This conversation explores artificial intelligence in contact centers with Amit Mathradas of Five9, chief executive officer and board member, at theCUBE Research Contact Center Summit 2026. Mathradas discusses the transition from pilot to production for AI in contact centers, focusing on agentic AI, agent assist, voice AI deployment, quality management and platform interoperability as well as Five9's approach to low-latency scalable AI-driven customer engagement.
Mathradas recommends organizations begin with clearly defined problems, prioritize agent assist and agentic quality management and select open platforms that integrate with existing infrastructure. They caution against attempting to address all use cases at once and advocate phased adoption. theCUBE Research hosts the conversation with Bob Laliberte and Zeus Kerravala, who emphasize governance, security, change management and measurement as key factors to achieve faster time-to-value for AI in customer experience. The discussion addresses scalability, cloud contact center deployment and platform interoperability and outlines strategies to move from pilot to production while maintaining quality and latency requirements.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
The AI ROI in Contact Center Summit. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for The AI ROI in Contact Centers Summit
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for The AI ROI in Contact Centers Summit.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
The AI ROI in Contact Center Summit. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to The AI ROI in Contact Center Summit
Please sign in with LinkedIn to continue to The AI ROI in Contact Center Summit. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Amit Mathradas, Five9
This conversation explores artificial intelligence in contact centers with Amit Mathradas of Five9, chief executive officer and board member, at theCUBE Research Contact Center Summit 2026. Mathradas discusses the transition from pilot to production for AI in contact centers, focusing on agentic AI, agent assist, voice AI deployment, quality management and platform interoperability as well as Five9's approach to low-latency scalable AI-driven customer engagement.
Mathradas recommends organizations begin with clearly defined problems, prioritize agent assist and agentic quality management and select open platforms that integrate with existing infrastructure. They caution against attempting to address all use cases at once and advocate phased adoption. theCUBE Research hosts the conversation with Bob Laliberte and Zeus Kerravala, who emphasize governance, security, change management and measurement as key factors to achieve faster time-to-value for AI in customer experience. The discussion addresses scalability, cloud contact center deployment and platform interoperability and outlines strategies to move from pilot to production while maintaining quality and latency requirements.
>> Artificial intelligence has quickly become one of the most significant technology shifts enterprises have seen in decades. While many organizations are still exploring how best to apply AI across their business, customer experience has emerged as one of the earliest and most successful use cases. Today's contact centers are evolving well beyond traditional automation. AI agents are beginning to augment, in some cases perform, complex customer interactions, helping organizations improve satisfaction while reducing operational costs and enabling employees to focus on higher-value work. Hello everyone, and welcome to theCUBE Research Contact Center Summit. I'm Bob Laliberte, Principal Analyst at theCUBE Research, and I'm here with my co-host Zeus Kerravala, Founder and Principal Analyst of ZK Research. Joining us today is Amit Mathradas, CEO of Five9. Amit, welcome.
Amit Mathradas
>> Well, thank you. Thank you for having me.
Zeus Kerravala
>> Yeah, I'm going to actually start the questions off here. when we talk with other IT leaders and business leaders, there's not one today that doesn't have AI at the top of the agenda. In fact, we always talk about how 2026 is the year that we went from AI pilots to production. But when I ask them to focus on where they're gonna be applying AI, the first place they seem to always bring up is customer experience. And so why do you think customer experience has been such a compelling use case for AI?
Amit Mathradas
>> Well, thank you, Zeus Kerravala, and really good question. Look, as you all know, AI is now being used in multiple practices across organizations, but it doesn't surprise me or anyone that AI's first use case or the most prevalent use case is in contact centers. For two key reasons, right? First, it started off as effectively, if you look at it, 9 out of the $10 that go into OpEx into a contact center is labor, it's humans. So when I think about where you want to use new tools like AI, that seems like a natural first starting point, right? How do I actually optimize? How do I make my unit costs better of what comes out of a contact center? What is strangely happening is, as AI has gotten better, you're actually finding that it is not just removing costs, but it is actually revolutionizing how companies are leveraging contact centers and the experience that they are now providing, right? Today's new AI, agentic AI, is effectively saying, hey, I can take all my agents and learn from all of them. Before when we did quality management, it was like, let's listen to 5% of calls or 10% and extend an understanding of what's happening in the rest of the contact center. Today with AI, you can listen to 100% of the calls and not just make corrections, but also route it to specific human beings to solve specific problems versus saying, I'm gonna send it to one part of my contact center and hopefully it gets solved there. So I think these are the two big fundamental shifts that are happening. It started with labor, but now it's moved to how do we drive better experiences? And lo and behold, what we're already seeing where interactions are starting to go up because as unit costs of serving a customer in the contact center come down, you can serve more and more interactions, more and more of your customers, and that's leading to much greater outcomes for our customers at the end of it.
