Episode Transcript
[00:00:22] Speaker A: Welcome, everybody, to our latest Metrocyte episode. I'm Robin Garris. I'm CEO and principal analyst of metrogy. And today I am absolutely delighted to be joined by Jonathan Rosenberg, who is the CTO and head of AI at five nine. So welcome, Jonathan.
[00:00:37] Speaker B: Thank you, Robin. Glad to be here.
[00:00:40] Speaker A: All right, so I want to start with just a little bit of background on Jonathan. For those of you who may not know him. I don't know if there's anyone out there but 90 patents. Co inventor of SIP, CTO of Cisco's collaboration business, and also chief technologist for Skype back in the day. You know, PhD in electrical engineering from Columbia, B's in electrical engineering from MIT.
Now, you lead five nine's overall technology and AI. So did I miss anything substantial?
[00:01:10] Speaker B: I think that's fine. It's all good.
[00:01:12] Speaker A: I know that. I know that about you. We've had some good restaurant discussions, yes.
But anyway, those are honestly some incredibly impressive accomplishments, as anyone can see. And honestly, some of the many reasons why I really, really always enjoyed talking to you. In fact, I think what our listeners and viewers will enjoy is your ability to take all the technology knowledge and understanding that you have that is obviously very complex and explain it in a way that everyone can understand. So I'd really like to just get started and jump right into it.
So, first, before we get deeper into the technology, I want to understand this, like, from a high level, about five nine. The company recently exceeded 1 billion. That's with a b in annual revenue run rate as of the last financial update. And you joined the company in 2019 when revenue was about 328 million. So I want you to talk a little bit about that near triple growth since you arrived. In your opinion, like, from a five nine technology standpoint, what do you think was most influential in driving that growth?
[00:02:17] Speaker B: Yeah. So lots of things have to come together for growth like that. It is not just technology. Right? It's sales, execution, services, support, marketing. Everything has to come together.
But if I look at the technology side and think about the answer to that question, I'd say it's a few things. First, it's that we've scaled the platform. One of the things that's happened in five nine trajectory is it started out its early days as like a small company dialer and it added inbound, but it was focused on the lower end of the market. And as time has progressed, it scaled up its platform. And so we made tons of investment. It'll be able to handle more and more agents, more and more complicated use cases.
And so that scaling of the platform has allowed us to address larger customers and grow our revenue as we move up market. That's probably the biggest thing. The second thing I would say is it's been all about this AI and innovation that we've poured into the platform and adding new capabilities to it that have allowed us to sell more, increase the ASps for adding on these additional products.
That's really helped drive a lot of this growth as well.
[00:03:22] Speaker A: Okay. I mean, that sounds reasonable. And I think it's everyone's goal, I'm sure at five nine to keep that growth and momentum going. And we see companies investing pretty wildly in CX and AI in general. I think it was like two or three years ago when the economy was very well, the economy's always up and down. But CX was the top area of investment for companies when we gave them a whole list of areas. And in our most recent research, the average of spending on AI technology for CX NEx alone is 3.3 million annually. And for obviously, larger companies are spending a lot more than that. Smaller companies are spending less. And this year, and going into 2025, more than 80% of companies plan to increase that spending. So that's huge. The appetite is there. It's going to continue to be strong for the near term. So I wanted to talk a little bit about your vision further out. And you and I have talked about a three pronged approach that you have of where you see customer experience heading in the next five to ten years. And I wanted you to give our audience a little bit of a high level overview of that and then we can drill down into some of the details.
[00:04:28] Speaker B: Yeah, so thanks for letting me talk about this. I just love talking about this. And let me. And so what's pillar one? Pillar one is I have this idea of sort of the CX industry, the CCAS industry, reclaiming terra firma as the place where we actually own the customer experience. Because I'm going to let you in on a little dirty secret. In our industry, we like to talk about how we are the CX for the business. We own all the customer interactions. In reality, a lot of that moved to the website, a lot of online that has nothing to do with the convex center. An example I'd like to give is airline bookings. Robin, maybe you remember these days when you wanted to book a flight, you literally called the Connex center, and you spoke to an agent and you gave them, and then you booked it. And that was it. Literally. The Connect center was the source of all transactional business unless outside of a retail store. Right. And then when Webb showed up, a lot of that left the Conex center and we were left with some. But this causes bifurcation of the customer experience, where you have this online experience, which is great for self service, by the way, but it's highly structured.
