Back

Webinar: Conversational AI - Transforming Your Contact Center from IVR to AI Voice Agents

transcript

All right, we'll go ahead and get started. Good morning and good afternoon, everybody, and welcome to our webinar with Tenyx. Thank you for joining us. We're excited that you're here. It's going to be an action-packed webinar with live demos and some really deep dives in some different features and really understanding how this new generation of technology can transform your existing contact center into a more conversational AI-based solution. just letting you know this webinar is being recorded so if you have some colleagues you want to share this with or you want to watch it later or review the demo you're able to and also you know feel free to ask questions throughout the webinar we'll answer them at the end and we'll be happy to follow up with you with any questions you might want to have offline And so with me today, I have Angie. Angie, say hello. Hi, I'm Angie. David, thanks so much for the intro. I've been with Tenyx for nearly two years and working in voice for just about 10 now. Tenyx is my third voice AI company. It's really exciting. And I am really stoked to be here with everyone today. Thanks so much. Yay. And I am David. I am the head of product at Tenyx. I've been working, building and selling automation technology for roughly the last 10 years, most notably at Salesforce for a few years, automating customer experiences with chatbots. Really excited about this new technology and all the great things we're going to show you today on voice. Just a little bit about Tenyx. We're an enterprise conversation AI company based in Palo Alto. We've been around for two years. And our founding team is very technical. They actually developed the solution that is now being used by McDonald's, automating millions of orders through drive-thrus. So as you can imagine, the most complex environment with one of the most legacy backend systems with a massive scale and complexity. So they understand, we understand these problems that customers are facing. And we're really excited to use this new technology to be solving problems. We have customers in multiple verticals and multiple domains. We're really excited to showcase some of those today. We're going to start off today's journey with Angie telling us a little bit about today's reality. Thanks, David. Yeah, so recently I needed to transfer some money and I needed to do it over the phone. And so I called my bank and I run into this IVR system, this interactive voice assistant, voice response system. and the IVR is like, oh, okay, how can I best help you? And I thought to myself, okay, well, I need to transfer some money. So I said, transfer. And the IVR says, oh, no, no, like in order to be able to transfer you, I need to know how best I can help you. I'm like, well, I need help with a transfer. Like it was so frustrating. And I just thought to myself, okay, like this has to be able to be better. What can we do to make this better? And I'm finding the more I need to call these different companies, I'm myself running into this poor experience and I'm becoming so frustrated, whether it be by the long wait times, by the repeating information, by the out of domain information. I would understand if, you know, transfer in most situations prompted the voice bot to talk about this, this like, how can I better assist you? But in the case of a bank specifically, it's like, well, banks do a lot of transfers. Why? Why can I not be helped in this way? and I'm finding this to be true for myself in lots of different situations. We're also having running into problems where existing solutions are failing to build trust in other ways. So, for example, as David mentioned, part of this team helped to enable, helped to voice enable McDonald's drive-thrus. I was part of that team as well and I've also worked on other quick service restaurant voice teams. And I was finding that when the drive-through bot was telling people that something was out of stock, for example, oh, you want a cheeseburger and we're out of cheese, so we can give you a regular burger, for example, people weren't actually believing the bot. They were like, oh, no, no, no, the bot must not know. Let me get escalated to a human. And then the human would be like, yeah, sorry that we don't have any cheese right now or whatever the situation was. But we're finding that we really need to be able to build trust with our customer base in order to be able to help solve these problems. And so some of the things that we at Tenyx are trying to do to help build trust is cover more scope, be more specific when in domain. So for example, if you call maybe a restaurant, transfer maybe would direct you to a human, whereas transferring the bank domain will maybe get you into like the bank transferring money mode. We're trying to do more active listening to really understand people and their problems and the questions that they might have. And I think almost most importantly, we're really trying to be able to recover and repair the conversations when things don't go perfectly. If I'm talking to David and David misunderstands me, I just don't hang up on him. I ask him to clarify. And having really building that trust and that repair in the conversations is something that Tenyx is really striving to do. And we're striving to do this in incremental stages so that we don't just kind of change everything all at once, but we're really striving to really build and foster that connection with people. That's great. I mean, there's so many new capabilities with this technology that we're using. And these three pillars of speed, accuracy, and reliability are a bunch of capabilities that are going to be required to kind of far exceed the previous solutions capabilities. We know that speed is incredibly important when building a solution with