Bob Laliberte
>> Amit, one of the things that I've been fascinated by and following is something I refer to as the time to comfort with AI. Are you seeing that organizations are becoming a lot more comfortable trusting AI with their customer interactions as well?
Amit Mathradas
>> Yeah, for sure. Look, AI, even in the span of the last 6 months, a year, a lot of it has moved from POCs or proof of concepts into actual deployments. Two things are happening there as well. One, the new generation of AI and the new generation of the models that are powering AI have only gotten better. Latency is improving, hallucinations are down, complexity of answering that you can get to is up. That is one. The foundational capabilities have gotten better and you can serve more and more complex use cases. The second thing that we're also noticing is organizations have become better at organizing around AI. Security, governance, better management, actual thought process of how do you deploy it. It's not just let's throw it on the wall and hope it sticks and we get an outcome. It's actually now being planned. There are use cases with very specific outcomes in mind. Those two things coming together is really kind of helping this transition accelerate.
Bob Laliberte
>> Got it. Got it. So I know there's a lot of organizations jumping into it. Zeus Kerravala mentioned about the whole pilot to production this year and so forth, but we also know there's still a number of organizations that haven't started yet. They're still on the sidelines. What advice would you give to organizations that are just getting started on that AI journey in the contact center?
Amit Mathradas
>> Look, the first thing anytime I talk to a customer or I get requests on this is focus on what is it you're trying to solve. A lot of people rush in saying, "AI can solve this. I'm just gonna throw AI at the problem and we'll figure out if it works." My strongest recommendation is first figure out what is the problem? Where is the longest, you know, pole in the tent that you are trying to solve with your customer experience? And then kind of consider does AI actually solve that and how do you actually plan around the implementation there. That's number 1. Number 2 is, once you've decided that, how are you organizing for the outcome? It's not just— you have to now have teams that are specialized or specific with how to drive that outcome, how to measure that outcome, how to deploy that outcome. So think through not just the capabilities of the technology, but the tooling that goes behind it. And how you get the best of breed. The third one for me, and this is, as I think about AI in the contact center, how are you actually choosing the product or the vendor or the capability that utilizes the infrastructure you have today? I think that is a big one because very often customers will go in and then they'll come back and be like, oh snap, I gotta actually change another whole set of tools for what I'm actually working with this vendor to go solve. So who can actually solve it with the infrastructure you have today and who can actually give you an open platform to go drive those connection points, I think is a really key piece. And I think that if you can nail those 3 things and go in clear-eyed, I think you will end up with, clear outcomes. Don't boil the ocean. Don't go say I'm solving everything with AI. Pick a use case, get it right, move to the next, move to the next, and you will have a pretty robust outcome.
Zeus Kerravala
>> Yeah, it's interesting you bring that up because within the CX ocean, I guess to use your analogy, there's a lot of things you could do, right? There's agent assist, there's GenAI agents, there's note-taking and things like that. And so do you find when you talk to customers, they're trying to do too much? And if so, when you talk to them, where do you recommend they start to be able to demonstrate that value quickly and get some good wins there?
Amit Mathradas
>> Yeah, look, if it's a customer who hasn't had experience with AI before they go into voice agents, which I think is the most complex to deploy and get right and tune, I always recommend, hey, start with something like an agent assist. Get the most out of your entire organization that is sitting there today. It is easier to deploy. You actually get better outcomes with your agents, you know, getting a whisper in their ear telling them, hey, you're missing this so you can add this. Maybe move to agentic quality management that can actually lift all boats across your organization and give you better servicing metrics on, hey, who's performing, who's not. And then once you're comfortable with how you actually deploy these, then I would recommend, hey, go into voice AI. There is tremendous outcome there. There's a tremendous lift, but it also requires a lot more engineering might, a little bit more. It's not out of the box. You have to fine-tune. You have to think about the use cases. often forward-deployed engineers working with you on your infrastructure to get this right. So that is generally the steps, but if you are comfortable, you've gone through the first one or two, you're seeing the outcome, start moving to voice. You will see the next big lift that's coming from it.