It lets you do just what you can do. If you want to do something you can't do, there you go to the other thing, which you call the Connex interview, chat or email, and there you have a conversational experience. So we've had this bifurcation, and I think what we're now going to see is with the arrival of generative AI, these things can come back together where we can bring the best of the online self service, which has been a phenomenal success by any measure, with the conversational, hyper personalized, interactive, customized, much more flexible experiences we can deliver in the client center and those get brought back together for a single unified experience for consumers, I think that's really exciting. And I think it wouldn't have been possible to even think about such a thing until we had this generative AI tool. So that's pillar number one.
Pillar number two, which I'll cover more quickly and we can deep dive on them, is really bringing humans and AI together in a great way. There's a lot of people who think that AI is going to kill a connect center agent and we won't have any more of those people. I don't agree with that. I think that role changes. It evolves into something that's more like a relationship manager. And what happens is they are empowered with Aihdem to be a part of this new experience that we're going to do by bringing these things back together.
Then the third part of it is really making all of this happen with data. There's this new type of data that we didn't have this data two years ago before Genaii called contextual data, that's specific to generative AI models. That's the new digital gold in the era of generative aih. And what we're going to see is a huge rush on this thing to build and grow your set of contextual data to power Genai to enable these other two pillars of this vision.
[00:07:33] Speaker A: So the first pillar that reclaiming CX as our terra firma, it almost seems like everyone have their own personal concierge working for them. You talk about web and contact center kind of integrated for this unified experience. Can you drill into that a little deeper for us? And provide some examples of how you would see this playing out for businesses and their customers and maybe some of the challenges that they might have to get there. Like, what would they have to do to get there?
[00:07:59] Speaker B: Yeah, exactly. Right. So I'm going to start with, like, what does that look like? And I'm going to give an analogy. Right. And let's, let's use our travel use case all over again. Like, if you wanted to book a trip back in the day, even before you, you would call up, you go to travel agency. Remember that. I'm David.
[00:08:16] Speaker A: I remember walking in the door to a trap to walk.
[00:08:18] Speaker B: Exactly. That's how I booked my honeymoon. You walk in a travel agency and you'd have a conversation. Oh, I'm thinking about a sunny via beach vacation and like, what's good this time of year? Oh, you know, maybe the Wine islands, maybe Fiji, and show you some pictures. And it would be this conversational experience that was personalized. And you end, you exit that thing with a trip book. Right. And so you can imagine, and this is the what if, and Robin, this isn't here today. This is this vision. We finally have the tools to do this. You can imagine that if we were able to really supercharge these bots, these chat bots, these voice bots, where instead of the situation today where people don't like them, they often just want to speak to a live agent. Imagine we enable them to have the same superpower and skills and flexibility of that travel agent, where you just have a conversation, have exactly the same discussion. Oh, I'm thinking about each vacation. It's, oh, how about these three different destinations? And it shows you pictures. Oh, what about this one? And said, oh, great hawaiian islands. You know, look at this. There's cheap flights in May, and it shows you flight information. Like, you can have this highly interactive, unscripted, super flexible conversation with like, the bot, the voice and chatbot of our dreams. Like the thing that we wish we had but we don't have today. Right. That's, that's how I think it's going to look when we've, when we built up this technology over the next few years. And the main litmus test is it's only going to work if people love it more than speaking to a live person. And, and there will be a role for a live people, too. Right? That's part of this also is bringing in a human to help close this deal. I can give this example. Wherever, again, you're talking to this bot to book a vacation, and you're hemming and hawing about the price and you're not sure. It says, hold on, let me bring in one of our advisors and then proactively it brings a human into the conversation who says, jonathan, let me tell you, I just went to Fiji last summer and, oh, my God, it was so amazing. The water was so blue. And you hear from a person, I'm not going to believe it when the bot tells me how beautiful the water is. But hearing from human, who can really sell me on it and get me over the finish line. Okay, great, let's go forward. They drop and I complete my transaction. So where there's an example where the human and the AI work together to achieve a goal, everybody wins. Everybody wins. In this situation I just described, I can't wait for us to deliver this to the world.
