In this interview from theCUBE + NYSE Wired: Mixture of Experts - AI AGENT Conference 2026, Jonathan Corbin, chief executive officer of Maven AGI, joins Sami Shalabi, chief technology officer of Maven AGI, to talk with theCUBE's John Furrier about reimagining enterprise customer experience by replacing siloed workflows with unified AI agents. Corbin explains why the real step function in value emerges not from layering AI onto existing processes but from rethinking the entire customer journey — from marketing through support — as a single, orchestrated experience. Shalabi details Maven AGI's no-copy data strategy, which connects directly to enterprise systems of record rather than duplicating petabyte-scale data, enabling deployments in weeks rather than months.
The conversation also explores the evaluation framework Shalabi built drawing on his experience at Google Search, which provides full traceability and explainability to eliminate hallucinations and hone in on the 35 APIs — out of a typical enterprise's 30,000 — that solve 90% of use cases. Corbin breaks down the three pillars every agent needs to deliver accurate answers: knowledge, context and actions, illustrating each through customer examples at Rho and Clio. Both founders highlight voice as the breakout modality of the year, while emphasizing that the future is not a single channel but a multimodal, always-available presence with persistent memory. From proactive engagement that resolves issues before customers notice them to a low-code platform that puts business users in control, Corbin and Shalabi provide a practical blueprint for how enterprises can move from chatbot deflection to genuine resolution at scale.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
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 this link to automatically sign into the site.
Register For theCUBE + NYSE Wired: AI Agent Conference
Please fill out the information below. You will recieve 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 theCUBE + NYSE Wired: AI Agent Conference.
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
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
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 this link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: AI Agent Conference
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: AI Agent Conference. 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
Jonathan Corbin & Sami Shalabi, Maven AGI
In this interview from theCUBE + NYSE Wired: Mixture of Experts - AI AGENT Conference 2026, Jonathan Corbin, chief executive officer of Maven AGI, joins Sami Shalabi, chief technology officer of Maven AGI, to talk with theCUBE's John Furrier about reimagining enterprise customer experience by replacing siloed workflows with unified AI agents. Corbin explains why the real step function in value emerges not from layering AI onto existing processes but from rethinking the entire customer journey — from marketing through support — as a single, orchestrated experience. Shalabi details Maven AGI's no-copy data strategy, which connects directly to enterprise systems of record rather than duplicating petabyte-scale data, enabling deployments in weeks rather than months.
The conversation also explores the evaluation framework Shalabi built drawing on his experience at Google Search, which provides full traceability and explainability to eliminate hallucinations and hone in on the 35 APIs — out of a typical enterprise's 30,000 — that solve 90% of use cases. Corbin breaks down the three pillars every agent needs to deliver accurate answers: knowledge, context and actions, illustrating each through customer examples at Rho and Clio. Both founders highlight voice as the breakout modality of the year, while emphasizing that the future is not a single channel but a multimodal, always-available presence with persistent memory. From proactive engagement that resolves issues before customers notice them to a low-code platform that puts business users in control, Corbin and Shalabi provide a practical blueprint for how enterprises can move from chatbot deflection to genuine resolution at scale.
In this interview from theCUBE + NYSE Wired: Mixture of Experts - AI AGENT Conference 2026, Jonathan Corbin, chief executive officer of Maven AGI, joins Sami Shalabi, chief technology officer of Maven AGI, to talk with theCUBE's John Furrier about reimagining enterprise customer experience by replacing siloed workflows with unified AI agents. Corbin explains why the real step function in value emerges not from layering AI onto existing processes but from rethinking the entire customer journey — from marketing through support — as a single, orchestrated experi...Read more
exploreKeep Exploring
What has changed in the past few months regarding whether agents are real and how they fit into strategy?add
What has changed in the past few months regarding AI agents, and what challenges are companies facing in adopting and implementing them?add
What execution gaps and blind spots do organizations commonly encounter when establishing and scaling agentic infrastructure, and how are you addressing those issues in your product and data strategy?add
How do you manage and use petabyte‑scale, distributed enterprise data for AI applications without having to duplicate and copy it into a single store?add
What are the implications for customers if AI agents replace humans as the reasoning layer between front-end interfaces (e.g., chatbots) and backend systems (e.g., Salesforce)?add
What is required to design, measure, and deploy an enterprise-grade intelligent agent system that delivers accurate, explainable, and automatable outcomes for complex use cases?add
What significant shifts in user expectations have emerged with AI—particularly around voice interfaces, multimodal experiences, and continuity of context/memory?add
>> Welcome back here. I'm John Furrier, host of theCUBE in our NYSC studios here for theCUBE, of course, part of the NYSE Wired program. A CUBE Original, also an open community, also a preview of the Agent Conference. Simon Chan and his community growing very fast. Builders, investors, people bringing in the agent era with force and a lot of value creation. We've got the founders of Maven AGI here. We got Jonathan Corbin and Sami Shalabi, CEO and CTO. Great to see you, Sami. Good to see you. Welcome to theCUBE.