your customers. And there's a lot of different facets of this, which is, you know, have you ever tried talking with Alexa or Siri? Or have you ever tried using like chat TPT with a voice integration? It's just really, really slow and you can't interrupt it. And like no customer experience wants to be waiting, like a customer wants to be talking and then wait a few seconds after every turn. And what we're learning from our customers is that there's a higher correlation of drop-off of people just hanging up when the solution is really slow and they get frustrated and they lose that trust. The other part is it needs to be intelligent speed. So, you know, we also learn as we're reducing latency, reducing latency in our solution, it almost felt too aggressive. Like I would say something and then it would immediately like respond to me. And so we develop capabilities to predict when the customers finish speaking. So we have intelligent speed so that we wait when I'm not finished speaking. And then we hurry up and know when I've completed my sentence to respond appropriately. And that's created a really great conversational experience. Accuracy is incredibly important. I mean, a lot of us, anybody who's been in this space for conversation knows that entity capture and capturing phone numbers and emails and dates, that's very, very important to get that accurate. It's also very important to be able to understand what your customers are asking about, whether it be, I want to reset my password or I want to look at my order status. Those are really common types of accuracy and solutions, and we're all familiar with those. Now with language models, it's also important to be able to answer questions appropriately and make sure your agent is saying things that are correct. And so there's more and more models being used in these solutions. And some of the questions you should ask yourself is, is the solution I'm using going to be able to get better? If there are failures in the solution, how is the solution getting better? And who's responsible for that? Do I have a team? If I'm a customer, do I have a team of AI experts who can retrain models and find the failures and retrain these classifiers? or would I want to partner with somebody like Tenyx who has expertise in these areas to continuously improve these solutions and make them better? And that's why accuracy is so important is that you want all these systems working with a very high degree of accuracy. And then finally, reliability. It's really simple. It's about being trustworthy and consistent. There's a number of solutions that are popping up in the market where people are hooking up a prompt and some ASR and some TTS, and they're expecting it to just work magically. I'll tell you, I don't know if you chat GPT, but you type something in, it's pretty unlikely. With all the variations, you're going to get consistent responses and outcomes. And you're going to be able to navigate a complex process like taking an order, looking up personal information, making a reservation. Because these solutions lack the outer loop systems that provide the safety and guardrails and provide predictability in the system behavior. over time and be able to test it and make it robust. So we're really excited to be able to show you how this works and show you why Tenyx has all these capabilities. So with that, we're gonna focus in now on a particular industry. We're gonna look at the hospitality and travel industry. So one of our customers, we're gonna show you two demos and we're gonna dig into this online travel agency that we're working with, who's a customer. And as we approach them, they're very excited about using AI to solve some of their customer problems. Like many customers, they have too few call center agents, they're really expensive, they wanna provide a great customer experience. And what this customer is doing is they're taking reservations over the phone as an online aggregator. So they have hundreds of thousands of properties they make bookings for, and they wanna make these bookings over the phone. And so what they found is that based on variations in season and time of day, They just weren't able to have enough predictability in the capacity for their team and provided long wait times, which resulted in lost revenue. And also they had a large number of people calling about questions about the properties, which makes a lot of sense. And today's solution, they're just using an IVR, which is press one, if you want to learn about a property. And then it says, call the hotel, or it says press two to make a reservation. Then you talk to an agent. And so what's happening is their customers are not really getting helped in the way that they should be. And that's where Tenyx was able to come in and help. So the big part of this, the big lift in this solution was being able to integrate into over 250,000 documents with all the property information, the amenities, the location. It's amazing how we're doing this. And Angie's going to show you in a second how we're able to look up the correct hotel, understand when we don't know what the correct hotel is, capture that information, ask the customer, you're checking in checkout, and then we're handing it off to the human agent to finalize the booking. And so what it's doing is two things. It's helping the customers who aren't necessarily interested in making a reservation by answering their question, by providing a great customer experience and essentially qualifying them out. And then for those who are interested in making a reservation, it's improving the efficiency of the agent so they're able to help more customers. And that is such a powerful thing to show. So with that, I'm going to turn it over to Angie to give a live demo of the solution. Thanks. And yeah, as David said, I just really want to call out that this is in fact a live