theCUBE + NYSE Wired: Mixture of Experts Series | Raju Malhotra, Certinia & Scott Hebner

Raju Malhotra, Certinia & Scott Hebner

In this theCUBE + NYSE Wired: Mixture of Experts segment from the New York Stock Exchange, theCUBE’s John Furrier sits down with Raj Verma, CEO of SingleStore, to unpack how the intersection of technology and finance is shaping enterprise strategy. Verma shares why SingleStore is “on course” for the public markets, reflects on brand-building through the company’s partnership with golf Hall of Famer Padraig Harrington and connects that ethos to how SingleStore helps organizations fix struggling data “swings.” The discussion zeroes in on what’s next as Wall Street watches the AI infrastructure buildout: after chips and systems, the software and data layers set the pace for value creation. Verma outlines why enterprises must modernize “brown” data estates into “green” ones to safely bring corporate context, governance and compliance into LLM workflows via RAG – and why commoditized data-at-rest puts the advantage at the query layer that unifies data in motion with data at rest. He predicts agentic AI will gain reasoning capabilities in roughly 18 months, cites industry indicators like Google reporting ~25% of its software now built by AI and argues that high switching costs will give way to disruption as buyers reassess legacy vendors. The conversation closes with concrete momentum: ~33% YoY growth, ARR in the ~$135M range, gross dollar retention ~98%, cloud NDR ~130, ~50% of business now in the cloud, landing ~3 new customers per day, a path to cash-flow breakeven in the next two quarters and a teaser for AI-related announcements in the next two months. Listeners will find notable stats, real-world use cases and forward-looking views on how databases power reliable AI at enterprise scale.

Share this session

Clips
More from theCUBE + NYSE Wired: Mixture of Experts Series

Scott Hebner

Principal Analyst, AI

theCUBE Research

Raju Malhotra

Chief Product & Technology Officer

Certinia

play_circle_outline Certinia as growth engine for services-oriented businesses selling, delivering, managing services

play_circle_outline AI disruption seen as opportunity; potential net job growth via augmentation

play_circle_outline Veta: suite of packaged agents and intelligent actions for workflows

play_circle_outline Native Salesforce integration and Agentforce partnership enable tight data and AI coupling

play_circle_outline RFPs Require Open Standards for Enterprise AI as Veta's 64 Invocable Intelligent Actions Enable Cross-Platform Workflows

Info

Raju Malhotra, Certinia & Scott Hebner

Scott Hebner

Principal Analyst, AI theCUBE Research

HOST

Raju Malhotra

Chief Product & Technology Officer Certinia

search

Gemma Allen

>> Welcome back to theCUBE studio here at the New York Stock Exchange. This is Mixture of Experts, one of our programs with NYSC Wired, and today we have a special segment. We're going to have a conversation on two topics that we talk about day in, day out here on theCUBE, and that is the AI factory thesis, creating AI production infrastructure outcomes at scale, and matching that with business outcomes that compel enterprise users and buyers. Joining me now is Scott Hebner, one of our key analysts at theCUBE Research, and Raju Malhotra, Chief Product and Technology Officer at Certinia. Welcome, folks.

Scott Hebner

>> Thank you.

Raju Malhotra

>> Thank you.

Gemma Allen

>> So Raju, I'm going to start with you because Certinia, talk to me a little about the value proposition of this company, especially from the perspective of really creating business outcomes in the world of AI and agentic.

Raju Malhotra

>> Yeah. So thanks for having Certinia on your show. Thanks for having me. Certinia, think of as a growth engine for services oriented businesses. Maybe one of the ways to think about what we do is to think from the perspective of our customers. On one hand, we work with customers like IBM, PWC, KPMG, companies that have a large pure place services business. And then we also work with companies where services is a big part of their overall business. So companies like Salesforce, Google, AWS, Philips, HPE, Cisco, et cetera, et cetera. And they have services, businesses as part of their overall solutioning. And for all of these customers, we help them sell, deliver, and manage their services business effectively.

Gemma Allen

>> And Scott, you've been doing a lot of research on this company, but also this space, right? Talk to us about the macro-economic moment where there's a lot of noise, a lot of promise, and sometimes people feel not a whole lot of actual outcome. What are your thoughts?

Scott Hebner

>> Yeah, I think the macroeconomic part is what I would call digital labor transformation. It's more than just a technology trend that we see developing now. It's about digital labor. And what I love about what these guys are doing is they're taking the agentic technologies and they're applying it to a very labor-intensive segment of the marketplace, which is professional services. And it's always been the fact that there's been an effort to automate all this. And with agentic AI now, I think they're taking it to a whole new level because it's not just automating repeatable tasks, but it's also helping knowledge workers make better decisions. It's decision intelligence. And I think that's what you need to really bring to life the notion of digital labor, right?