Jonathan Corbin
>> Thanks for having us.
John Furrier
>> Love when founders come in because it's like the founders, they're the ones spilling the blood, eating glass, spitting nails.
Sami Shalabi
>> Blood, sweat and tears.
John Furrier
>> Blood, sweat, and tears. Get the idea, build a company, get the beachhead. Pretty good market for the agents right now. I think the Agent Conference growth that Simon's team are putting on shows that it's still a small community of innovators. And you guys are leading that. So one, thanks for coming on. Appreciate it. My first question is that what's changed in the past couple months because you're starting to see the strategy of, well, are agents real too, agents are real. People have been doing it for a year. We talked about it last time, Sami, when you were on and took advantage of it. But now it's even more real, but now it's not a strategy problem, it's an execution challenge. And people who get it right get the benefits because you got to do a lot of grinding. Compliance, privacy, identity. There's a lot of blocking and tackling. If you do that right, then the agents can really go to challenge, so to speak.
Jonathan Corbin
>> I think there's a couple of challenges that companies are facing right now. One, you mentioned it, security and legal. What the heck happens if things don't go as planned? The second is that you have AI agents that are capable of doing so much. And there is a natural fear that, "Hey, is this going to replace my job?" It's not clear for everybody how AI is going to help them evolve to a new role that allows them to do so much more. And I think that's one of the things is the change management and the human perspective is something that people don't talk as much about. And so one of the things that we do pretty frequently with our customers, we actually go in. We have a conversation about what is it that you want to do with AI agents in terms of creating a customer experience, you start with your customer and then the technology and the people all have to be part of that conversation.
John Furrier
>> What are some of the gaps on the execution that you're seeing or what are some of the areas people either see and don't get too fast enough or blind spots in going at figuring out how to establish and scale an agentic infrastructure solution?
Sami Shalabi
>> Yeah. One of the things we constantly see is people pretty much bring in AI agents to improve an existing workflow as opposed to stepping back and rethinking the entire process. You start to get incremental value, but the real step function emerges when you're reimagining the entire workflow in a way that's driven by agentic capabilities.
John Furrier
>> One of the things we talked about last time was a lot of the data. It was, no copy data strategy was a big discussion point. Explain what you guys are doing, because I want to set the context of a deeper dive, which is you guys taking a pragmatic approach to agents. Take us through the product, the successes you've had, where you guys are at.
Jonathan Corbin
>> We're building AI agents for customer experience, and so we're working with some very large companies and we'll be announcing some more in the relatively near future. But when you look at the customer journey of today, for most companies, you have silos that have been created. You want to go buy a product, you're informed about it by marketing. You want to purchase it and you swipe your credit card or pay for it, you have to talk to the sales team. You have to go through these multiple steps in order to be able to access the product to use it and to get support with it. And so when we looked at, what's creating that structure that we have traditionally kind of had where we have these silos of systems and data and people that are really inhibitive to a great customer experience. And so we looked at that and said, there's an opportunity for us to fundamentally change that with AI agents. What is the customer experience of the future going to look like? It's going to be proactive. It can engage with you, understanding who you are, creating this very personalized experience, and then reactive, I can ask it questions and it would do things for me. And so we've kind of reimagined that entire customer experience from end to end using AI agents that are capable of engaging and forming, interacting and taking actions on your behalf.