Zeus Kerravala
>> Yeah, so on the topic of agents, that's really the topic du jour at every event we go to today, and really every CIO conversation I have. From your perspective, how do you think about it? Just how do you define it? And how does that differ from a lot of the chatbots and virtual agents that we've frankly been using for years, right?
Amit Mathradas
>> Yeah, look, the prior generation, right? The new generation is agentic, AI agents. The prior generation was very much— it was predominantly built on deterministic, right? And what I mean by that is it was a decision tree. Effectively, you went in and you programmed your bot. Rules-based. Yeah, it was rules-based. Hey, if-then statements. say you're asking for, do you have a credit card? Yes, it takes you down one track. No, it takes you down another. So it was effectively a decision tree masked with voice capabilities and would take you down that. The new generation of agentic is probabilistic. And what that means is it will contemplate all different probabilities of what you are asking and then narrow it down based on the knowledge that it has gathered from your organization. organizational knowledge it has from the entire market, how you've trained and tuned the bots. Effectively, what that allows you to do is get into more complex solutioning. It can actually understand what you're asking for. It can probably go through a probabilistic understanding and say, based on this, it actually wants— the customer is asking for that and I can narrow it down to this route or transfer to a human or transfer you to a content site. All those things are what's in the new generation of agentic AI. And it is truly, in my view, revolutionizing what is happening in the contact center and serving a lot more complex use cases.
Bob Laliberte
>> Now that sounds really fascinating. And one of the things that you always get is that, what's the right balance between humans and the agent and so forth? Where are you seeing customers? How do they determine what that balance is between, leveraging the agents and automation and human engagement?
Amit Mathradas
>> Look, I— we, Five9, have a very, very strong belief that the world, the future world is what we are calling Humantic, right? And Humantic is the combination of humans and agentic sitting in your contact center. I think there, from my perception, from my perspective, I think there is a misconception that all humans are going away in the contact center. That is not true. Talking to our customers that, deal in complex and regulated industries, there are 3 very clear use cases: complexity, value, and vulnerability. When you are facing one of these 3 use cases, you want a human involved, whether it is your highest value customer, whether it is someone in a healthcare environment who may have just been dealing with the death of a partner or a family member and is dealing with an insurance case, or it's highly complex and you want to understand what's happening with all my stocks and trades and someone needs to show you the different optionality. So for us, I think this combination of how does AI solve a lot of the base cases, the high-volume cases— the two areas we see AI solving is high volume, password resets, what's my bank balance, things like that, or when I need to be available 24/7. I can call in the middle of the night. Someone will take my call and transfer it in, and it's in a high kind of desirability code where you need someone available. And where humans are going to be is in this complexity. And so as we think about the world, Bob, we are building a world for this human-AI era where architecturally our voice AI agent is connected to our contact center as a service platform so that you can get low latency, you can get humans to actually jump into an AI call and take it over if there is an issue. You can get these levels of service that drive the next level of containment across the board. I'll close it with this one thing. From the research we have done, we found one stat really amazing. 99% of business practitioners, companies who are deploying, think that their contact centers and their serviceability has gotten better. Only 66% of actual users think the contact center has gotten better. That means a third of all your customers are actually saying the experience is worse, and more than 50% of them are saying the reason is I want to access a human and I don't get that. So just put that into perspective, and that's what's in the back of our minds as we build.
Bob Laliberte
>> Excellent. Yeah, and it's interesting as these deployments occur, a lot of people are so focused on the technology. You brought up before, it's also about people and processthat need to be involved as well. So we know the technology alone doesn't determine success, but a lot of times that implementation speed, integration, change management can really help make the difference. So how does Five9 help customers move from their pilot projects to production and get to that point where they're realizing business value quickly?
