 Hello everyone and welcome to this CUBE conversation, which is part of our startup and founders series. I'm Dave Vellante, I'll be your host and we're pleased to welcome two guests. Kreet Prasad is the CEO and co-founder of AptEdge, a firm that's using natural language processing to automate customer service. And Bobby Napaltonia is an advisor to the company's former Twilio and Salesforce and some others. Guys, welcome to the CUBE. Thanks for taking some time out of your busy schedule. Thank you so much, Dave. Appreciate the time. Kreet, let's go back to the beginning. Why did you start the company? What's the problem that you saw that you're solving today? Yeah, so I've been in the industry for about 20 years. I spent early years at several different great software companies like Salesforce, AppDynamics. And my responsibility of these companies was around product, R&D and working with our customer service teams. So for me, one of the biggest pain points and challenges that I saw for our customer service teams is around customer escalations. How do we help our customer service teams reduce the escalation impact when they're solving a customer problem? And at the heart of that pain and that problem, for us, was always what we call the knowledge problem. Customers escalate when they can't get answers quickly. When we're starting to see that trend become more and more common. And so what we wanted was to look at solving that problem by building a technology that helps customer service teams get to knowledge information more quickly. Great, we're gonna get into that. Bobby, what's your role? How did you get involved? So consider me sort of like a chief helper. I've known a crit for a couple of years and got reintroduced for through their investor over at National Grid. And I have to be honest, I've been at this 35 years since the green screen days. And what got me involved was the excitement of where we're about to enter what I think is one of the biggest changes that we've seen in the industry. And you've mentioned my background. So when you think about those changes of APIs and cloud and platforms and most of the world runs on it, when I saw what we're up to, we're not going to change the landscape. We're going to change the way that people work. And I believe that there's a day that just like at Salesforce, we have no software. There'll be a day where we have no applications. And we're looking forward to ushering that day in with you. Oh, interesting. This is a good conversation here. Let's talk a little bit about the products. Answer GPT, I think is your flagship, right? Tell us a little bit more about your products at Creek. Yeah, so answer GPT was built around the idea that if we can connect to different areas of information and knowledge, essentially different areas of knowledge that customer service teams need to access to get to the most complete correct information the first time when they're working with a customer. If we can connect to those areas and be able to create answers, find the knowledge, find the answer, get to responses immediately. We can arm our customer facing teams with instant answers and accurate answers to questions from customers in real time. And that was really the genesis of why we coined it answer GPT. It's really about reducing the need to go look in different places and getting to the answer as quickly as possible. Yeah, so I want to learn more about this. I mean, at theCUBE, we have this 13 years of content through conversations like this. And it's all in the cloud. We've got GPT, we've got vector databases. We have open source tooling and we put it all together and you can actually talk to, you know, decades of CUBE. So what do you actually sell to a customer? And how do I license it? Yeah, yeah, that's a great question. So we, you know, our customers, often their support services teams are required to use a variety of different products for capturing knowledge, information and looking for answers. They'll usually be using a support system that could be Salesforce, send asks, there's now, there's different products that are used for that. They'll be using, there'll be an external knowledge repository and a system of knowledge for internal information. They might be accessing bug tracking systems to understand, hey, is the support problem related to an existing bug that our incident or enhancement that our development team is working on. And they're also tracking knowledge and collaboration channels like Slack or Teams. And then on top of that, we often also are now finding that there's knowledge in community systems, LMS learning management systems. And so because of this, we find that the customer service, modern customer service rep today is bouncing between all of these different systems of information to find the most relevant context and the right answer to a customer question that they're working on as part of their support requirement. And so what we're really focused on is how do we bridge all of that knowledge to create the most intelligent experience that makes every support rep in a company the smartest rep in the company. And Bobby, when did you exactly join a Crete in this journey? You lost the company in 2018, right? When did you start? Approximately two years ago. And to be honest, we were making good progress, but it wasn't until ChatGBT got to show the new delivery mechanisms. And if you think about it, what were the excitement for me is we're living in the answer economy. And we all know that because we're done waiting for data scientists to produce a report about those 13 years of show. When you and I want to go somewhere, we pick up our phone and we type in the Uber and I get an answer at this amount of time, this amount of money and I get there. We believe customers should have that same experience even if you're asking that question. And this is where the trust comes in. Today we're starting with the company's own internal information. So we're gonna hope that your trust where the information is gonna be exposed in the most expeditious manner to serve your customers better. And