Sami Shalabi
>> And then from the data perspective, which is the question you were asking us, John, is that when you're entering a large enterprise, this is petabyte scale of data. And there's data in lots of systems of record, there's a lot of duplication and having an approach where you have to duplicate that data so an AI understands is incredibly prohibitive. Any large enterprise meeting, it's probably a nine-month project just to even copy it-
John Furrier
>> ....
Sami Shalabi
>> and by the time you're done, it's inaccurate. So we've taken an approach where your data stays in place and we are connecting to all the systems and orchestrating across all the systems and built a lot of IP around finding the right information to solve the problem that the user is looking for across the entire go to-market experience.
John Furrier
>> I love that angle because data gravity can kill projects. We see that all the time. And the other thing that's been in the news the past month, everyone's been talking about, of course, I wrote three posts on it when it came out, the SaaS-pocalypse, which was the thesis that AI is going to kill software. What you just said, Sami, is basically the moat. Enterprises have preexisting infrastructure. Now the interfaces can change because you're in the customer experience business with AI native. Okay, customer expectations comes from their experience and what they expect, that's been like a first principle in all UX since I could remember. The customer expectation becomes what they want and/or what they see. So I can see the front end certainly changing, call it the front end/backend, but these backend moats has been largely ignored by the SaaS-pocalypse argument. They just say, "Oh, someone's going to build Salesforce." Yeah, I could technically replicate Salesforce and have it look like the same, but where the hell's the data? Or I could start a Reddit, CUBE-Reddit. It's open source software, I can just say, "Here's our Reddit threads. We're the people." So you can't just copy ideas with software. I bring this up because I think the user experience piece is the most critical because that's what AI has shown people, I can go from a chatbot to a fully reasoning coworker, not Copilot. I could have a better answer based upon the data access. This is what you're doing. Explain that piece. I think this whole UI, UX, old school first principle can be applied to AI native.
Sami Shalabi
>> I'll give you a little bit of my perspective on this is that if you actually look at the history of SaaS software, most of the software has been really built around this concept of, how do I route work to people? So your support systems, a ticketing system is of, I get a question in German, it goes to the person who speaks German. If I'm asking a question about a technical piece, it goes to an engineer. And so if you look at a lot of the customer experience, it's really been around building company centric experiences about organizing human work. For the first time in software, agentic experiences have unlocked the ability to have one user experience that transcends the entire customer journey, that's able to orchestrate by writing zero code. Because of that, you can start to change these silos. And when we think about it from a user experience standpoint, the experiences are no longer my support experience, my pre-sales experience, my success experience. It is a unified experience that is both proactive, meaning you're reaching out to people before problems even happen when you detect them, reactive, which is kind of like what we think of support today that is interacting across the entire customer journey.
John Furrier
>> That's why I like the fact that you got AGI in the name, even though people will debate what AGI is. What you just said is interesting because like Jensen's been on stage at NVIDIA, GTC for now three years, the past two years in particular, hardcore explaining that this is not a static internet anymore.
Sami Shalabi
>> Oh, no.
John Furrier
>> And what you were just saying about this path goes here, those are rules. Those are rule-based systems. It's just too big. So what we're getting at is intelligence is being injected into the system. If you assume that it's going to be some sort of intelligent system, AI, at a platform level, then it's easy to say, "I will have a generative front end." That's what you guys are doing. Did I get that right?
Jonathan Corbin
>> Yeah. I think-
John Furrier
>> Or am I off?
Jonathan Corbin
>> Kind of, right?
John Furrier
>> Yeah, okay. So ...
Jonathan Corbin
>> And so when you think about it like, okay, wait, if we were to think about the layers to the existing systems today, you have your ... You mentioned Salesforce earlier, right? So we have Salesforce and then we have a front end for that. And you have a reasoning layer between the front end, which might be a chatbot and your Salesforce. That's people today. And so when you think of, what does the evolution of that look like with AI agents in play, all of a sudden, as opposed to humans being the reasoning layer, the people are actually giving AI the information it needs in order to be able to follow the thought patterns that they used to use in the past.