[00:10:56] Speaker A: I mean, and in your example of travel, I miss that. I miss that about being able to go into a travel agency now. It's like you just rely on web searches and things like that. And that would be a really cool way to just, in your example, you could think of all sorts of examples in other industries where you could have that interaction between an intelligent and interactive bot and then a live agent coming in when needed. So I like that.
[00:11:20] Speaker B: Conversational, unstructured, the kind of things that are difficult to do.
If you go to travel site, you're like, oh, you have to pick where you want to go. And then crisis. It's that discussion to help you select and do unstructured discovery of options that we've lost but can bring back now with the power of generative AI.
[00:11:43] Speaker A: And one of the challenges I see anyway is with bots themselves. Like consumers, when we do our consumer research, they don't have a lot of confidence. Now, businesses are moving forward aggressively with bots, no doubt internally for agency and externally for customers. But in our research, only about 13% say, I prefer using a bot, 40% actively avoid them, the rest them selectively turn this around. I mean, how do we convince the consumers, those 40%, that bots have value? How do you see their role changing in the coming years? What do businesses need to do?
[00:12:16] Speaker B: Yeah, that's a great question. So I think it's been a challenge. Right. We've in our industry that these bots would be great. People would love them, but they haven't performed, and that's why people hate them. And there's a long list of reasons for it. I'd say the main reason is they just can't get the job done for you. They're highly structured, they have limited access to information and controls they can do to get the job done.
And that's what's got to get fixed. And we fix that by using generative AI to exponentially improve the ability of these things to carry on a natural conversation. This is the biggest thing. I think if we didn't have Genai, what I'm describing would be science fiction. But now with Genai, you can sort of see this possible. So if we use generative AI to improve natural language capability, if we integrate all these different rich data sources and give these Genai bots access to all the information they need to do these jobs, to be personalized to, to know who you are and what your past history and preferences and experiences were, so that it felt more like a concierge experience.
If we can do all of that, then we can deliver this almost super agent experience where as a consumer, this is even better than talk to a person because they can operate at the speed of information.
That's something you can't even get today with the human experience. So if you could get the best of both worlds, you get the self service capabilities. You have to wait online, you can get natural conversation, like talking to a human, the speed of digital, like, wow, that would be exciting. That's what it's going to take. And the way to get there. You don't jump right to that destination.
The journey to that is the hardest part. And it starts little by little, by introducing more and more generative AI into your contact center, giving it more and more contextual data, increasing the scope of use cases more and more, so that you work up and up towards these things with successes, with use case at a time. That's how you get there.
[00:14:24] Speaker A: And I think even from almost like a marketing standpoint, I guess I think companies need to be more deliberate about telling consumers, almost like marketing the technology to consumers, and even doing little subtle things. I've just gotten done with a chatbot experience, my problems resolved. I get a message back from the chatbot saying, are you satisfied? Yes, everything's good. Okay, we've been able to resolve your problem in two minutes and 15 seconds. Had you had to call into our contact center today and only speak to live agent, it would have taken twelve minutes and 18 seconds, things like that, so that consumers are starting to internalize. Maybe I should give this technology a chance, you know? But anyway, I want to go to your second pillar, because I love this pillar of human and digital working together, because I think that's really been the vision of AI for a while. You know, we're starting to see some evidence of that happening already with agent assist. Our data on agent assist is through the roof of how successful it's been for companies. But I think your vision of this goes a lot further. So can you see, talk a little bit more about where you see this heading? How do multimodal and perpetual channels even play a role here moving forward?