Amit Mathradas
>> Yeah, look, our heritage, we have over 20 years of experience in being voice-centric. We know that this is where the complexity in the ecosystem is. Like our name suggests, five nines of uptime, 180 countries, 3,500 customers, nearly 90 certifications and regulatory needs that go with that. The way we help is we start partnering and working with customers on two fronts. One, our forward-deployed engineers, our capabilities around PS or professional services, understanding the need, starts with understanding the complexity you're dealing with, and then how are you deploying, building, and going from there. The second big one for me is the open platform and the open architecture. You will never hear us at Five9 saying you have to end this technology or toolkit that you're using for our capabilities to work. It's about how do you actually open up and drive our solutioning on top of what you already have today. As we get better and better and service you more, you will effectively start picking up, greater pieces from us. That is traditionally what customers want as they are deploying new technologies and a helping hand as they go through this shift.
Bob Laliberte
>> Yeah.
Zeus Kerravala
>> And Amit, you mentioned that data and that was interesting 'cause obviously if some customers are seeing value, right, there is value in it, right? I think that's safe to say. So from the deployments that you've seen, What are the common characteristics of those deployments that make them successful?
Amit Mathradas
>> Yeah, some of it is a combination of, just to kind of summarize what I've been saying. One is a lot of them start with an organization that is ready for AI and have actually thought through the use case and how they want to deploy it. Two, a lot of them start with a use case. Even if they start directly with voice AI agents, they will start with one simple case, build it, get it right, expand to the next department, the next department. And the last one is the ability for them to really pick and drive initiatives with the ecosystem and infrastructure that they have today. I think that is another key reason why they are successful and what kind of drives back that they're not forced to make changes on the whole entire backend stack for their AI to work. They can kind of build it on top of what's available today and go drive and test from there.
Zeus Kerravala
>> All right, well, that being said, can you give me a customer example that demonstrates the business outcome they were hoping to achieve and then the result that the deployment had?
Amit Mathradas
>> Yeah, look, there are a lot of them. And the one I will probably cite is there's a large moving and logistics company that effectively, and storage company, been in business for 25 years, has been a customer of ours for a long time, came to us first when AI became real. Started to deploy AI agents, started to deploy AQM. Pretty recently they moved to our GenAI agentic voice bot. And over time now we're on track by the end of this year to handle about 100,000 calls a year for them running through our agentic stack. And what this has led to is not just the improvements around the overall agent ecosystem with Agent Assist and AQM, but even with our AI bots now, they have gone over 50% containment on their specific use case. It's moving towards 53. It's higher than what the original goal was. Their CSAT has gotten better. And strangely enough, their agent churn rate has come down even in that time because you've taken all the manual work out and moved a lot of the grunt work to AI agents. So that's just one example of how we're seeing customers use the full stack. And keep maturing along with it.
Bob Laliberte
>> Yeah, that's great. I love to hear those customer examples, right? It really brings it home to a lot of people who are watching. clearly we're still in the early stages of people adopting AI and really understanding how they're going to get all the value from it. So I'm wondering, as I wrap up, I like to look at, looking out a couple of years, how do you see AI transforming customer engagement? What should organizations be doing today to prepare for what's coming next?
Amit Mathradas
>> Look, I have a huge belief that the contact center 5 years from now is not gonna look anything like the contact center does today, right? And that is, and I hope a lot of your listeners are really grasping that and saying, well, yes, humans will be around, but what they do, how they service, there's a whole new world of customer memory that is coming to light. Let me give you an example. Tomorrow with the capabilities of an AI agent being able to listen to every single call. You can now start capturing customer sentiment, which is the biggest part of the interaction layer, right? If you have called in and we both have bought the same pair of shoes from the same vendor, same size, a CRM will capture that. The sentiment will capture, did I have a great experience with that agent when I was talking to them or did I not? And that can surface in the next engagement when you call back the next time, I can open it up and say, "I'm sorry, you were talking to Amit, he completely sucked, I'm gonna give you a $20 coupon." Well, these new capabilities around the platform, around how contact centers are going to evolve are new. So that's one area it is going to pivot. The next big thing is around how these systems of record all work together. To actually enable the next shift that is coming, right? If the interaction layer, digital, voice, AI coming together is going to create these new experiences, well, you have to be able to connect to the right systems of record to enable the orchestration to take place. What are you driving with these outcomes? So orchestration is the next big layer that I think a lot of customers should be thinking about and companies like us are thinking about in terms of where it needs to go.
