 From Palo Alto, California, it's The Cube, covering the conference board's sixth annual Innovation Masterclass. Hey, welcome back, everybody. Jeff Frick here with The Cube. We're at the Innovation Masterclass at Xerox Park in Palo Alto. Really excited to be here. Never have been here, surprisingly, for all the shows we do, just up the hill next to VMware and Tesla, and, you know, this is kind of the granddaddy of locations and innovation centers that's been around forever. If you don't know the history, get a couple books, you'll learn it pretty fast. So we're excited to be here, and our next guest is Anthony Brighton, four-time founder and CEO, which is not easy to do. Again, check the math on that. Most people are successful. A couple times, hard to do it four times, and now he's a co-founder and CEO of Directly. So Anthony, great to see you. Good to be here. So Directly, what is Directly all about for people who aren't familiar with the company? Most companies are excited to, and pursuing the opportunity of automating up to 85% of their customer service. That's the ambition and giving customers a delightful answer in their first experience. Most of those companies are falling down out of the gates because there are content gaps and data gaps and training gaps and empathy gaps in the systems. So we build a CX automation platform and it puts experts at the heart of AI, letting these companies build networks of product experts, and then rewarding those experts for creating content for AI systems, for training AI systems, for resolving customer questions. So let's back up a step. So Zendesk is probably one we're all familiar with. You send in a customer service, notice a lot of times it comes back, customer service is Zendesk. But you're not building kind of a competitor to Zendesk. You're more of a partner, if I believe, for those types of applications to help those apps do a better job. We are, we're a partner for Zendesk, we're a partner for Microsoft Dynamics, for Service Cloud and the like, and essentially building the automation systems that make their AI systems work and work better. Those are pure technology systems that often lack the data and the content to deliver AI at scale and quality. And that's where our platform and the human network, the experts in the mix, come into play. So we could probably go for a long, long time on this topic. So what are some of the key things that make them not work now? Besides just the fact that, you know, it's kind of like the old dial-in systems. Like I just want to hit zero, zero, zero, zero. I just want to talk to a person. I have no confidence or faith that going through these other steps is going to get me the solution. Do you still see that on the online world as well? No, there are very clear gaps. There are four or five areas where systems are falling down. AI project mortality, as I refer to it. Very few companies have the structured data that systems need to work at scale. On the back to feed the whole thing. That's right, labeled, structured, organized data. So that doesn't exist. Many companies don't have the content. That's a secondary. They may have enterprise knowledge bases, but they're five years old. They're seven years old. They're outdated. They're not accurate. You know, many companies don't have the signal. When an automated answer is delivered, they have to wait for a customer to rate it. And that tends to be really poor signal on whether that answer was good or not. And then last, many companies just don't have the teams to maintain these algorithms and constantly tune them. And that is where experts at the heart of a platform can come into play by building a network of product experts who know the products inside and out. These could be Airbnb hosts for one of our customers. These could be Microsoft Excel users in the Microsoft example. Those experts can create that content, train the data and actually resolve questions. You know, filling those gaps, solving those problems. I'm just curious on the expert side, how many, or I don't know if there's best practices or if there's certain buckets depending on the industry, of those expert answers are generated by people inside the company versus a really kind of active, engaged community where you've got, you know, third-party experts that are happy to participate and help provide an info. Over 99% of the answers and the content is actually generated by the external network. 99%? 99%? Now we can, now we certainly start, you know, you start with sources of enterprise knowledge, but it's a long, hard, arduous process to create those internal knowledge bases and companies really struggle to keep up. It's Britannica. By the time you ship it, it's outdated and you have to start all over again. The external expert networks work more like Wikipedia. Content constantly being organically created. The successful content is promoted. The unsuccessful content is demoted and it's an evergreen cycle where it's constantly refreshing, overwhelming the external. Overwhelming, and how do, I mean I can see where there's certain types of products, I was talking to somebody else the other day about, you know, Harley-Davidson, one of the all-time great brands, people tattoo it on their body. You know, there aren't very many brands that people tattoo on their body, right? So easy to get people to talk about. Motorcycles are some of these types of things, but how do you do it for something that's really not that exciting? What are some of the tricks and incentives to engage that community or is there just always some little core that you may or may not be aware of that are happy to jump in so passionate about those types of products? There are definitely some companies where there's very little expertise and passion in the ecosystem around it. They're few and far between. You know, if you find a product, if you find a company, you can find people that rely, love, and depend on that company. You know, I gave some of the B2C examples, but we've also got networks for enterprise software companies, folks like SAP, folks like Autodesk. And those networks have experts that are developers, resellers, VARs, systems integrators and the like. You know, in the overwhelming majority of cases, the talent and the passion exists. You just have to have a simple platform to onboard and start tapping that talent and passion. So if I hear you right, you use kind of your encyclopedia Britannica, because that's what you have to start to get the flywheel moving. But as you start to collect inputs from third-party community, you know, you can start to refine and get the better information back. And I ask specifically that way because you mentioned the human factors and making people part of this thing, which is probably part of the problem with adoption, right? Is