Episode Transcript
[00:00:03] Speaker A: Welcome to INA Insights, where prominent leaders and influencers shaping the industrial and industrial technology sector discuss topics that are critical for executives, boards and investors. INA Insights is brought to you by Ina AI, a firm focused on working with industrial companies to make them unrivaled. Segment of one leaders to learn more about Ina Aihdev, please visit our website at www. Dot Ina dot AI.
[00:00:40] Speaker B: Good morning everybody. Welcome to another episode of our INA Insights podcast. Today we will continue our focus on early stage AI companies.
Our guest is Mister Adam Earl, who is the chief technology officer of Tenix AI. Tenix AI represents an innovative artificial intelligence system created to transform the methods of engaging with technology.
It utilizes sophisticated machine learning algorithms and natural language processing methods, allowing it to comprehend, adapt and provide intelligent responses to user commands. Tenex has raised a total funding of $15 million to date, with the last funding raised in May 2022.
As per Adam, prior to Tenex he was working as a senior director in IBM and before that worked as the chief technology officer of MCD Tech Labs.
Adam has a PhD in computational and applied mathematics from the University of with Water Strand and completed the Stanford Business School executive education program.
Adam, a very warm welcome to our podcast. We are very excited to have you with us here today and look forward to talking about Tenex and your own journey as a CTO.
[00:02:04] Speaker C: Yeah, thank you. I appreciate the introduction. It's so good to be here with you.
[00:02:07] Speaker B: Perfect. So let's get started. Adam, why don't you tell us a bit about Ten X, especially why and how was the company started?
[00:02:16] Speaker C: Yeah, absolutely. So first and foremost, at Ten X, we deliver voice solutions for enterprise customers. So for listeners, you really want to think of the form factor as kind of just your cell phone customer picks up their phone calls, a number to talk to Walmart or Delta Airlines or any big enterprise, and as a full, naturalistic, high intelligence, engaging conversation just with an AI rather than a person. So that's always what we're going to be talking about. And really what got us started at Ten X is building these high fidelity, high intelligence, but very naturalistic sounding voice agents and voice assistants. So obviously, a huge part of what's gotten us started is the LLM revolution. Both my co founder and myself have worked in the machine learning space for many, many years, and so we were observers of the improvements in transformer based language models before kind of chat, GPT and those sorts of public technologies came around. But it's been clear for some time that those technologies would be a huge step, function, improvement and really empower this next wave of innovation. So again, both Edamai and I, that is, my co founder and CEO, have worked together for many years in the speech and natural language processing space, and we've seen the rise of these technologies and we really think now is that pivotal moment and the right time where we're going to see these step function changes in voice systems for enterprise applications really across the board. And it's an exciting time to be here and building those kinds of things out in the real world.
[00:03:48] Speaker B: Right, sounds exciting. Adam, can you tell us a little bit more about the end markets that this is applicable to? I mean, which are the industries or end markets that you are seeing most traction for these voice solutions?
[00:04:03] Speaker C: Yeah, great question. So first, just the technologists answer. I think we've been pleasantly surprised to see just how flexible and broadly applicable these solutions are. So we've come from many years of building NLP and voice based solutions, and really with machine learning tools in the previous generation, way back in 2021, really before large language models, providers would spend many years chasing down this long tail of problems and really having to build for one and only one vertical. Whereas today, with the flexibility of these technologies, we are far more adaptable and can really service many different domains. So that's the first point to land.
The second is, of course, as a smaller company, as a startup, you kind of really have to find your niche and pick a spot to start. And so where we've gotten started at Pennex is really in a couple of places. So hospitality, travel, reservations, that sort of space for all these, I think flight bookings, rental cars, hotels, also then in financial services and healthcare and insure tech space as well. And then as a third big pillar in sort of surveys and research, you can think of big sort of polling agencies and those sort of applications. So again, we have to get started somewhere, land and win from there. But the technologies really empower something far more broad, and that's really in the long term vision for ten x.
[00:05:30] Speaker B: Got it. And what are some of the use cases that you are observing in these industries that you mentioned?
[00:05:37] Speaker C: Yeah, so sometimes we like to think about the taxonomy at the top, starting with inbound versus outbound use cases. So in the inbound case, it's sort of the canonical customer calls into the company, and then you're thinking of customer service use cases. So people who either need to make alterations, their reservations, or make bookings or get assistance with products that kind of canonical customer service use case. In the outbound use case, it's really the businesses reaching out to their customers this is always obviously consented and solicited, but people who need help with abandoned car problems or midflower need to be connected for some other reason. So really, I think in a lot of these industries, there's often both an inbound and an outbound use case, either for customer care or for reaching out to your customers for secondary sale or some other approach like that. So in all instances, I would say there's at least these two halves of the use case.