Bob Laliberte
>> I think, I think those are really valid points. It makes a lot of sense. Amit, thank you so much for joining us today. Zeus Kerravala, thank you for co-hosting.
Amit Mathradas
>> Thank you both. Yeah, really enjoyable.
Bob Laliberte
>> Absolutely. Well, it's really clear that AI is moving beyond experimentation, beginning to deliver measurable business outcomes across customer experience. The organizations that combine the right technology with the thoughtful implementation and a focus on business outcomes are likely to realize value much faster than those approaching AI as simply another automation project. So I wanna thank everyone for watching. If you enjoyed the conversation, be sure to explore the rest of the sessions from theCUBE Research Contact Center Summit, where we're talking with industry leaders about how AI is transforming customer engagement, operations, and enterprise technology. For Zeus Kerravala, I'm Bob Laliberte. Thanks for joining us, and we'll see you next time.
>> Artificial intelligence has quickly become one of the most significant technology shifts enterprises have seen in decades. While many organizations are still exploring how best to apply AI across their business, customer experience has emerged as one of the earliest and most successful use cases. Today's contact centers are evolving well beyond traditional automation. AI agents are beginning to augment, in some cases perform, complex customer interactions, helping organizations improve satisfaction while reducing operational costs and enabling employees to focus on higher-value work. Hello everyone, and welcome to theCUBE Research Contact Center Summit. I'm Bob Laliberte, Principal Analyst at theCUBE Research, and I'm here with my co-host Zeus Kerravala, Founder and Principal Analyst of ZK Research. Joining us today is Amit Mathradas, CEO of Five9. Amit, welcome.
Amit Mathradas
>> Well, thank you. Thank you for having me.
Zeus Kerravala
>> Yeah, I'm going to actually start the questions off here. when we talk with other IT leaders and business leaders, there's not one today that doesn't have AI at the top of the agenda. In fact, we always talk about how 2026 is the year that we went from AI pilots to production. But when I ask them to focus on where they're gonna be applying AI, the first place they seem to always bring up is customer experience. And so why do you think customer experience has been such a compelling use case for AI?
Amit Mathradas
>> Well, thank you, Zeus Kerravala, and really good question. Look, as you all know, AI is now being used in multiple practices across organizations, but it doesn't surprise me or anyone that AI's first use case or the most prevalent use case is in contact centers. For two key reasons, right? First, it started off as effectively, if you look at it, 9 out of the $10 that go into OpEx into a contact center is labor, it's humans. So when I think about where you want to use new tools like AI, that seems like a natural first starting point, right? How do I actually optimize? How do I make my unit costs better of what comes out of a contact center? What is strangely happening is, as AI has gotten better, you're actually finding that it is not just removing costs, but it is actually revolutionizing how companies are leveraging contact centers and the experience that they are now providing, right? Today's new AI, agentic AI, is effectively saying, hey, I can take all my agents and learn from all of them. Before when we did quality management, it was like, let's listen to 5% of calls or 10% and extend an understanding of what's happening in the rest of the contact center. Today with AI, you can listen to 100% of the calls and not just make corrections, but also route it to specific human beings to solve specific problems versus saying, I'm gonna send it to one part of my contact center and hopefully it gets solved there. So I think these are the two big fundamental shifts that are happening. It started with labor, but now it's moved to how do we drive better experiences? And lo and behold, what we're already seeing where interactions are starting to go up because as unit costs of serving a customer in the contact center come down, you can serve more and more interactions, more and more of your customers, and that's leading to much greater outcomes for our customers at the end of it.
Bob Laliberte
>> Amit, one of the things that I've been fascinated by and following is something I refer to as the time to comfort with AI. Are you seeing that organizations are becoming a lot more comfortable trusting AI with their customer interactions as well?