as it go to market executive, I believe that's the best place in times like these in which we'll be able to grow our businesses by better serving the customers that believe in us, showing the others that they too can join in sort of this journey because change is happening. And they will either get blockbustered or they're gonna come along for the journey. So the AI heard around the world as we sometimes call it. So did it change the way you thought about the business or did it just wake everybody else around you and say, okay, finally the market is coming toward us? Yeah, I can speak a bit. I mean, it really amplified what we wanted to do from day one. Our goal was how do we help the customer service teams get to information, get to the right response as quickly as possible because they're getting flooded with a lot of customer issues and escalations happen when answers aren't found quickly or the wrong information is sent. And so we looked at this from a point of view of how do we help the customer service teams get to information context as quickly as possible in a world where knowledge is only growing. We're probably creating two in the next couple of years create more knowledge in the history of knowledge. On top of that, you've got fragmentation of information, siloed knowledge from different systems of information including information that might be in PDF repositories that support teams access for instructions or processes that they have. And on top of that, you've got this signal of customer impatience. We as consumers of products expect immediate responses when things break or things don't go as planned. And so with that, we're moving to a world where this technology really allows us to deliver the end experience for our customers in a much more compelling way that ultimately completes that feedback loop for their use case. So where are you getting traction? I know from your website, you got a bunch of logos for customer logo, Cisco, CloudBees, Mile IQs, several others. I'm interested in why they buy and why they specifically buy from you. And Bobby, of course, you know product market fit, macro bears your lectures all the time on this. Where are you in that whole product market fit journey? We could not have set this up better. So you said something that I love to speak about. We don't have product market fit. We have something far better. We have something called founder market fit. Having delivered products the world consumes over the last 35 years, there's nothing better than knowing that you've walked a mile in those shoes, you then solve that problem. And you heard the story of the beginning where Akrit created a product and the next day had 3000 tickets. I guarantee you, if the product leaders were to be honest on this call and answer it, they've all had that same experience. So it's not as if this is a one-off problem. That's probably the first part. The second part is the budgets have been constrained. This is actually a CFO sale. I believe where you say, look, your business needs to grow and you only have $5, what will you do now? So you have to do that. And I believe that you're gonna see technology not be about, you know, let me go to this site, let me hear about Gartner G2 and let me hear what other people, I wanna know where it's being released in the wild. Look, that building that you see in the background, that's all about trust. We're trust disciples from this religion of we truly care about that experience. And I think you're gonna see where we grow. You mentioned that our current customers, they're the early adopter leaders. At Salesforce, our first big customer was Cisco. It then dominoed that we owned high tech. We're following in that same path because those are typically early adopters of technology. And we're seeing big uptakes in manufacturing that realize I can't keep throwing money or bodies at the problem. I must embrace technology. This is so right on. I mean, you see it in the data. IT budgets, you know, technology budgets are compressed. Maybe they grow in 2.5, 3% this year. And yet post the chat GPT announcement, budgets are going up on AI. You can see it very clearly in data. So it's taken away from everything else. And the bottom line is, see if I was like, how can I not hire more people? Because labor is my biggest cost. And so that is a trend that is a tailwind for you guys. You've raised, I want to say about 13 million in venture capital, which is not a huge amount by kind of pre 2022 standards. So, you know, where are you in that regard? Where are you investing now? How is the tech downturn, you know, affected your business? If at all, has it helped? Is it still a headwind? Maybe talking about that a little bit. Yeah, I'll share a little bit on that. You know, we actually are finding that the tech downturn in some ways is definitely a very positive headwind or tailwind for us. And that's primarily because every company we're talking to is getting the direction of do more with less. And, you know, they still want to keep growing their business and achieving greater outcomes every year. But the CFOs are not giving budget to linearly scale. Let's say a support organization or a customer facing organization. So we're seeing clear signals where every team is, every executive in a business is being asked to show how they're empowering their own team to do more with the same staff, right? And that's been a great equation that we've been able to demonstrate with our technology of how we're actually helping every member of their team work faster, smarter, work more efficiently, be the best rep in their organization. That's been one component of it. And the other that we're also seeing is businesses are looking at kind of in some organizations, you know, support is looked at as a cost center. And so how do we drive more efficiency out of the organization that ultimately ties back to the gross margin of the business, right? And how do we show those metrics that show us on a clear path to faster profitability? And so because of where we're focused, we're actually starting to see really, really great