[00:15:34] Speaker B: Exactly right. So I think the first thing I'd say is that we have to think about it as the role of the agent shifts, because the type of work that is coming to the agent shifts today, a lot of stuff comes to the agent, and it's like this boring routine, password reset. I almost have to roll my eyes every time I say that. It's still the case that a lot of calls are passed or reset. I'm like, hello, it's 2024. Why are we still having to solve with humans? It's unbelievable.
Once these initial steps of this transition and this market evolution happen, and we can pull those use cases off. Right. The role of the agent becomes much more about this relationship manager. And to do that job effectively, they need a lot more context, a lot more information, a lot more tools and data at their disposal. So again, here's where Jennai can really help is it can take all this information that we have about the customer who's connected to the connect center. All their past, the transcripts of their past chats and emails and transcripts of their voice calls, their current account status, their last 20 bills, all of this monster information. Now, this contextual data can be fed into the genai models to provide agents with, like, concise summaries of the state of the customer so that when you call up, like, they know what's going on. And people complain about this today, tons. Like, it feels so impersonal. I'm your number one customer. I call up and you're like, I don't even know anything about you. Let me look up your orders. Come on.
If consumers could feel like it's, again, this concierge idea I'm connecting, this person knows me. They know what I've done. They know what I need. And we can solve that by giving agents all this great contextual information and then help them work faster by giving them access to answers that customers are asking in an instant so that we want to get rid of call hold. This is another thing. I'm on a rampage sometimes.
How much time does the world spend sitting on hold on the contact center? It's, like, terrible. Everybody hates it. And a lot of the reason it happens is the agent is spending time doing something or looking up information. We can use generative AI to give them that information to automate those tasks that they're spending time doing. So that, you know what? If you go visit the travel agency in person, you don't get put on hold. You have their attention the whole time. Why can't I have that back?
So this is the technology and the people coming together to deliver this personalized concierge style experience.
Again, only possible now that we have this superpower of Gen AI.
[00:18:16] Speaker A: I think that last pillar about contextual data being the new gold is perhaps, maybe you agree, maybe you don't. But for me, it's perhaps the most important because we can't really accomplish any of this without that integrated, contextual, relevant real time data. In our last round of research, I think we had about just over 50% of companies said they had integrated at least some of their disparate databases. But I think that level of integration will be really much more difficult to sustain as time goes on and we start, we're pulling data from more and more places. So how do you see companies achieving that nirvana of truly integrated contextual data that AI can then leverage in real time? What are the ultimate benefits here? And maybe even as you're explaining it, give us some real world examples of this piece of the vision.
[00:19:05] Speaker B: Yeah, so just again, just to level set, because I think this is, like I said, sort of a new concept, and you and I get it. But for listeners out there, what is contextual data in the world of generative AI? I'm going to put it in technical terms, the way generative AI works is this model. And you give it a prompt, which you can think of if you go use chat GPT, it's the thing, you type at it. Here's a five page document. Please provide a summary of this document. That thing, the document you gave it, and you request to summarize it. That's the prompt. Contextual data is information you provide in that prompt to help the generative AI model do its job. In my super simple example I just gave, where I asked the genai to summarize the document, that there was two pieces of contextual data. The document I asked it to summarize, and then the instructions to please summarize this document, and I might give more detailed instructions. Summarize a document, put it in pirate speak. In your summary, those are instructions. So instructions and content. So that's all contextual data, and that's the simplest form where things get really exciting is these generative AI models can take a lot of contextual, lots of information, and the more you give it, the better it is at its doing its job. So contextual data is at the time of the call, at the time of the chat, at the time of the interaction, taking all this information about the consumer, about the business, and feeding it into the generative AI model. So it has all this information to do its job. And again, it's the analogy of just like a person. If you were working with a concierge and they knew you, they knew all of your past orders, they knew all of your past products that you bought, they knew your preferences, they knew that they remembered the last five conversations you had together. They're going to be much better at personalizing that interaction.