Amit Mathradas
>> Yeah, for sure. Look, AI, even in the span of the last 6 months, a year, a lot of it has moved from POCs or proof of concepts into actual deployments. Two things are happening there as well. One, the new generation of AI and the new generation of the models that are powering AI have only gotten better. Latency is improving, hallucinations are down, complexity of answering that you can get to is up. That is one. The foundational capabilities have gotten better and you can serve more and more complex use cases. The second thing that we're also noticing is organizations have become better at organizing around AI. Security, governance, better management, actual thought process of how do you deploy it. It's not just let's throw it on the wall and hope it sticks and we get an outcome. It's actually now being planned. There are use cases with very specific outcomes in mind. Those two things coming together is really kind of helping this transition accelerate.
Bob Laliberte
>> Got it. Got it. So I know there's a lot of organizations jumping into it. Zeus Kerravala mentioned about the whole pilot to production this year and so forth, but we also know there's still a number of organizations that haven't started yet. They're still on the sidelines. What advice would you give to organizations that are just getting started on that AI journey in the contact center?
Amit Mathradas
>> Look, the first thing anytime I talk to a customer or I get requests on this is focus on what is it you're trying to solve. A lot of people rush in saying, "AI can solve this. I'm just gonna throw AI at the problem and we'll figure out if it works." My strongest recommendation is first figure out what is the problem? Where is the longest, you know, pole in the tent that you are trying to solve with your customer experience? And then kind of consider does AI actually solve that and how do you actually plan around the implementation there. That's number 1. Number 2 is, once you've decided that, how are you organizing for the outcome? It's not just— you have to now have teams that are specialized or specific with how to drive that outcome, how to measure that outcome, how to deploy that outcome. So think through not just the capabilities of the technology, but the tooling that goes behind it. And how you get the best of breed. The third one for me, and this is, as I think about AI in the contact center, how are you actually choosing the product or the vendor or the capability that utilizes the infrastructure you have today? I think that is a big one because very often customers will go in and then they'll come back and be like, oh snap, I gotta actually change another whole set of tools for what I'm actually working with this vendor to go solve. So who can actually solve it with the infrastructure you have today and who can actually give you an open platform to go drive those connection points, I think is a really key piece. And I think that if you can nail those 3 things and go in clear-eyed, I think you will end up with, clear outcomes. Don't boil the ocean. Don't go say I'm solving everything with AI. Pick a use case, get it right, move to the next, move to the next, and you will have a pretty robust outcome.
Zeus Kerravala
>> Yeah, it's interesting you bring that up because within the CX ocean, I guess to use your analogy, there's a lot of things you could do, right? There's agent assist, there's GenAI agents, there's note-taking and things like that. And so do you find when you talk to customers, they're trying to do too much? And if so, when you talk to them, where do you recommend they start to be able to demonstrate that value quickly and get some good wins there?
Amit Mathradas
>> Yeah, look, if it's a customer who hasn't had experience with AI before they go into voice agents, which I think is the most complex to deploy and get right and tune, I always recommend, hey, start with something like an agent assist. Get the most out of your entire organization that is sitting there today. It is easier to deploy. You actually get better outcomes with your agents, you know, getting a whisper in their ear telling them, hey, you're missing this so you can add this. Maybe move to agentic quality management that can actually lift all boats across your organization and give you better servicing metrics on, hey, who's performing, who's not. And then once you're comfortable with how you actually deploy these, then I would recommend, hey, go into voice AI. There is tremendous outcome there. There's a tremendous lift, but it also requires a lot more engineering might, a little bit more. It's not out of the box. You have to fine-tune. You have to think about the use cases. often forward-deployed engineers working with you on your infrastructure to get this right. So that is generally the steps, but if you are comfortable, you've gone through the first one or two, you're seeing the outcome, start moving to voice. You will see the next big lift that's coming from it.
Zeus Kerravala
>> Yeah, so on the topic of agents, that's really the topic du jour at every event we go to today, and really every CIO conversation I have. From your perspective, how do you think about it? Just how do you define it? And how does that differ from a lot of the chatbots and virtual agents that we've frankly been using for years, right?