I want confidence that there's some person behind this, even if the AI is smart. I want to at least feel like there's some human-to-human contact when I reach out to this company. Yeah, that is critically important because the empathy gap is real in almost all of the systems that are traditionally out there, which is when an automated answer is delivered, you know, in a traditional system, it typically has a much lower CSAT than when it comes from a human being. What we found is, when you have an expert author that content, when his or her face is shown next to the answer, as it's presented to the user, and where he or she is there to back it up, should that user still need more help, there you retain the human elements that personalize the content, that humanize the experience, and immediately get big gains in CSAT. So I think that empathy piece is really important. Right. I wonder if you could share any specific examples of a customer that had an automated, kind of dumb system, I'll just use that word compared to what they can do today and some of the impacts when they put in some of the AI-powered systems like you guys support. So one of the first immediate impacts is often when we go in a automated or unassisted system, we'll be handling a very small percentage of the queries, you know, and percentage of the customer questions coming in. And people are going straight to zero, they're just like, I gotta go to a person. Yeah, we're mostly in digital channels, so less phone, but yes, because the content there, because the content isn't there, it doesn't hit and resolve the question in that frequent array, or because the training and the signal isn't there, it's giving answers that are a little off base. So the first and lowest hanging fruit is with a content library that gets created that can get 10, 50, 100 times broader than enterprise content pretty quickly. You're able to hit a much broader set of questions at a much higher rate. That's the first low hanging fruit and kind of immediate impact. And is that helping them orchestrate, coordinate, collect data from this passionate ecosystem that's outside the four walls? Is that essentially what you're doing? It essentially is. It is about companies having these ecosystems of these users, millions of hours of expertise in their head, millions of hours free time on their hands, and the ability to tap that in a systematic way. Wow. Shift gears a little bit. You are participating on a panel here at the event talking about startups working with big companies. And there's obviously a lot of challenges starting with vendor viability issues, which is more kind of selling to big customers versus necessarily partnering with big companies. But what are some of the themes that you've seen that make that collaboration successful? Because obviously you've got different cultures, you've got different kind of rates of the way things happen. You've got, you know, be where the big company will meet you up in meetings all the time when you're a little startup, right? They'll kill you accidentally just by scheduling so many meetings. What are some of the secrets of success that you're going to share here at the event? So we've got experience in that. Microsoft is a partner of ours. Microsoft Ventures is an investor. You know, I think the single biggest key is an aligned vision and a complementary approach. You know, the aligned vision where both the startup and the partner are aiming for a similar point on the horizon, you know, and, you know, for example, the belief that automation can delight, you know, a very large set of customers by providing them a good instant answer, but complementary approaches where the core skill sets of the companies round out each other and become less competitive. In this case, you know, we've partnered with Microsoft's Best in Class, AI platform and Cognitive Services, and we're able to tap and leverage that. We're also able to bring something unique to the equation by putting experts at the heart of it. So I think that architectural structure in the first place is a great example of kind of getting it right. Right. Is your experience that's been pretty easy to establish at the head end of the process so that you have a kind of smooth sailing ahead? No, I don't think it's easy to establish at the head of the process. And I think that's where all of the good work and investment needs to happen, you know, up front on that kind of shared vision and on that kind of complementary approach. I think it is probably 20% building that together, but it's also 80% just finding it. You know, the selection criteria by which a corporate partner picks a startup and the startup partner picks the corporate partner, I think just selecting right, you know, is the majority of the challenge rather than trying to craft it, kind of mid-stream. Going and saying, this is not gonna, you know, yeah, so if it doesn't feel good at the beginning, it's probably not gonna work out. Right, it's about finding it. It's a little bit like the venture analogy. Do they find great companies or do they build great companies? You know, probably a little of both, but that finding, you know, that great company is a large part of the process. Yeah, it helps. So, before I get to the last question, so again, four successful startups that does not happen very often with the same team. I look at your background, you're a psychology and philosophy major, not an engineer. So I just love to get kind of your thoughts about, you know, being, you know, a non-tech guy starting, running, and successfully exiting tech companies here in Silicon Valley. What's kind of the nice thing being from a slightly different background that you've used to really drive a number of successes? So I think the, yeah, I think two things. I think one, coming from a non-tech and coming from a site background has given us an appreciation of the human elements in these systems that tech alone can't do it. I'd say personally, one of the impacts of being a non-tech founder in this valley is a heck of a lot of appreciation for what teams can do. And realizing that what teams can do is far more important than what individuals can do. And I say that because as a non-tech founder, there's literally nothing I could accomplish, you know, without being a part of a, you know, a team. Right. You know, so that I think non-tech founders have that in spades. You know, a harsh and frank realization that it's about team and they can't do anything on their own. Well Anthony, thanks for taking a minute out of your time, good luck on the panel this afternoon and we'll keep an eye, watch the story unfold again. Yeah, I appreciate it. Thanks very much. He's Anthony, I'm Jeff, you're watching theCUBE with the Master Innovation Clavids, Xerox PARC, thanks for watching.