John Furrier
>> So what does that mean for the customer? Because I love the abstraction layers, I agree 100%. You can actually build abstractions and leave the existing stuff where they are, so you kind of take the silos away, but you're not really taking the silos. There's storage. At that point, it's just storage.
Sami Shalabi
>> I mean, from a user experience standpoint, you end up with a ... I mean, one thing when we think about the user experience is that why do customers get frustrated with any ... Like anyone I've ever spoken with, they hate their support experience. And the reason is, one is it's slow, but what really destroys it are the handoffs. You go from person to person. It's like nobody has ever known who you are and it's just this repetitive experience where it takes many, many turns to be able to resolve someone's issue. And in this new world, when you kind of blur all the different go to-market functions into one unified experience, you end up with an experience that your answers are being responded to in one turn, first question and you immediately get the answer, zero handoffs because AI is able to remember all your historic interactions. And if you do get the need to talk to a human, the handoffs are contextualized where everything the human agent needs to know is handed off. So you end up with a very nice user experience that is seamless. And then the other thing is that we're seeing a tremendous amount of variability of how people want to interact, some people like to use voice, some people like to use chat, some people like to use email, and the modalities that they're looking to interact when they're trying to get an issue resolved change as the day progresses.
John Furrier
>> It's like the old endpoint discussion. Hey, what is the end point? If it's email, then give a great email answer.
Sami Shalabi
>> But the thing is, it's not just what the endpoint is, the endpoints change. You can start an email, move to voice, move to chat in one seamless ... It's one conversation.
John Furrier
>> All right. So explain you guy's philosophy on the agent and multiple agents because there's two things going on, you got intelligence, got reasoning, you got the abstraction, you got access to all the data, but there's been a discussion around multiple agents, agents for agents. And then in this community for agent conference, a lot of MCP, A2A. So you're starting to see the infrastructure piece get really worked on. We had Chase Harrison on from LangChain. We had Arcade on, the founder of Arcade. There's only a handful of people really driving this right now. So as builders, what's going on at the infrastructure level? Because when you nail that agent to agent, fleet of agents, teammates, coworkers, all kinds of words for it, how do you guys view this whole agent landscape? Chaos, structured?
Jonathan Corbin
>> I think it's interesting, Sami is an engineer. And so if you give Sami a problem, I'm going to speak for you for a second, Sami.
Sami Shalabi
>> Yeah.
Jonathan Corbin
>> He's going to break it down into smaller pieces. He's going to say, "Great, what are the components that have to be solved in order for me to get to the outcome I'm looking for?" And so when we're talking about agents and agentic frameworks and things like that, what we're talking about is like, what is the problem? How do we break it down into smaller pieces so we can make sure that's right and then have something checking to make sure that what they're coming up with is accurate. And so when we looked at that, I think we started working on something like this about two years ago. We said, "Hey, this is something that's going to allow us to be able to move incredibly fast and meet people's needs while at the same time ensuring that the answers are high quality." And that's what people care about. At the end of the day, the customers we're talking to about, they don't care about whether it's agentic or how you et it.
John Furrier
>> Yeah, what's the outcome?
Jonathan Corbin
>> How do you get me the right answer fast?
John Furrier
>> I don't want to reduce my churn. I want to have happy customers. That's right. I want people to self-serve and get answers they want.
Jonathan Corbin
>> That's right.
Sami Shalabi
>> And in order to answer a question, based on our analysis with enterprise customers, it is typically any agentic experience needs to communicate with at least six to nine different systems, at least, if not more. The CRM, the billing system, the contracts, the telemetry, maybe some product stuff. And the fact that it is a distributed orchestration is the problem that needs to be solved. Now, these protocols, A2A, MCP are a step in the right direction. What we've found is that they're on the way there, they're still not ready for true enterprise use cases, even though we see a lot of folks supporting them. But I mean, we embrace standards, we participate in them. We're actually participating with the A2A standards committee to help elevate them to actually solve real world, real use cases, real scale, real complexity-
John Furrier
>> And you've got to put a plug in for Linux Foundation-
Sami Shalabi
>> Oh, absolutely.
John Furrier
>> Both of those.
Sami Shalabi
>> Yeah, we're members of that working with them to evolve the standards.