Amit Mathradas
>> Yeah, look, the prior generation, right? The new generation is agentic, AI agents. The prior generation was very much— it was predominantly built on deterministic, right? And what I mean by that is it was a decision tree. Effectively, you went in and you programmed your bot. Rules-based. Yeah, it was rules-based. Hey, if-then statements. say you're asking for, do you have a credit card? Yes, it takes you down one track. No, it takes you down another. So it was effectively a decision tree masked with voice capabilities and would take you down that. The new generation of agentic is probabilistic. And what that means is it will contemplate all different probabilities of what you are asking and then narrow it down based on the knowledge that it has gathered from your organization. organizational knowledge it has from the entire market, how you've trained and tuned the bots. Effectively, what that allows you to do is get into more complex solutioning. It can actually understand what you're asking for. It can probably go through a probabilistic understanding and say, based on this, it actually wants— the customer is asking for that and I can narrow it down to this route or transfer to a human or transfer you to a content site. All those things are what's in the new generation of agentic AI. And it is truly, in my view, revolutionizing what is happening in the contact center and serving a lot more complex use cases.
Bob Laliberte
>> Now that sounds really fascinating. And one of the things that you always get is that, what's the right balance between humans and the agent and so forth? Where are you seeing customers? How do they determine what that balance is between, leveraging the agents and automation and human engagement?
Amit Mathradas
>> Look, I— we, Five9, have a very, very strong belief that the world, the future world is what we are calling Humantic, right? And Humantic is the combination of humans and agentic sitting in your contact center. I think there, from my perception, from my perspective, I think there is a misconception that all humans are going away in the contact center. That is not true. Talking to our customers that, deal in complex and regulated industries, there are 3 very clear use cases: complexity, value, and vulnerability. When you are facing one of these 3 use cases, you want a human involved, whether it is your highest value customer, whether it is someone in a healthcare environment who may have just been dealing with the death of a partner or a family member and is dealing with an insurance case, or it's highly complex and you want to understand what's happening with all my stocks and trades and someone needs to show you the different optionality. So for us, I think this combination of how does AI solve a lot of the base cases, the high-volume cases— the two areas we see AI solving is high volume, password resets, what's my bank balance, things like that, or when I need to be available 24/7. I can call in the middle of the night. Someone will take my call and transfer it in, and it's in a high kind of desirability code where you need someone available. And where humans are going to be is in this complexity. And so as we think about the world, Bob, we are building a world for this human-AI era where architecturally our voice AI agent is connected to our contact center as a service platform so that you can get low latency, you can get humans to actually jump into an AI call and take it over if there is an issue. You can get these levels of service that drive the next level of containment across the board. I'll close it with this one thing. From the research we have done, we found one stat really amazing. 99% of business practitioners, companies who are deploying, think that their contact centers and their serviceability has gotten better. Only 66% of actual users think the contact center has gotten better. That means a third of all your customers are actually saying the experience is worse, and more than 50% of them are saying the reason is I want to access a human and I don't get that. So just put that into perspective, and that's what's in the back of our minds as we build.
Bob Laliberte
>> Excellent. Yeah, and it's interesting as these deployments occur, a lot of people are so focused on the technology. You brought up before, it's also about people and processthat need to be involved as well. So we know the technology alone doesn't determine success, but a lot of times that implementation speed, integration, change management can really help make the difference. So how does Five9 help customers move from their pilot projects to production and get to that point where they're realizing business value quickly?
Amit Mathradas
>> Yeah, look, our heritage, we have over 20 years of experience in being voice-centric. We know that this is where the complexity in the ecosystem is. Like our name suggests, five nines of uptime, 180 countries, 3,500 customers, nearly 90 certifications and regulatory needs that go with that. The way we help is we start partnering and working with customers on two fronts. One, our forward-deployed engineers, our capabilities around PS or professional services, understanding the need, starts with understanding the complexity you're dealing with, and then how are you deploying, building, and going from there. The second big one for me is the open platform and the open architecture. You will never hear us at Five9 saying you have to end this technology or toolkit that you're using for our capabilities to work. It's about how do you actually open up and drive our solutioning on top of what you already have today. As we get better and better and service you more, you will effectively start picking up, greater pieces from us. That is traditionally what customers want as they are deploying new technologies and a helping hand as they go through this shift.
Bob Laliberte
>> Yeah.
Zeus Kerravala
>> And Amit, you mentioned that data and that was interesting 'cause obviously if some customers are seeing value, right, there is value in it, right? I think that's safe to say. So from the deployments that you've seen, What are the common characteristics of those deployments that make them successful?