signs of being able to deliver on that value for customer service organizations specifically and it's really been more of a tailwind for us with the advancements of this technology, maturing and where, you know, companies are wanting to invest in empowering their people through technology. Now, I mean, it makes sense. I mean, if you can show that you can reduce labor cost and everybody gets panicked to say, oh, what does that mean? We're going to lose jobs? Well, maybe, but over time, it's going to increase jobs. We know we have a productivity problem. I want to talk more about your product in your GPT. When I think about the traditional customer care, you have that human touch, your relationship building and you know, how do you strike the balance between the efficiency that we've just been talking about through automation and that human centric, you know, high touch approach? Are customers concerned that they're losing that personal touch? Do you sort of, are they tiering? Like if you pay us more, you get the human touch or and or is your product, you know, functional of that you can actually hot mask some of those, you know, traditional bot inconsistencies. Yeah, yeah. So, you know, with the way that we've designed answer GPT and our core product and to end from day one, it's to be an A-native product that while we're understanding the context of what's happening between different systems of knowledge that our support team is accessing to be able to get to answers more quickly, we're also powering that experience around a personalization layer. We're being able to create customer-facing responses that are personalized to the reps that are working on a support case based on what's happened in the support case previously. So taking into account conversation and context of the support case or the issue that a customer is reporting, being able to marry that context with the knowledge that the team has internally in their domain about potential solutions. Maybe we've seen this problem before, maybe this is tied to a recent workaround or a process update, gathering that context. And then really, you know, in some cases providing the full complete response that a rep can use to send some information back to their customer to close the issue or getting them, you know, 90, 95% of the way there where they can personalize that experience a little bit more and allowing our system to kind of understand that context and eventually keep improving the experience for the frontline teams based on how current teams are using the product and what they're learning and tweaking the answers. Thanks to that. So Bobby, you know, let's talk about metrics a little bit. Executives want their customer SAT scores to be better. NPS score, however they measure it. How are customers telling you that they want to quantify the impact of products like answer GPT? Is it customer satisfaction? Is it NPS? Is it sort of we're cutting labor costs? What differentiates you from some of the AI driven customer service tools that are in the marketplace and how do you quantify that? That's a really good question and I love getting this and again, this is another one where we didn't even prep you for this but we've got some phenomenal video testimonials of one of our CMOs and she's been that in that role for probably more than two decades and for the first time she got a 100% CSAT score. Now, if you know me and those people watching this do I'm an OBS person. So when I say that that's a pretty meaningful statement and we dug into how she was able to achieve that and then they literally get a response for everyone that exit to call and the experience was such that through the technology and your point, if you really look back at your question the bots was an evolution of outsourcing that started 25 years ago, your mess for less and I shipped it offshore. Now I just shipped it to this thing that I thought was smart but it really wasn't smart and that's where we bring this intelligent layer and to make that bot just be another touch point multi-channel attribution sort of what you know this industry is now growing up on. And so as we see through that I actually believe that with the help of folks like you were going to educate hey, you could trust it, it's safe today here's the awareness, here's what you can do it here's those use cases and that's where we're going to talk into the marketing budget. Again, you heard me say the CFO budget these are not typical customer care places that would send it because those are cost centers and actually maybe at the end of this we'll annotate some of those videos so your audience can actually hear from our customers themselves that the proof is in the pudding. Yeah, send those to us we'll put them in the show notes. I'm interested too in how you integrate I think about security it's like another tool and I can see AI becoming oh wow it's going to solve this problem or that problem or that problem. So you've got to, it's one thing to understand algorithms and large language models but you got to integrate into existing systems and I know on your website you've got this graphic I see service now I see JIRA, Confluence, GitHub, Slack, Microsoft tools how does your product ensure that it's interoperable with all these diverse ecosystems and tools and platforms out there in this space? Yeah, so security and trust is our number one value and we have been very pragmatic and forward thinking about that since day one but the way that we have designed app is in experience where we can plug into the systems and provide value within the first week and part of what we've had to do in ensuring that complete picture is a solution that is compliant with all the necessary industry requirements like SOC2, HIPAA compliant but also on top of that the way when we connect to these systems of information we connect in a limited sense where we're only connecting to the information that's pertinent for customer service organizations. So when we connect into a system like Slack or Teams we're not looking at all the information we're only looking at information that's