John Furrier
>> So props to the Linux Foundation, big supporters of them. Let's talk about the workflows piece because one of the things I learned last week in Barcelona from Mobile World Congress or MWC, I was talking to the senior engineer at AT&T. Now it's a telecom example, so bear with me. I'm like, "How's it going?" It's like, "Well, we found that the frontier models didn't speak telecom." So they were trying to use the frontier models to create these agents for people in the field to fix towers and stuff. They just didn't have the crawl, they didn't have the data. So they donated 19 models, it was a big to do. I'm super pumped to support that. So they basically donated 19 small or specialized language models to the frontiers and to open source. What that talks to is the fact that they had unique requirements for AT&T. Now the enterprises have a similar thing going on where this frontier data, but I have my own workflows, I have my own data that was not yet, it's still locked in their premise. So you start to see the enterprise side. How do you guys see that? Because you guys have been doing this for a while with enterprises. How are they reacting? Oh, by the way, and Mark from AT&T said, I asked him, "Is it working?" He goes, "Our agents are working super great." I said, "What was the key to success?" He said, "We grinded out," my word, he didn't say grind. He grinded out the identity, privacy, and compliance because Intel comes with a lot of departments of no, "No, you can't do that." So he had to actually just grind it out and then the agents were highly effective. He had the workflows, they were known workflows and they were just really crushing the performance because he did the work. There's no shortcut. So comment on enterprise example of they got their own models, they got to grind out some things.
Sami Shalabi
>> Yeah, so this is the thing a lot of people don't talk about is if you're trying to build an intelligence system, you need a way to measure its intelligence and then use that as a feedback loop. One of the first things we built as part of our agentic platform was an evaluation framework inspired by a lot of the work I used to do at Google when we're kind of building Google Search and Google News. And it's one of those things that people don't talk ... You cannot make anything intelligent, you cannot fix an AI system in any way, shape or form if you cannot measure it. And measuring it is how we have been able to deal with no hallucinations, the ability to actually validate that it's doing what we think it's doing. And then as part of ... So that's the first part of the evaluation framework. The second piece was actually we spent a lot of energy on the explainability. So whenever you get an answer, so if you see an error, why? And in order to resolve it, you need to understand what data it's using, where it's getting the information, how it's reasoning across that, why it's calling this system versus that system, so that you can tune it to achieve that. So this is why in our case, we have been consistently with all of our customers achieving 90 plus percent automation rates because we're focusing on where the problems are.
John Furrier
>> You designed it in?
Sami Shalabi
>> Yes.
John Furrier
>> You designed basically-
Sami Shalabi
>> It has to be designed in.
John Furrier
>> When you say measurement, is that like observability or what does measurement mean to you? Just tracking movement?
Sami Shalabi
>> So it's a couple of things. So we need a framework to understand reasoning, this answer came from this paragraph, from this API call, from this data source. So overall traceability.
Jonathan Corbin
>> And you think about what's required in order for you to give an accurate answer to an enterprise customer. You talked about, they had to train small language models in order to capture all the information. What we found in enterprise is there's three things you have to be able to do. First, you have knowledge, what is a product for? Why does it exist? How does it work? The second thing you need is context. So one of the banks that we work with, a new bank called Rho, they had customers coming to them and they said, "Hey, what's my balance?" You can't answer that question if you don't have context. And so the second thing you need is context. Okay, context in this case would be, you have three accounts, you made a deposit in your savings account two days ago, now I can answer that question. The third thing that you need is the ability to be able to take actions. And so when we looked at-
John Furrier
>> And you've got to know who they are too.
Jonathan Corbin
>> Correct.
John Furrier
>> Because we've seen people do -
Jonathan Corbin
>> Because those haven't been -...
John Furrier
>> a software engineer and rewrite the code for them, they're like, "Okay, I'll do it. Be my account manager." I'm foreign.
Jonathan Corbin
>> So you're right, you talked about security, it has to start with that. Identity and security, those are first and foremost. You have to know who the user is, you have to be able to understand context around them, the knowledge around the products they purchase, and which actions do they have access to. Now we can go out and create that personalized experience that we were talking about earlier.