demo. So we'll see how it goes. Thank you for calling. How may I assist you in finding the ideal hotel today? Hi, I was looking to book a holiday inn in Kansas City, please. Sure, I can help you with that. Could you please provide me with the address of the Holiday Inn in Kansas City that you would like to stay at? You know, I'm not sure. I think it's the one on Admiral Boulevard. Sure, I can look that up for you. Could you please let me know your preferred check-in and check-out dates? Yeah, I was hoping to stay from Earth Day for three nights. So checking in on April 22nd and checking out on April 25th, and how many adults will be staying in the room? Just me and my husband. Excellent. What a fantastic vacation plan. Just to confirm before I transfer you, you're looking to book a stay at the Holiday Inn in Kansas City for two, checking in on April 22nd and out on April 25th. I'm about to connect you with a travel agent who will ensure you receive the best rate available for your trip. Is there anything else you need or other question I can help with right now? Oh, I'm sorry, I forgot. Does that location have a pool? Yes, the Holiday Inn in Kansas City on Admiral Boulevard does have a pool. It's an outdoor pool that is available 24 hours a day, and there is a towel service at the pool. Great. Thank you. So that's an example of our currently deployed solution. As you can see, I talked about multiple things in this demo. So as a starting point, the agent asked me what hotel I was looking to book. And I said something to the effect of, oh, I'm looking to book a holiday inn in Kansas City. So this is providing some unprompted data. I volunteered the city myself in addition to the hotel name. But then the agent said to me, you know, like there are a lot of holiday inns in Kansas City. I'm not sure which one you want. Let me try and disambiguate that for you. And so I said, oh, OK, I think it's the one on Admiral. But, you know, lucky for me, there was only one Holiday Inn on Admiral. It's not like a Starbucks, maybe. But fortunately, like that was enough information to be able to disambiguate. And I have to say, as one of the natural language processing engineers on this project, building out that disambiguation was both challenging but also quite a fun project. And I'm really proud that we've been able to disambiguate across the 250,000 hotels, as David mentioned earlier. Another thing I'd like to call out from this is the date capture. You may have noticed I said on Earth Day and for three nights. So I didn't really give a specific check-in date or a check-out date. The agent was able to infer, oh, hey, I know Earth Day is April 22nd. Three nights from April 22nd is April 25th. And at the end of the call, the agent was able to summarize all of this. The agent said, oh, hey, you want to stay on Admiral Boulevard? You want to stay from April 22nd to April 25th? And again, when the agent asked, hey, how many adults are in the room? Rather than providing a number, two or zero or what have you, I said, oh, you know, just me and my husband. To which the agent was able to infer, oh, you plus one other person is two. And so it was really wonderful to be able to use these maybe nuanced or natural language-like expressions in order to be able to have the agent really understand what I was trying to say. And I didn't have to modify my speech or talk to it like I was talking to a robot, like my previous bank example, when I couldn't get it to transfer for banking reasons for the life of me, things like that. So that was really wonderful. next up David's going to be demoing a different solution still in the hotel space but he's going to be touching on some other capabilities that we're able to support yeah this is this is really great so as Angie showed this was sort of a agent assist type example where we're sort of augmenting augmenting some of the agent's time by collecting information and passing that on to the agent In this demo, we were working with a single property, an IHG property, and they wanted to use this service on their front desk phone line so that they were able to answer questions about a specific property and make reservations with that direct property. So what I'm going to go through is a portion of the call, making reservations a little bit lengthy, collects a lot of information. But I'm going to highlight a number of other features, including some interruptions and some of that intelligent speed that we're talking about around waiting long enough, and then also some additional knowledge-based features. So with that, we're going to get going on this demo. Call is being recorded. Thank you for calling the San Francisco Holiday Inn Express. My name is Penny. How may I help you today? Hey there, how are you doing? I'm doing great. Thanks for asking. How can I assist you today? I want to make a reservation from June 8th through the 11th. From June 8th to June 11th, we have a queen room. You know what, I want to actually check out on June 15th. From June 8th to June 15th, we have a queen room, a king room, and a suite with a spa bath available. Would you like to make a reservation for one of those rooms? Yeah, I'll take the suite. How many adults will be staying in the room? Me and my wife. And how many children will be staying in the room? Six. I apologize, but the number of children must be between zero and two. All right, we'll leave them at home. No kids. I see. Could I please have your first and last name for the reservation? Sure, it's David Clark. Thanks. Do you have an IHC reward number? Um, I don't, but what's the distance from the hotel to the nearest airport? The hotel is approximately a 30-minute drive from the San Francisco airport. And the best way to get to the hotel from the airport is by ride-sharing apps or taxi use. And does the hotel have free Wi-Fi? Yes, there is free Wi-Fi available for all