pertinent to what's related for a Q and A repository in a Slack channel that our support team might be referencing to get to an answer to a customer facing question and really tapping into that collective knowledge of a customer service organization so they're able to get to the most complete information as quickly as possible in a way that where the data and the information is in a very controlled environment. So another challenge that we see with just AI in general I was talking to somebody the other day, they said you know, have you noticed that the chat CPT is kind of getting worse as more people use it? I said, I don't know, I use it in certain ways but somebody else said to me entropy is winning and so as knowledge bases evolve you get more randomness and businesses are constantly updating their offerings, their policies so how do you adapt in real time to these changes and make sure that customers always are receiving the most accurate and up-to-date information? Yeah, so that's part of how we built Apted from day one is a technology that's specific to the domain information of a business and a team and a technology that's continuously evolving and understanding the context of answers that are being created, answers that are helpful when answers are not helpful learning from that interaction. And so in our case, you know when a technology like chat GPT for the consumer and the wide web is different because it's kind of a consumer general purpose solution for us it's a very domain specific experience so our AI and our technology actually goes and finds where the context and the answer is to a customer question we use generative technologies to be able to provide the delivery layer of this experience so we use generative technologies like a chat GPT technologies to be able to create the customer facing response we're not leaning on it to find the information or find the answer and that's what allows our technology our AI tech to continue to get better and better based within a customer environment the more the team our customers use it. Yeah we wrote a post the other day my colleagues and I came up with the power laws right came up with the power law the 80-20 rule essentially we had all the you know the consumer and the big cloud, LLMs you know the left side and then we had the the X axis was sort of long tail and we called it domain specificity so it sounds like that's really where you intend to play a lot many more large language models playing out there that are very specific to a domain very tuned to that domain versus sort of the generalized of the problem that we're seeing potentially with chat GPT is that a fair sort of mental description? Yeah and I would even add to that you know not only are we specific to the domain we're also specific to the domain of customer service and the types of questions and challenges customer service teams are seeing so are even more tailored and specific in that domain that allows us to create the most acceptable answers accurate answers in the first first time when a customer service team is using our solution to find an answer to a customer question. So Bobby you're kind of the conciliary here you got rapid advancements in AI and natural language processing landscapes changing all the time ahead and invested in the other day say yeah we get these term sheets they sign the term sheets by the time we turn around somebody else is disrupting here the business model so how do you plan to stay ahead of the curve what advice do you give to your CEO here to maintain relevance and leadership in this domain? Great question and you know what? You've heard me say this and I'll state it consistently is that the customer is always right because they have the pocketbook it doesn't mean they're always right technically and so if you look at the way in which we're engaging our early customers and you ask a question I want to come back to in terms of the amount of money when you have founder market fit you don't need it as much money to go fishing in the ocean to see if somebody likes your product and what benefits they might have you take and you show that I can acquire a customer in 35 days get you up and running in less than 24 hours by the way we haven't even talked about the system integrators at Salesforce we change out landscape and everyone from Accenture to people you didn't know had to re-adop business models we will do the same thing here because I don't need you to screw and glue Dave I know your age so if you remember the days that the tipcos and the integration layers became the most important part and then when they got the understanding of what the data was and what it was doing that's essentially what we're doing with knowledge which is far more powerful than data data's just the gold dust knowledge is the jewelry we can wear and take advantage of in the shiny, shiny valuable pieces. So you'll see a drumbeat of success focus on ecosystems and then I'm very eager to see how the ecosystem of deployment gets up and running because I would like maybe a crit to say within 24 hours you're seeing value and sometimes when we get a prospect on the phone we've actually scraped their site and I let you ask me a question and so the 30 day trial doesn't even work anymore because I can have a trial up and running in 30 seconds 30 minutes probably. Anything you want to share on that a crit because seeing is believing showcasing those testimonials and the proof is in the pudding what more do you need to the customers to be foolish not to move forward? Yeah, you know one thing I'll add to that Bobby sharing was very top of mind for us from day one was how do we get to fast time to value? Building a solution technology that's low lift low touch easy time to value and that is the ethos of how we built the technology from day one. So for us to connect into even like four or five different systems of knowledge or information we can do that in a few minutes and be up and running within a day where we're going to a motion there where this gets to be within a few minutes not just a few hours and getting people to real