Sami Shalabi
>> And then the evaluation framework allows us ... When you walk into an enterprise, there's 30,000 APIs. Which ones do you integrate with? 30,000, that's two years of work. So because of the evaluation, we hone in on the 35 that are necessary to actually solve 90% of the use cases. So it's a very ROI driven, very practical, very fast journey to impact.
John Furrier
>> I love that you brought up the Google experience because in the search business, as you know, everything is contextual and behavioral, ads and results, kind of the similar thing going on with agents. You got to know the context. There's behavior, workflows, and then the architecture really, what you're getting at is, understand the system, identify the key levers, work there. Give some examples of where you guys have done this so people can understand how to take that on. What would be the approach? Walk us through a day in the life of a customer.
Jonathan Corbin
>> Yeah, I think a really good example is a company called Clio that we work with. It's a legal tech company. I think they're really forward-thinking in terms of the experience that they want to design for their customers. And so they went out, they looked at the market and they said, "Hey, a lot of players are in this space, how do we know who's the right one?" They evaluated 32 different players and they said, Maven is the best product for complex use cases and that's what we have.
John Furrier
>> Talk about your target. That's a good point. Are you guys targeting large enterprises? What's the target market?
Jonathan Corbin
>> Yeah, we really focus on complex use cases. And so for a-
John Furrier
>> What would be a complex use case?
Jonathan Corbin
>> Complex use cases is multiple systems. And so we were talking about earlier, what's required in order to create an accurate answer for a customer? You have knowledge, you have context, you have actions. Now for every single question that's asked, you have to orchestrate across all those. You have to know which systems am I pulling knowledge from? What's the piece of information that I need to be able to provide the outcome, the answer here? Is there additional context required to answer that, and where am I pulling that from? And which actions of hundreds of possible actions you should take is the appropriate one? And so you can see how it gets incredibly complex.
John Furrier
>> It's a system. It's a system.
Jonathan Corbin
>> And for a lot of the other players out there, they're very good when it comes to, let's put a chatbot on top of a knowledge base, great, I can give you answers. When it comes to complexity, there's very few players who are capable of doing what we're doing. And so you were asking who is our ICP? And we were talking about Clio, legal tech, it doesn't get more complex than that.
John Furrier
>> Yeah. A lot of database.
Jonathan Corbin
>> You really have to get your answers right and you got to-
John Furrier
>> A lot of structured and unstructured.
Jonathan Corbin
>> That's right. Yeah. And so they came to us and they said, "We want to redesign our workflows. We want to create a customer experience that is second to none."
John Furrier
>> All right, take me through an engagement because you guys are taking obviously a horizontal scale approach to the enterprise. You mentioned silos earlier. I'm imagining that there's a lot of work going into identifying those APIs. Is this a six month, six week, six day pilot? Take us through, what's the engagement look like? Do they start with some use cases? I mean, you don't want to get too narrow and say, "Okay, throw a RAG system on a chatbot." I mean, that's easy, but like, okay, I want answers that pull from ... Is it a database problem? I mean, take me through the engagement. What does a customer engagement look like, a prospect?
Jonathan Corbin
>> Yeah, it really kind of depends on where they're at in terms of their evolution internally. Sami talked about 30,000 APIs. Some people would like to connect 30,000 APIs to it. So the first thing is understanding what is it that we want to achieve here, and then how do we get started? For the majority of our customers, they're up and running in weeks. They sign the contract, we're up and running within weeks-
John Furrier
>> That's fast....
Jonathan Corbin
>> in front of their customers.
John Furrier
>> I mean, compare that to what it used to be. I mean, if this was like 10 years ago, that's a six-month POC.
Sami Shalabi
>> Yeah.
Jonathan Corbin
>> Well, I think there's a lot of players out there that are really focused on the services aspect, they're going out and doing custom builds that are really challenging for them to maintain over the long term. We created a system that allows us to be able to be up and running at a very rapid pace.
Sami Shalabi
>> And it's because of this concept of data in place. We connect, we don't try to duplicate. So it becomes like, "Okay, where's the most impact? What are the key integration points?" Connect, connect, connect, connect. And we're able to connect with legacy systems, modern systems, whatever it is, and we have an architecture-
John Furrier
>> Do you find, Sami, that they're doing it to get the cost reduction out first or get efficiencies or is it top line revenue based or both?