guests throughout the hotel. All right. Sounds great. So I was just going to give you a little sample of some of the capabilities, but I wanted to highlight that again, just changing your mind. you saw how some of the phrases it waited for me to finish and other ones, it was just like lightning fast. So I'm showing a lot of capabilities there. I know Andrew's going to go deeper into it as we talk about our voice platform and some of those knowledge capabilities. Yeah. Like David said, simulating the end pointing. So the longer duration of time when he wasn't finished speaking versus maybe replying very quickly when he was, I want to call out, he said something like, oh, I want to check out from June 11th until June 15th. And it's kind of like this long pause where in the beginning, we actually found that the agent kept interrupting us. And we're like, ah, agent, stop talking. Like, clearly, I have not finished my thought. And so, again, another really cool engineering solution was to be able to figure out a way to really have the agent make longer duration pauses when it was less likely that someone was finished speaking and to be able to respond quickly or promptly when we anticipate the person to have been finished with their sentence. Another thing that David demoed really flawlessly was his ability to interrupt the agent when he changed his mind about the checkout date. And both of these things are made capable by our voice platform. We originally started using open source voice platforms in order to really manage the simulation of turns. So the agent's turn speaking or myself, the customer turn speaking. And we found that we really weren't able to do all of the things that we really want to be able to do. And so our solution was to build an in-house solution. And that's actually been working quite flawlessly, quite well, not necessarily flawlessly, but quite well as David just demoed there before. Another thing I'd like to call out is the knowledge base. So David asked the agent a couple of questions. He asked about the distance to the hotel, a few other things. And the agent was able to go and pull a document and look up that information and then repeat that information to David without an industry. The term is hallucinate without making up an answer. And so it was really nice to be able to see the agent going and pulling that information and repeating it. I want to call out that on some turns, such as when we look up this information, it might take a little bit longer. And that's because we want to really ensure that we're like maybe fact checking the data, if you will, rather than making up an answer or worse lying about something. And, you know, saying something that is factually inaccurate. And so it's, again, want to call out the fun engineering solution. So I'm really fun to be working on a project that covers so many different niche use cases and is able to execute on them with so much fluidity. That's great. And I'm sure many of you on this webinar are wondering where you can get started on how you might be thinking about adopting some automation in your call center. I know there's a lot of agent assist tools out there to make your agents more productive, but putting a solution like this on the front lines and taking a lot of these basic use cases is just a great way to get started. And so what we're finding is that the, what, what have normally been like the traditional IVR solutions where we're going to figure out who you need to talk to and like maybe answer some basic questions. That's really, really easy for us to do. Um, and so what we're finding with customers is that we can do, you know, a POC with them, take their knowledge base, um, apply some routing rules. We can, we can deflect a lot of calls and make the experience so much better in just a few weeks. And that's just a really, really great way to get started because, um, it's a bit unintrusive. It's not like you're trying to, you know, displace agents in your call center. It's not like you're, you're stepping on anybody's toes. It's just providing a lot of value and a great customer experience above and beyond. My experience this morning when I called Kaiser, which is my, my hospital for my daughter, where it says, you know, press one to talk to a nurse and get advice and, or say advice and press two to say, you know, it just, it feels like I'm always on this very narrow path and I'm not really, I don't really get to say the thing I need to say. And we don't have to have solutions like that anymore. So let's get rid of the phone trees, please. The second level on this journey is kind of what we saw in the first demo, which is expanding beyond that really base case and being able to think about how you can provide or collect information, excuse me, from your callers and then hand that off to agents to make them more productive. This is something agents love because, you know, they're asking the same three or four questions every time a reservations may, like, where do you want to book? What are your dates? How many people, how many rooms, what are you looking for? If we can collect all that in a really automated way, agents can just get into the fun part, which is, okay, let me help you select the right room or tell me a little bit more about like the type of property you want, you know, and they can like get into that fun conversational selling, which is still great. You know, that's, that's a great human experience to be having. And we don't have to automate that. If for your particular industry or use case, you feel like full automation is a great use case, like a password reset or looking up an order status, you know, those are all things we are capable of. We've, we've been there, we've done that, like the, the whole McDonald's journey. You know, my journey, you know, automating 2000 different customers with bots, you know, I