[00:06:38] Speaker B: Adam, since we focus a lot on industrial sector at InA, just curious if you have seen any traction from that sector or any use cases that some of those companies have reached out to you for.
[00:06:53] Speaker C: Yes, we haven't spent a ton of time in the industrial space thus far, but really just a function of the fact that we're a small company that has to start somewhere, if you like. So again, our focus specifically has been more in hospitality, fintech, and in the survey space, but projecting ahead, even just a little bit. One has to imagine that the same sort of customer service use cases and outbounding are similar there as well. So while my answer is not yet, I think there's high probability that we would see application in those sectors as well.
[00:07:24] Speaker B: Right? That sounds about right, because we hear that a lot of these industrial companies are now interested in using technology, especially for their customer service areas.
[00:07:35] Speaker C: That makes a ton of sense. Send them our way.
[00:07:38] Speaker B: Will do, will do. Awesome. So Tenex till date has raised $15 million. Adam, so can you tell us a little bit more about how's that played out for ten x so far?
[00:07:53] Speaker C: Yeah, of course. So, as you kind of mentioned in the introduction, our seed was somewhat on the larger side at 15, and that was in early 22. And a lot of those investors are investors that we, myself and my co founder have worked with before. So a few too many to name all. But like Army Cloud Ventures, Coda Capital, Murado 72, and then some individual angels as well in Jan Talen, Zach Colias and so on. It's really a great investor group and a group that we've worked with before. I think what all of these investors share and the reason we work together is because they all have kind of a shared vision for the long term potential and application of this kind of technology. So we all think that while there's obviously a lot of excitement around Gen AI and automation today, we see a very clear view of the future, where in the not so distant future, this idea of kind of being on hold as you wait for a customer service representative will just be anachronistic it shouldn't be present in our day to day vocabulary, and that is for sure where we're going to end up. So again, we've kind of established investors that we've worked with before, but also those who see this clear path towards really disruptive innovation in the customer service space across multiple sectors and multiple domains. So that's, yeah, that's sort of what's gotten us started, how we found those specific investors and what keeps us working. Well today.
[00:09:18] Speaker B: Sounds exciting. Adam would like to switch gears now a little bit here and talk about your own journey to date. Tell us a bit more about your background and your professional life before ten x.
[00:09:31] Speaker C: Of course, I somewhat tongue in cheek like to start by just saying the accent you're hearing is actually from South Africa. So I get a lot of Australia and I get british occasionally, and every day a little bit more american, but really it's emergently from South Africa and that's where my story starts. So I did my graduate studies there, as you described, at the University of the Witwatersrand, where I did, I came up through physics and applied math. I did my master's in operations research and ultimately a PhD, applied math and machine learning with a focus on reinforcement learning. So I really came up through math and machine learning and then joined a startup in South Africa, a smaller company called All Life. And all life was focusing on building life insurance products, but really specifically for people who were HIV positive. And so some really fantastic high social impact work. And I got to do a lot of exciting work on the risk retail pricing and work with some large reinsurers as we were trying to build out a product for HIV positive people in South Africa. That technology led to some really interesting health management protocols and risk management protocols, and olife ultimately then extended that product offering into diabetes. And I got to work in some of the international expansion of those life products protocols into diabetes in a bunch of other countries in Spain, Australia, Germany and so on. But that brings us to probably the last seven years or so where I've been in the Bay Area working on machine learning, working on speech and natural language processing, and that's really taken me through a pretty interesting journey. So the start of that journey is really with a startup apprentice, which is where I started working with Itamark. This was back in 2018. We were building voice technologies for enterprise applications there as well, but with a specific focus on quick service restaurants, McDonald's, Yum brands, stockbucks and so on. And that ultimately resulted in a subsequent acquisition by McDonald's and then while at McDonald's, I sort of led the core technology and machine learning teams and played a very brief role on our product team as well. And then that entire 100 person unit was subsequently reacquired by IBM, and I served there for about a quarter to two to facilitate that acquisition and transition.
But really in my blood, in my DNA startups, is what I love doing, small, impactful teams. And that occasioned us to get started with Ten X again. So really a journey through some insurance experts back in South Africa, and then machine learning and NLP here in the bay through startups and a few larger companies through acquisition in between, but again, really in our DNA is to do big, impactful things with small yet mighty teams. And that's what we're doing here at Tenex.