Sami Shalabi
>> It depends on ... I mean, for a lot of our ... When it's a complex use case, usually our customers have very high LTVs. So they're much more concerned about quality of experience than they are about cost reduction. Cost reductions are inevitable, but at the end of the day, many of them are measuring CSAT scores, actual upsells, deeper integrations. And so-
John Furrier
>> So velocity is a key thing for them. And that's one thing that's coming out of the agent conversation is that what used to take six months, you can do in six days, six years. So the timeframe with the compute available, there's a lot more gettable things.
Jonathan Corbin
>> That's right. Well, what it allows them to be able to do is because all of a sudden your cost per interaction has gone down, I can engage with a larger swath of my install base, right?
John Furrier
>> Yeah.
Jonathan Corbin
>> I can make sure their questions are being answered. I can introduce them to new products. I can have real conversations about what their use cases are and how they can do a better job of monetizing my product. And so what we're seeing is people saying, "I actually want to reinvest in my customers with the cost savings that I haven't realized. I want to talk to them more frequently." CSAT scores are going up, used to just going up, cross sells, upsells are all being realized as a result of the interactions that they're having with their customers.
John Furrier
>> Well, guys, it's great to see you and I love the name, Maven. I love that AGI's in there. Dave and I always debate on AGI on theCUBE pod, whether it's here or not. In fact, Ali Ghodsi from Databricks said, he had a good quote, he said, "If you asked me four years ago what AGI was, I'd say what's in the frontier models is AGI, because it's so powerful." Even today, it's still scratching the surface. Talk about some of the things you'll be talking about, Jonathan, at the agent conference. What's the talk? What do you expect? And what's the buzz? Because it's a lot bigger.
Jonathan Corbin
>> Yeah. I think for us, there's really kind of two areas where we're seeing meaningful impact from AI agents today in the enterprise world. One is on coding. Sami's crew is, I think, using everything out there and we're developing some of our own stuff in order to celebrate our growth. But the other area is in customer experience. And so for a lot of companies out there, they're focusing on aspects of the customer journey that are like, "Hey, how do we deflect the conversations and the questions our customers are asking us?" So what we're really focused on, making sure we're interacting and engaging with the customers and resolving their questions. And so that's one of the things we're going to be talking about, "Hey, what's the difference between a deflection and a resolution?" Well, one is you get to the door and they say, "Sorry, you're not allowed in." A resolution is like you're walking through and you're a pretty happy customer. So that's one of the things that we'll be talking about there.
John Furrier
>> How would you explain, maybe Sami, you can chime in if you want too, on this one, because it's kind of a generic question I mentioned earlier at the top, that first principles of UX was expectations and as things shift, so do expectations. What is the big expectation shift that you guys seen with AI? Voice, answers?
Sami Shalabi
>> So I think this year is going to be the year of voice. We've seen a step function and capabilities, and it's a very different beast compared to chat because of the latency requirements. People are expecting very natural speech. We believe we've cracked it and seeing tremendous results and we are at scale and deployed and we're seeing some interactions where the people are actually saying thank you and are treating it like a human-like experience. And as a-
John Furrier
>> We've all yelled at our Siri, "Bad AI."
Sami Shalabi
>> But as a user experience, I mean, I would argue we're still in the early days of deciding what a great user experience ... And I'm not saying answering a phone call, I think voice is going to be much more integrated into actual products. I mean, we're seeing this in the coding world. I have some engineers on it who are literally talking to their computer all day as a means to drive the development of a-
John Furrier
>> They're . They're artists at this point. It's almost like a craft, back to being a craft, it's interesting, but if you go back to the early days of the web, some of the simple answers were on user experience was make it simpler, easy to understand and reduce the steps it takes to do something.
Jonathan Corbin
>> That's right.
John Furrier
>> Versus click, click, click, click. Now you have AI. Similar, again, these first principles don't really change.