understand that there are so many use cases you can fully automate and that, you know, whether it's HR, insurance, you know, they really are well suited to automation. One of my favorite examples of full automation is paying your bills. You know, if you have to call and you're late on a payment, whether it's, you know, maybe your car payment or your insurance, you call in and you're not necessarily comfortable talking to a human about being late on your bills, but if it's just a really conversational friendly AI solution, it's great. And what we actually found is that people are more comfortable with that than in humans. And so just finding those great use cases that are open for full automation. And then the last thing I'll say here is at Tenyx, we're here for wherever you are on your journey. We can do the really simple stuff, like a lot of people can. But then as Angie pointed out, we can really start working on the long tail of complexity and nuance because we are a team of technical experts. I think there's, I might be the only non-PhD in the company, you know, because ML is in our background and automation is in our background. And so we really embrace the hard problem and look forward to solving things like hotel disambiguation or hotel recommendations, things you'll just never find in another voice platform. And so with that, I want to kind of bring us home with this last one, which is just, you know, your next step with Tenyx, as I alluded to earlier, like just contact us and give us your use case. We're we'll do a POC within two weeks and we'll show you, we'll give you a phone number. You can call it. It'll be fast. And we'll tackle some of these, these small, smaller use cases that you can then internalize with your organization. You can get them to believe that this, this is a solution that's going to work. And then the part I love about this is just it's white glove. You're buying a solution. It's going to work. And you're working with experts who can continue to manage, manage it and improve it. So you don't have to stay up to date on the latest, you know, hugging face model or how to retrain or fine tune, we have that covered for you. In fact, we have several patents in that space, one particularly in model fine tuning. So we are very technical in that team. And so with that, contact us and let us know what your use case is. And with that, we're going to turn it over to some questions. So one question I have here in the chat is do you have any other voices than the one you showed in your demo? Angie, you want to take that one? ANGIE COOK- Yeah, I can answer that. Yeah, so we definitely do. So right now, we're using 11Labs as our voice provider. And I believe we're using the same voice in both of the demos that we showed today. But theoretically, we could use any other 11Labs voice or any other voice from a different provider. For example, we've been looking at PlayHT amongst some other vendors. And also, we can also do something like voice cloning if you have like a particular voice in-house staff member whose voice you wanted to use, or if you have a voice actor whose voice you wanted to use. We also just want to call out that the text-to-speech augmentation is like this is the worst it's going to be. It's getting better and better every day, or every week maybe more accurately. And so we're definitely able and willing to either use a vendor that you prefer or to swap the vendor for us if it makes more sense as well. No, that makes a lot of sense. Maybe this question, there's a lot of questions, so I'm just filtering through them. Maybe this is another one that's similar. What to look for in a voice provider? I can answer this just because I've evaluated a lot of voice providers. So there are several things that you want to look for in a voice, a good voice provider. One is definitely just the fact that it, you know, sounds on brand for you. It sounds natural. It sounds maybe humanistic or maybe more robotic if that's what you prefer. It's like really up to the customer in this particular case. One thing that I really strongly evaluate for is that it's able to say complex phrases efficiently and naturally. So, for example, a lot of the voice providers that I've analyzed sound really, really great, but then can't read back email addresses or sound really great, but like can't spell names. And you think, oh, that sounds like such a simple thing, like A-N-G-I-E, you know, kind of like let me spell your name back to you. Or I find it's more important when they can't spell the words in email addresses when you're like ensuring like, oh, did you get my email right? Or, oh, did you get my phone number right or my credit card number right or what have you, anything like that. And it sounds kind of like gibberish sometimes when the agents read back a weird mixture of non-English sounds or non-whatever language you're working in sounds. And so amongst other things, those are two really key things I think are important in voice providers. That's great. Here's another question here. I think I can take this one, which is, are language models safe? We're seeing mixed things in the media. Yeah. I mean, yes and no. I mean, depending on how you use language models, they can be very unsafe. The good news is when you're using a solution like Tenyx, we use open source models that are domain specific and fine tuned for solving very specific problems. So they have some of the powers and capabilities of these large language models like chat GPT, but we essentially train them to work just for your use case. And so we don't run into the problems that you would typically see if you were just using any language model off the shelf or sending your data to chat GPT. In addition to that, we develop safeguards essentially around our solution to prevent any types of hallucinations. And so there's a combination of language models being