[00:12:26] Speaker B: Wow, that's a very interesting journey. Adam, what attracted you about this voice solutions market? Of course, you're applying the same core principles of the underlying technology, but it's a pretty big shift that you've made over the years from medtech to, to what you're doing right now at ten X. Right.
[00:12:47] Speaker C: It is a shift, and I think I probably start with the word sort of impact. We talk a lot about the fact that voice is just the most natural way for us to communicate. I think we've seen probably a decade of companies pushing consumers into chat applications, onto websites because the labor cost is just really high. But voice is still the way we communicate. When you're sitting at home and you're talking to a spouse, or to your kids, or to your friends, voice is how we do it. And I think it's just been the case that the technology hasn't been good enough. And so when you work with voice in the past, it really means outdated IVR, you know, please say claims. Did you say claims? And we all know what that experience is like. But on the flip side, when we engage in our personal lives, voice is how we do it. And so it's felt for some time like the technology pieces were really slowly coming together so that we could empower really natural, really high intelligence voice experiences. And once that happened, it would just proliferate very broadly, very, very quickly. And that's the vision that we've been building to even before ten X, frankly. And so impact release is the keyword. The second thing I would just say very quickly is even back in the insure tech space with all life, we had to build our call centers. And something that was really interesting, there was the fact that our customers were sometimes much more comfortable talking to a call center agent about some of their disease and health implications, rather than the traditional way insurance was sold was having a broker come and sit in your living room. And so they need the first taste, a little bit of the ease and comfort, that kind of, a little bit of separation or a little bit of AI and automation can provide for many, many use cases. So I think there's going to be, again, a wave of adoption. I think it's likely to be much better for consumers, of course, but also alleviating for a lot of those industries. So a big shift indeed, but an exciting one, and one that ties together maybe a little bit more than it seems on the surface.
[00:15:06] Speaker B: Right, Adam, this also provides a good segue to the next section where I want to focus a little bit more about the impact that these voice technologies are having on the different industries. So from your own perspective and what you've observed at Tenex while serving, let's say, the hospitality industry or the financial industry, how are some of these technologies shaping the problems that you're trying to solve for your customers?
[00:15:36] Speaker C: Yeah, a broad question, but a really important one. So I'm going to try to cover a couple key threads, but happy to go deeper wherever you'd like to take us next. But the first thing I think I would say is just maybe help your listeners to understand the scale of what we're talking about here. So we often talk about, just to pick a round number, the call centers in the US alone being about 100 billion a year, and the vast majority of that going into labor costs. So at and t alone is something like a billion dollars a year in their call center. Almost every big company has some need of a call center because every big company has some need to service their customers or their clients. And so when we talk about the way that these things are coming in, it's really useful to keep in mind just how broad sort of a playing field we have in front of us. I think the second thing that's happened, and this will lead to more specific answer next. But the second thing that's happened is we've had 20 years of automation here already. It's just been bad. It's been ivrs, limited call trees, and we've really been trying to cost cut ourselves into the most cost efficient, hollow experience for customers, rather than an experience that kind of delights because it's just really expensive. And so that's, I think, the first thing that we're seeing with the adoption of some of this new AI technology, one enterprises are considering a shift from existing IVR solutions to this next generation of much more fluid language model based natural interaction. There's a kind of a shift of that budget allocation from old ivrs that were cost cutting and consumer frustrating into this next generation voice experience, which is consumer delighting and much more fluid and much more likely to solve.
I think the second thing that's happening as well is there's been kind of a push for a long time to offshore at the lowest possible cost these kinds of roles. And even if you are a lot of this work into the Philippines or wherever else, we're seeing kind of labor costs rising there significantly as well. $1012 an hour is pretty common even for offshore solutions. And I think the reality is just coming to the fore, that automation is now ready and timely to start playing a significant part either for replacement or even just for enhancement and support of the human offering to make sure that calls that go to people are high impact calls or those that really need the human touch, rather than keeping consumers sitting on hold for a long time. So I think those are, those are sort of backbones. But again, it would anchor us as well in sort of an inbound versus outline dynamic. We're seeing some customers wanting to pick up outbound flows where they have lots of historic leads or need to connect with their customers, and some in particular, those looking to scale picking up inbound needs, because the alternative is either long wait times or dramatically scaling up headcount in these other areas. So hopefully that gives you a good overview for that. But a big question, and I'm happy to take it further in any specific direction you'd like to from there, right?