And then they have access to everything. They can do stuff. They can access all these systems to book your flight, to look up photos of different locations, to find out services at hotels, and all that's runtime data that the generative AI model gets access to via this prompt. So the technology we need is to do more of pulling this data sources. And you're right, Robin. In some ways the situation has gotten worse and that enterprises have even more systems now that have even more data. And in many ways, this is still one of the hard problems. It's not like you still got to get app to have APIs, you got to get that data.
What I think you're going to see is increased investment from vendors like us. And we are doing this in making it easier and easier to integrate into these third party data sources, to pull on their APIs, to push on data into them, and to collect that data and integrate it into these generative AI models, this layer of AI middleware that serves as the glue between these sources data via APIs and the generative AI models. That's the place that's going to differentiate vendors and needs all this investment to bring this data in, to deliver this experience. Because like you said, Robin, the contextual data is the enabler for all the things that we've been talking about.
[00:22:12] Speaker A: Yeah. And it's such an important investment. Like you said, for vendors, it's happening kind of in the background. It's not this thing that everyone's like, oh, wow, this is so. No, it's stuff that has to happen in order to enable the oaus, you know?
[00:22:22] Speaker B: Yeah, exactly.
[00:22:23] Speaker A: But a couple more questions for you. I know that five nine recently acquired Equionde, and I want to understand, like, if you see the Aquaman technology, helping with your vision here, will it enable you to expand into other areas, maybe sales and marketing or other different areas of companies?
[00:22:41] Speaker B: Yeah, absolutely. In fact, it's directly related to this first piece of this vision where I said, our vision here is for the CCAAs industry to reclaim the terra firma as the central point of CX, between a customer and a brand. Right. That's the vision. It's a bolden old vision, and today we have just a slice of it, which is like inbound calls and some amount of outbound. And what we've seen over the years is that other forms of customer communication have gone to other places. Acreon brings another form of interaction between a brand and a consumer. That's something that's not in five nine's portfolio up till now, which was this sort of more proactive outreach kind of communication across multiple channels.
And that's not something we've been doing, but it's another piece of this overall experience of communication between a customer and a brand. And so that when we add that in, we increase our width of the set of things that we're doing that encompass all the exchange of information between a customer and brand. So it's directly related to this vision, and we see it bringing more contextual data, more interaction history that is all folded into the system, so that if you respond to one of those pieces of outreach, the system knows about it and it's tied in with the inbound campaign. So it's a very, very natural next step towards this vision.
[00:24:01] Speaker A: Yeah, that's interesting to see how that's going to be expanding. I'll be really watching that to see that the new capabilities you guys come out with in these areas. So just wrapping up then, we talked a lot about big picture and a lot of forward looking vision. But if there were one or two things maybe that CX leaders at businesses could do today, right now, to start helping to bring this vision to life, what would those be?
[00:24:28] Speaker B: No. Yeah, two things I would say out of the gate to do that you could do right now. One, Robin, you mentioned this. The runaway success story for generative AI in the Connecticut center right now is post call summarization agent assist. It's been like supercharged. It's like a no brainer Roi. It helps CX, it saves you money. You click and you turn it on.
Us and our competitors all have this product now. Everyone's doing. If you're not turning this thing on, you are just missing the boat. Do it now. Okay. That's like an easy first step. And the second is guys like accelerate your self service, like some basic use cases with iva, natural language, better voices.
You can do that right now. And as a step towards this journey, know if I have to dial one more IVR where it says press one for sales and press two for support. And, you know, I'm going to just come on, guys, let's deliver for the world. Let's make Connex center a delightful experience. We are far from it. Let's work on it.
[00:25:29] Speaker A: Yes. And I'm sure if people are already beyond that, you guys are happy to talk to them about what else they could be doing.
All right. Jonathan, well, thank you so much for joining us today and for sharing your vision. I think it's fascinating to see what we'll be doing. It'll be great to be talking about this in the coming years and seeing, watching it sort of come to life.
[00:25:49] Speaker B: Exactly. Robin, it's great talking with you this morning as well.
[00:25:52] Speaker A: Thank you.
[00:25:53] Speaker B: Thanks, listeners.