Jonathan Corbin
>> No, no they don't. For me, I think it's great that we're evolving from a voice perspective. I don't think that the future is a single modality. I think it's AI that is present, that's available when and where I want it. If I want to talk to it on my Dell laptop, if I want to interact with it through an app on my phone, if I want to call it, if I want to text it, I get the same response, the same capabilities and the same outcomes that I'm looking for. It's multimodal, it's with you-
John Furrier
>> Multi-platform....
Jonathan Corbin
>> and it has a memory.
John Furrier
>> Yeah. Memory's the key because this is where the intelligence comes in, because the context switching between, "Okay, rule, put packet A over there, move things around," is gone. It doesn't really scale. Well, the user is on email and then now they're in their car, which they can't really type, or they're typing at a park, walking around, going to the store, all these contexts, situational awareness.
Jonathan Corbin
>> What we saw historically was if you were going to interact and engage with the brand, it was on their terms. "I'm open between 9:00 and 5:00 and you can email me if you have a question," and that sucked. And what we're seeing is the evolution of the customer experience with an agentic platform allows you to be able to say, "Actually, I can meet my customers where they want to be met." I can provide them with the channels they're looking for and the evolution of that experience becomes this very personalized engagement that you can engage and interact with.
John Furrier
>> It's like the old metaphor, is it an aspirin or a vitamin? The aspirin is call center. Everyone hates call centers because I want to actually get faster, I don't want wait times. So that's an easy low hanging fruit. I think we talked about it last time. The vitamin is the new use cases that surprise and delight. That's where I think is the opportunity is where I think we'll see new cases.
Sami Shalabi
>> Oh, absolutely. And I think it'll get accelerated with proactive use cases because most of the time you're calling because of the problem that you have observed. If you can get in front of it and turn it into a surprise and delight moment, all these negative experiences suddenly become positive experiences.
John Furrier
>> Well, the great news with the low code, no code platforms like you guys, you don't need engineering, a business user can come in once you're set up and they could iterate on the fly and generatively, that was process.
Sami Shalabi
>> That was a core design center for the actual product. Yes, we have all these AI agents that work with our customers' customers, but we have to build a platform to enable business users to manage these things at scale. The world is changing, products are evolving, the things they want to communicate are evolving and they need a mechanism that isn't engineers writing code.
John Furrier
>> All right, we've gone 30 minutes. A great conversation. Again, it should be a great event in May agentconference.com. Get the final word in, Jonathan. Some highlights on the momentum. If you can share some stats, how many customers, what's the revenue? How big is the team? What's your focus?
Jonathan Corbin
>> Yeah, we're about a hundred customers right now and this year is going to be amazing. We have some really amazing customers who we're going to be announcing shortly. So excited to be able to make-
John Furrier
>> Funding, levels?
Jonathan Corbin
>> Funding, our last raise was a $50 million raise led by DTC. And so we've got lots of cash in the bank and we;re continuing to grow rapidly.
John Furrier
>> You've got customers.
Jonathan Corbin
>> Yeah, I'm really excited by the customers. I think being able to go to the customers, see the impact that they're having, the improvement in CSAT scores and hearing from their agents, "Wow, we're so thankful for you bringing Maven and helping us improve the customer experience that we're delivering."
John Furrier
>> It really must be fun as founders to look at a customer and see the magic happen because they're seeing stuff for the first time that like, "Whoa, whoa, that happened." Because they're, in their mind says, "Oh, I got to call IT. I got to build an app on the cloud." So when you start to see the value to the end users.
Sami Shalabi
>> Something we do with a lot of our customers, because everyone kind of has their perception of the world, we do the art of the possible. And we spend a whole day with them to just loosen their thinking and start to expose new use cases because this technology is transformative and it can fundamentally change how you think about how you operate your business and we're just seeing tremendous impact as a result of that.
John Furrier
>> Guys, thanks for coming on theCUBE, really appreciate you. Have a good conference. Thanks for coming on, John. Great to see you.
Jonathan Corbin
>> .
John Furrier
>> Thank you. Sami, great to come back. I'm John Furrier, host of theCUBE. This is a preview of the agentconference.com event that Simon Chan and growing communities putting on, of course, part of our mixture of Expert Series" Agents in the Real World. Thanks for watching.
John Furrier