Raju Malhotra

>> Yeah.

Gemma Allen

>> And when we think about knowledge workers and we think about the world of agentic, where you have agents working alongside you, agents working for you. And I mean, if we go with the fear-based thesis, agents potentially replacing you, right? What are your thoughts? What are you both seeing and hearing in the market, especially from your customers directly and your actual user base?

Raju Malhotra

>> Yeah. So first of all, I think this is a very cliched statement to make, but I'll just start something very foundational. AI is transforming every business. Interestingly, AI is transforming businesses where you have a lot of manual work, a lot of bespoke, custom, human focused sort of work. And immediately you kind of think about professional services is the most bespoke business that can be. So the disruption from AI on professional services is probably way more profound than it actually is feeling anywhere else. So you're right. I think there's a thesis of a little bit of a doom and gloom. Is AI taking my job? What's going to happen? We actually did this study and we hired a third party researcher to work with us last year. And what we found, there are a couple of very interesting things. One is that many of our customers and companies in the professional services are already using hybrid teams of humans and agents. Now what that means is that it actually opens up the aperture on the demand side to get the addressable sales pipeline up to six to seven times of what is otherwise possible. And you may wonder, why is that? It comes down to the whole nature of professional services it has been. You can't really sell something in professional services unless you have capacity to deliver it. And who's delivering it? It's the human beings that are delivering traditionally. And that comes down to, do you have the skills? Are you available in the right time zone? Do you have availability, et cetera, et cetera. And that constraint goes away to a large extent when you think about this as a hybrid team. So that's where I think our thesis is, yes, it's a disruption, but it's a very positive opportunity for customers who really take that disruption in the right way and use it to the advantage. So I think coming back to this, it is a fantastic opportunity for anyone in the industry that is working with the professional services, and particularly if you really not only adapt, but really lead that change.

Gemma Allen

>> Wow. And Scott, what are your thoughts, especially in terms of we think about things from the perspective of being an orchestrator in this world, right? Like the build versus buy thesis. There was a time when it was very straightforward, you bought SaaS tools, they created siloed, but very effective based outcomes for your business and particular practices worked really well in silo, right? But business grew and everyone was happy. We know that's not necessarily the future. What are your thoughts?

Scott Hebner

>> Yeah. Again, I go back to the digital labor thought, which is by definition, almost every business you can think of is built upon people working together and collaborating and orchestrating their work. And I think this is what technologies like Certinia is bringing to market here with beta is going to give these knowledge workers superpowers. And this theme that's been over sort of the cloud over AI for over a decade has been, it's going to take everyone's jobs. And our research shows that it's actually going to be a net job growth. In fact, I think there was just a poll out by Fox News where 59% of people thought it was going to eliminate jobs and only 7% thought it was going to grow. And I think what this is going to do is give more people, more knowledge workers, more superpowers. They can do more. They'll be able to do more jobs and people are going to invest those savings into strategic parts of the business. And in the end, I think it'll be a net job growth, right? And what you're doing is not only helping give these people superpowers, but getting everyone working together within a workflow.

Gemma Allen

>> Can we talk a bit about Veta? I was going to call it Veta. So I'm glad you came in there, Scott. But talk a little bit about what this announcement exactly entails. Give me the lowdown on what it is that you guys are building and growing on here.

Raju Malhotra

>> Yeah. So Veta is a suite of packaged agents and intelligent actions that we are making available to our customers. And think about, I think the transformation we've had as an industry, the traditional workflow automation-based SaaS was fantastic to work with human users, but that is sort of in the past because the age that we're moving requires us to use humans and agents working together. As Scott was saying, the digital labor working with the human teams. And that's where Veta comes in. It's really our thesis is that it's going to be a collaboration between human and digital workers, and that means we're going to need the workflow automation from software as a service, a lot of those capabilities, but we're also going to need agentic capabilities. We're going to need package agents that are role-based, that are specialists, that are trusted, they have the guardrails, they have the right capabilities, but we also have intelligent actions that can be invoked anywhere in the enterprise based on the workflows that they have. So it's going to be a combination of sort of the software as a service and the agentic and the way the launch that we have really calls out that differentiation that we are driving for our customers. And it's that agentic suite of those actions and the package agents.

Gemma Allen

>> Wow.

Scott Hebner

>> Yeah. I mean, I would add to that. One of the things I really love about what you're doing is, and I'm going to tee you on the word you used, which was trust. In all the work I've been doing, all the research, all the conversations, if I was to describe enterprise AI in one word, it would be the word trust. I think that's going to be the defining word going forward. And at the same time, an LLM only architecture, when you're building agentic workflows and agents, I think becomes a liability trap on its own. You have to build value around that LLM. And what you've done with context and logic and you've extended that architecture to make it more trustworthy. Just talk a little bit about that.

