that sort of system design really take hold in the market and handle more and more and more calls. So that's the exciting part for me, and I really think this market is just getting started.

Dave Michels: 26:12

So you've got the architecture, right? You've already figured out the cloud services. So it sounds like you're pretty much done. Is there anything to look forward to in your endeavor? What's on your roadmap? So look, it'd

Evan MacMillian: 26:23

be, I think this mark is gonna continue to grow. I think augmentation, automation of voice calls is going to continue to pick up, I think, in some areas of the market collaboration, you know, incumbents will roll out new offerings, and they will really have the upper hand. And then in other parts of the market, the legacy folks are going to really struggle to compete with the startups. I think what we're most excited about is going from what was really an r&d project to taking on our customers biggest problems. And that's actually helping them through these workflows in the voice channel in the most satisfying way possible, and in the most efficient way possible. And, you know, what we're seeing were deployed in healthcare systems is that sometimes the most pleasing and efficient way to ask a Medicare patient a bunch of questions is actually to slow down and have a much slower call and a more detailed call about their situations. And this is possible with virtual agents and not so possible when you have nurses running between the contact center room and the patient rooms. What you can do now is some pretty incredible empathetic stuff with both customers are financial service companies and patients. Nobody likes talking with incumbent IVR. So this is a huge opportunity for God's face. So my

Dave Michels: 27:51

algorithms may not be as effective as yours. But I'm detecting negative sentiment. I heard you say that it's going to explode. I heard you talk. You say that. Nobody likes talking ivrs. So are you sure that things are looking up? Oh, well, I

Evan Kirstel: 28:06

think you're running the old model. You got to upgrade it. Well, I'm sold. And I've been around voice technology for 30 years. My first job was dialogic and 1992. So super excited. I'd love to get more info. I'd also love to invest in the company. I'd have my own virtual cryptocurrency called b2b on the rally blockchain and would you accept that maybe as a form of payment? Only if it's carbon neutral? Better than Bitcoin? That's better than Bitcoin?

Dave Michels: 28:34

I think we need to wrap this up. This may have actually been the first Talking Heads episode with an intelligent Evan I think that's when you're done with grid space. You may want to consider a career in podcasting. Edmund McMillen pays as well as money startups do pay zero. Yeah, absolutely.

Evan MacMillian: 28:52

Wow, well, maybe some CreateSpace gray spots can join me for the ride. And we can have we can have a the first virtual agent podcast on enterprise communication software.

Evan Kirstel: 29:06

Wow, that was impressive. I'm beginning to believe anyone named Evan is kind of super intelligent.

Dave Michels: 29:11

Well, I had a similar thought with a slightly different conclusion. But yeah, I think that was a really interesting interview. And I have to say, voice technology is hot. You know, in my insider reports, I write about all the companies that are getting funding. There's so much going on in voice technology right now. Yeah, and

Evan Kirstel: 29:28

Evan mcmillon. That grid space is really shaking up quite an old school sector and contact center and customer experience. It's nice to see innovation. It's nice to see a young new entrant into the space.

Dave Michels: 29:42

I forget what you ask them, but you ask them something about nuance, and theory. And he says, Oh, yeah, those two are part of the Sri graduates and so are we it's like, oh, wow. Well, we're not going from

Evan Kirstel: 29:53

that was pretty impressive. He's a Stanford Graduate. So like I said, you know, Evans are the smartest folks on the planet. But we have another guest coming up who also is pretty smart.

Dave Michels: 30:03

We now return to our regularly scheduled podcast schedule. This is the end of our bonus episode and I think we haven't mentioned it in our last episode it was Kira maca gone from ringcentral now heading up ringcentral ventures will be our next guest.

Evan Kirstel: 30:17

Wow should be lots of interesting insights into ringcentral is partnerships and investments and a peek into their future should be fascinating.

Unknown: 30:28

You can make a conversation. Oh man. I gotta get out of here.