time harness the value of their collective domain information in a way that is scalable to create answers get to knowledge that ultimately today is going to drive faster higher quality customer service experiences. You know, I got to ask you guys so I was looking reading the other day that leaked Google memo even Google says we have no moats and neither does open AI and I was listening to a friend of the cube Jerry Chen who's been on many times and he was talking about moats in this new world and yeah, there's maybe some changes but his point was, you know what? The old moats still matter. It's time to market. It's your go to market prowess. It's your brand. It's your customer relationships. How do you guys think about moats in this new world? What's different and what's the same? Yeah, I can speak to it more from the product technology lens that we're learning. So, you know, one of the things with Apt Edge that allows our system to get better over time is we have workflow automation capabilities built into the product where if you're seeing a recurring customer service problem, you can actually use our system to create the ability to cluster common support problems and then be able to take repeatable action workflows on it and that allows our system to get smarter where if new issues are coming in we know that this is the most consistent repeatable answer for this customer, right? But in a way where it can still be personalized to that specific customer and their response and their experience. And so we have underlying areas in our technology where we're building the data mode by understanding the relationships between the data across these systems of knowledge with respect to customer questions and service issues, being able to understand which cases and support problems are related, which knowledge is often linked to that information and related that may be outside as help desk system and being able to allow our system to get smarter and smarter for a domain specific company based on the answers that are being provided from the reps using Aptage day to day. Yeah, Bobby, we're way over the time I said it was gonna take us so interesting but I'll give you the penultimate word. Oh, moat. So if you again know anything about my love moat builders I think one of the biggest best moats we've seen that this industry happens to be the app exchange that we build at Salesforce and everyone would have believed that the moat was the delivery mechanism. And I wanna be very, very, very, very articulate in those words. All GPT is a delivery mechanism just like the cloud is a delivery mechanism. It doesn't fundamentally do anything except how we consume it. So we have to remember that it's like an exit ramp to a highway. It just gets us there faster, easily digestible. I believe our moats are gonna hinge on two things, knowledge and answers. Look, our kids do this today. I don't wanna go ask a person to get a question to get a good question to get, I want the answer, I want it now. That's where we're going into. So our ability to deliver answers to your knowledge today to project how you can grow your business tomorrow will be the moat that matters. So time to value, time to install. And more importantly, that CSAT score we spoke of earlier, let's see if you can get to 100%. All right, thank you. And last question for you. So where do I get more information? I know I can go to aptedge.io. Anything else you wanna share with us? Yeah, I mean, we're constantly investing in more material. You know, a lot of times we find everyone's excited about how they can apply generative technologies to their use case. And we have a very practical use case for applying this to the frontline reps that deal with frustrated customers day in and day out. How do we empower their experiences? So we have knowledge that we're creating in our own blog content. We've got information we're publishing on social. We have videos we're creating with all of our customers, video testimonials of value. We're launching a new series. Bobby can speak a bit more to it, but an education series. So we're continuously looking to allow more of the industry to understand, hey, this technology is here to stay, but it's here to be used in a very practical way for certain use cases that allow your team to be better, allow your people to be better, faster, more efficient. And it's a channel that I saw in my early career happen in the world of automation and manufacturing. Manufacturing and automation tools led to more efficiencies for people. That was in a kind of a different environment. Now we're seeing that same pattern happen in the world of customer service. I think generative AI and AI technology is really making every single person in a company, you know, five to 10 X more efficient and capable that ultimately leads to better customer experiences. So we'll be publishing a lot more content around the excitement we have in the space and the value where we're going. And I actually do have one more question. You guys, where are you in money? Are you raising money? Are you good? Yeah, we recently raised a decent size round of funding from a few investors. We got preempted interest. So we took that as an opportunity for us to really double down, triple down, move faster and bringing the value of what we built to the rest of the market. We're not fundraising actively at the moment. We're going to look to come for focus on that at some point in the future as we keep growing but we've just kind of beaten goals on our milestones for quarter. So we're just executing and what goes in customers. Yeah, good. If you don't have to raise money right now, that's a good thing, especially for a company who started around when you did and a lot of companies are, you know they talk about the zombie corns, you're not one. So guys, thanks so much for spending some time with theCUBE and best of luck to you. Thanks for having us. We look forward to seeing you around in the follow-up series. Thank you so much. Okay, and thank you for watching. This is Dave Vellante for theCUBE Conversations. We'll see you next time.