used with an outer loop system. So there's redundancy and safeguards in place to make sure that nothing is being hallucinated. And also, you know, we're also protected against, you know, racist things or really inappropriate things that you wouldn't want to have represented in your brand. Yeah. And just to kind of tag on to what David just said, in addition to that, I just want to call out that we do have our SOC 2 Type 1 compliance. And so don't worry that we're, you know, if you're in a regulated industry, we're not sending your private, your PII data to, you know, OpenAI via chat GPT or anything like that. We have those kinds of LLM safeguards in place as well. Yeah, that's right. Here's another one, a comparison to your solution versus Dialogflow and others. Boy, where to start on this one? I think the biggest thing to really focus on is that Tenyx is a voice-first solution. So everything we just showed you, whether it's our voice platform or the endpoint features or the speed in which our solution is delivered, it's all designed for voice. Dialogue flow is not. It's primarily a text-first solution, which means that everything that's going on around it is sort of stapled onto it. And so our theory on it and our mindset is that if you solve voice, you can kind of do everything else versus if you start with text, you can't. Because with text, you're always converting back into voice and there's just so much nuance to speech, latency, the interruptions, the waiting, the change in your mind. You don't have clean turns. So you got to maintain that context window. I mean, there's just so many things that really make voice unique. And that's why Tenyx is the best solution on the market for this. Multilingual stuff. Angie, do you want to take that one? Customers, I guess, wondering, you know, what are our multilingual capabilities? Yeah. To be determined is really the best way to describe that, I'd say. But But I will say that there are a couple of partners that we're currently working with that we are considering working on multi-language solutions for them. The other thing that's just kind of a fun call out but is also accurate is the first 10 TENIX staff members were all from different countries and all have different language backgrounds. And so it's like quite helpful on staff to just have a lot of in-house language varieties and knowledge of the nuances of different languages. Like for myself, for example, in addition to English, I speak German. And it's I think would be really fun to work on a German voice based solution. So we're definitely open to the opportunities. So just to add on to that, like our solution today is English primarily. And we know that the technology works with any language. we're just focusing right now on English and it's on our roadmap later this year to add additional languages based on customer needs. Thanks, David. Here's another great question. How do you integrate with backend systems and do you have any integrations with call center providers like Genesys? Yes. Yeah. I mean, we have to integrate with a bunch of different call center providers. The majority of them have great APIs, so we can hand off the call, we can hand off the context, we can do the transfers. So there's the part of it where you're integrating with the call center stack. So you're integrating with the agent experience. And then there's the part of it where you're integrating with other backend systems. And so there's sort of two different things. When you're integrating with backend systems, the most important thing is that there's APIs that are available that we can use and access so that we can get information in a timely manner and then use it as we showcase in our demo. And just kind of one of the things to add there as well is we actually are already doing these, I call them warm transfers. So in the demo that I provided right now, it's a partial automation solution. So we capture all this information and warm transfer it to the live agent. But I think a nice benefit of that is on the off chance or on the case where your customer needs to be escalated to a human prematurely, we can warm transfer the information that we've already collected. So if they've gotten halfway through the flow, but maybe not quite completed it, when they get transferred, they don't need to repeat everything they've already said. That's right. That's right. This last question, it looks like, can you tell me a little bit more about the journey from an IVR to this next-gen solution? The short answer is, yes, let's talk. I'd love to hear more about where you are at in your journey with implementing and managing an IVR. And obviously, there's parts of today's presentation that you can go back and sort of think through as you think about like, what are the right use cases? The reality is that every business is just a little bit different in the business dynamics and also the types of solutions that are going to move the needle for their business. And clearly just routing via an IVR is just not enough alone. So, you know, it could be answering questions and reducing call volume, or it might be finding a specific task within the IVR tree that we just took up our solution to, and then we're able to automate that flow. And that really moves the needle for businesses. So we look forward to talking more with you. Yeah. And just to kind of add on to that from what David just said, like he said, for us, we are building POCs in two weeks. Previous companies I've worked at, it takes around two years to build this kind of technology. And so it's really, really exciting to be able to be working on technology that moves this quickly, really allows us to integrate with truly next-gen technology. Well, thank you, everybody, for the great questions. We're looking forward to chatting with you more. Enjoy the rest of your day. See ya. Have a good one.