[00:19:00] Speaker B: Absolutely. So, Adam, especially given the magnitude of the numbers that we are talking about here with the labor cost area. Right.
It's an important area for companies, enterprises to focus on. So what advice or recommendations would you offer for companies who are looking to adopt the technology and successfully deploy it in that case?
[00:19:28] Speaker C: Yeah, I think the first thing I would say is, even though we see kind of gen AI and chat GPT everywhere sometimes joke that sort of, when your grandmother asks you about it, it's gone a little far. Even though we see it in the news so often, I think it's still somewhat poorly understood technology. I think there's still very real concerns about what it can and can't do, where I should and shouldn't be using it, which models, how do I get the most out of it? And I think the first piece of advice that I would kind of give enterprises looking for adoption is just to kind of take those questions head on. And when you're looking at providers or you're looking at potential solutions, to bring those questions to those providers and just to have them answered very frankly, I think the right way is not to kind of pretend that those aren't real concerns or really important, but to make sure that the teams that you're working with have thought deeply about those problems and have kind of intelligent and grounded solutions to those things. All that is to say is that while chat GBT can be a very cute demo where and build simple applications, the requirements for enterprise are much higher. They're real. And whomever you end up working with should have very good answers to how they keep your solutions safe. Guardrails or on script or whatever your preferred language is. Just how do I know what I'm getting? How do I make improvements? How does your technology support that? So that's sort of the first thing I would say very pragmatically, as well as sometimes the saying, how do you eat an elephant one byte at a time? And I think especially for big enterprises, it can sometimes be quite daunting to think about overhauling every IVR they have or every place in the automation space. And I think there's a more gradual approach that can still be incredibly powerful. And so where we found a lot of fantastic early success is with companies that were sometimes thinking of automation in the space already, but had carved out a high impact use case that was sort of a subset of their whole business as a way to get started. And as they built familiarity with the team, the tools gotten a felt sense of the safety and flexibility of these tools, then they've found it much easier to kind of rally that adoption internally. So also, second piece of advice would be to start with a smaller but high impact use case and roll forward from there. Yeah. And then finally, I guess, just try not to get caught up with the hype. There's nothing like a real solution or real demo in your hands to tell the story. That whole demos worth a thousand words sort of analogy. And so I would say to people, just make sure you get a demo in your hands as quickly as you can. And if seeing is believing, when that works well, you'll have some comfort beyond the headlines that there's something real that you can work with. So a few small pieces of advice for companies looking to get started with, with voice automation.
[00:22:29] Speaker B: Perfect. And Adam, that kind of also takes me to my last question. Since the world cannot stop talking about Gen Ei, what does it mean for ten x? Are there any new areas that you are looking to explore as a company in the future.
[00:22:48] Speaker C: Yeah, I love the way you said it. The world can't stop talking about it. Yeah, I think there's two important things. The one is the world is changing very rapidly. And so any company that's not set up to rise with the tides and continue to incorporate the state of the art, continue to take the best practices and the new available tooling, I think is likely to get lost very, very quickly. And so the way that we've set up is precisely to ensure that we continue to rise with the tide, if you like, and continue. We are architected in a way that lets us interplay these components as they continue to improve and develop. So again, it's another important thing, I think, to ask of your providers when you get there. But that's sort of the first thing for us is make sure that we are not one way decisions lock doors into situations that don't let us rise with the tide. I think the second thing is, if you visit our website, you'll find that we do quite a lot of research as well. We have a small applied research team that made contributions to open source models published at large ML conferences as well, focusing on the outstanding problems we still think are there around domain alignment, around safety, around fine tuning these large models with enterprise data. These are just some of the problems that we think are still present and we think they're important for large enterprise adoption. So again, domain alignment, safety and trust, and then actually using enterprise data through fine tuning in a way that really works. So folks are interested in what's next for us. I encourage them to check out our research page on the website as well, just to get a flavor of what we're working on beyond kind of just the voice solution as is right now.
[00:24:34] Speaker B: Awesome. Thank you so much, Adam. This was really insightful in terms of what Ten X is building on the voice technology side and also how the company is thinking about Genai going forward.
[00:24:48] Speaker C: Fantastic. Well, so appreciate the time today. It's been a pleasure chatting with you as well. And we love questions, so always happy to hear from folks and happy to help in any way that we can. So again, I really appreciate the time today. It's been great chatting.
[00:25:00] Speaker B: Thank you, Adam.
[00:25:07] Speaker A: Thanks for listening to INA insights. Please visit Ina AI for more podcasts, publications and events on developments shaping the industrial and industrial technology sector.