Amit Mathradas
>> Yeah, some of it is a combination of, just to kind of summarize what I've been saying. One is a lot of them start with an organization that is ready for AI and have actually thought through the use case and how they want to deploy it. Two, a lot of them start with a use case. Even if they start directly with voice AI agents, they will start with one simple case, build it, get it right, expand to the next department, the next department. And the last one is the ability for them to really pick and drive initiatives with the ecosystem and infrastructure that they have today. I think that is another key reason why they are successful and what kind of drives back that they're not forced to make changes on the whole entire backend stack for their AI to work. They can kind of build it on top of what's available today and go drive and test from there.
Zeus Kerravala
>> All right, well, that being said, can you give me a customer example that demonstrates the business outcome they were hoping to achieve and then the result that the deployment had?
Amit Mathradas
>> Yeah, look, there are a lot of them. And the one I will probably cite is there's a large moving and logistics company that effectively, and storage company, been in business for 25 years, has been a customer of ours for a long time, came to us first when AI became real. Started to deploy AI agents, started to deploy AQM. Pretty recently they moved to our GenAI agentic voice bot. And over time now we're on track by the end of this year to handle about 100,000 calls a year for them running through our agentic stack. And what this has led to is not just the improvements around the overall agent ecosystem with Agent Assist and AQM, but even with our AI bots now, they have gone over 50% containment on their specific use case. It's moving towards 53. It's higher than what the original goal was. Their CSAT has gotten better. And strangely enough, their agent churn rate has come down even in that time because you've taken all the manual work out and moved a lot of the grunt work to AI agents. So that's just one example of how we're seeing customers use the full stack. And keep maturing along with it.
Bob Laliberte
>> Yeah, that's great. I love to hear those customer examples, right? It really brings it home to a lot of people who are watching. clearly we're still in the early stages of people adopting AI and really understanding how they're going to get all the value from it. So I'm wondering, as I wrap up, I like to look at, looking out a couple of years, how do you see AI transforming customer engagement? What should organizations be doing today to prepare for what's coming next?
Amit Mathradas
>> Look, I have a huge belief that the contact center 5 years from now is not gonna look anything like the contact center does today, right? And that is, and I hope a lot of your listeners are really grasping that and saying, well, yes, humans will be around, but what they do, how they service, there's a whole new world of customer memory that is coming to light. Let me give you an example. Tomorrow with the capabilities of an AI agent being able to listen to every single call. You can now start capturing customer sentiment, which is the biggest part of the interaction layer, right? If you have called in and we both have bought the same pair of shoes from the same vendor, same size, a CRM will capture that. The sentiment will capture, did I have a great experience with that agent when I was talking to them or did I not? And that can surface in the next engagement when you call back the next time, I can open it up and say, "I'm sorry, you were talking to Amit, he completely sucked, I'm gonna give you a $20 coupon." Well, these new capabilities around the platform, around how contact centers are going to evolve are new. So that's one area it is going to pivot. The next big thing is around how these systems of record all work together. To actually enable the next shift that is coming, right? If the interaction layer, digital, voice, AI coming together is going to create these new experiences, well, you have to be able to connect to the right systems of record to enable the orchestration to take place. What are you driving with these outcomes? So orchestration is the next big layer that I think a lot of customers should be thinking about and companies like us are thinking about in terms of where it needs to go.
Bob Laliberte
>> I think, I think those are really valid points. It makes a lot of sense. Amit, thank you so much for joining us today. Zeus Kerravala, thank you for co-hosting.
Amit Mathradas
>> Thank you both. Yeah, really enjoyable.
Bob Laliberte
>> Absolutely. Well, it's really clear that AI is moving beyond experimentation, beginning to deliver measurable business outcomes across customer experience. The organizations that combine the right technology with the thoughtful implementation and a focus on business outcomes are likely to realize value much faster than those approaching AI as simply another automation project. So I wanna thank everyone for watching. If you enjoyed the conversation, be sure to explore the rest of the sessions from theCUBE Research Contact Center Summit, where we're talking with industry leaders about how AI is transforming customer engagement, operations, and enterprise technology. For Zeus Kerravala, I'm Bob Laliberte. Thanks for joining us, and we'll see you next time.
Bob Laliberte
Zeus Kerravala