Leverage voice AI agents to increase your revenue, improve efficiency, and strengthen your brand.

Facing revenue challenges, staffing shortages, and outdated IVRs, businesses are grappling with how to deliver the next generation of voice experiences for their customers. The solution? Conversational AI. Watch our on-demand webinar where we delve into how Conversational AI is reshaping contact center operations, from lagging IVR systems to immersive customer experiences, including live demos showcasing the power of Tenyx’s voice AI that leverages our unique investment in foundational voice and AI technologies.

Meet the Experts:

Angie Howard (Staff NLP Engineer): Angie's career in voice artificial intelligence spans over ten years and includes supporting the development of over 22 languages and 25 partners at SoundHound. She also played a key role in voice automating McDonald’s drive-throughs at McD Tech Labs, later acquired by IBM. At Tenyx, she specializes in NLP/NLG algorithm development, conversation design, evaluations, and more.

David Clark (Head of Product): Boasting over a decade of experience in building and selling enterprise automation for Sales and Service, David's tenure at Salesforce stands out. As a Product leader at Salesforce Service Cloud, Salesforce grew its Einstein Chatbot program from a startup to over 2000 customers.

Browse more

News
January 11, 2024

Generative AI Report

News
December 12, 2023

Tenyx aims to fix LLMs’ catastrophic forgetting problem

News
May 17, 2023

Driving innovation in auto finance with AI, transformer models, big data analytics