Raju Malhotra

>> Yeah. So you're, by the way, spot on. I think trust is not only a differentiation, it's a foundational requirement for anything to work, not just in enterprise, anywhere, even in a personal consumer way. I want to work with companies that are, I can trust, and I know my data and everything else is actually trusted. And you're right, we moved very fast from experimentation to being in production for AI. And we started out as LLM, as being able to solve a lot of those problems. And we very soon found out that's actually a very baseline. That's a very foundational requirement, but you really need some more capabilities that drive that trust. So first of all, trust in LLM is key, and we want to make sure I think the LLM providers we work with are trusted, are very transparent, auditable. But on the top of it, I think how we generate trust is really think about this sort of like a five layer cake. Who doesn't like cake, right? So it's about sort of the cake. And before I get into a little bit more specific deep dive, I think if you think about number one is the customer data, could be structured data, unstructured data. Second layer is about metadata and telemetry. That's the data about data. Third is really about a business logic layer, all the code, the actions that are needed. The fourth layer is about reasoning, which basically says which actions you should pick in what context. And then the fifth layer is about user experience. And the user experience could be in your workflow, could be in the Certinia app, could be Slack, could be a lot of different places. But if we think from that perspective, the bottom two layers really become context graph or context engineering because that's based on data and metadata. The next two layers become your business logic actioning and reasoning layer. And then the third part is about the actual, how the agents look like, how the actions surface, et cetera. That's the user experience part of it, which is very important. But it's really, I think the trust is built into all of them and the trust is built on the LLM that we really think about. And that's why we've taken the approach of having role-based agents, having specialist agents, having the actions, but all of this within the trust and governance guardrails that we have baked in.

Scott Hebner

>> Yeah. And I would say that is what is really unique here. LLMs are good at automating repeatable tasks.

Gemma Allen

>> For sure.

Scott Hebner

>> And they're as good as you want your future to be the same as your past, right?

Gemma Allen

>> Yeah.

Scott Hebner

>> When you move into the world of knowledge work where people are making judgments and they have to understand context and they have to know what other people are doing, that's where you need this extended architecture. And you're setting a model for where I think the industry needs to go because the vast majority of agents today are built on LLM only architectures.

Gemma Allen

>> And let's talk about like the LLM comparative for a second, right? In terms of pointing an LLM at your services data in a PWC or any of these large organizations, right? There are risks that, yeah, we hear that it hallucinates. It doesn't have domain context, right? That's another risk. And then we know that there's certainly guardrail risks broadly, okay? You talk about grounded AI. How do you build a model that has domain context, can talk silo to silo and connect that layer of technology stacks that has not really ever been connected historically and still have a level of agency, what is secure?

Raju Malhotra

>> Yes.

Gemma Allen

>> How do you technically meet that challenge?

Raju Malhotra

>> Yeah. So I think first of all, what you just described, absolutely key. And just to give you an example, in the services business, when we think about selling, delivering and managing, there's a lot of financial information, there's a lot of compliance, there's a lot of government regulations that come into play. And in those cases, very simple thing I would point out is it's not okay to be 99.9999% right because that's when somebody gets fired, you have to be 100% right. And by definition, LLMs are probabilistic. They're fantastic. They're probabilistic. So how we really technically build that in, we have a reasoning layer within our architecture, which brings the determinism to the inherent probabilistic nature of the LLMs. And that's where we bring in 15 plus years of experience that we have as Certinia to work for our customers. When we think about a lot of policies, a lot of objects that we work for our customers, there are certain policies about how a timecard should look like, how a P&L should look like, how account payables, how account receivables should look like. And that's the policies, actually the determinism combined with the probabilistic power of LLM, especially when you used in a trusted context, becomes very powerful for customers.

Gemma Allen

>> Wow. So when we think about the kind of front of house, back of house model of like large professional services firms, right? I know from my own experience, you will see scenarios whereby you could be selling a fantastic product and positioning and rolling out and deploying a great product, but your own backend can sometimes be running cobalt or it can be doing all sorts of things that are based on legacy systems, right? And a lot of the challenge there is around structured data. It's around actually having the right data to feed an agentic model in a way that adds value. How have we suddenly solved for this? I think as somebody who's been in tech for 20 years, it's hard to understand how we've had a magical solution overnight. What are your thoughts?

Raju Malhotra

>> So first of all, I don't think we've solved this problem yet. And I also think it's a journey. It's not quite a destination. And I think sometimes it becomes important to point out that as we un-silo the siloed fractured data within enterprise, the quality of recommendations, the quality of impact, the quality of tasks completed, the quality of, frankly, autonomy becomes better and better. So it's really a journey from that perspective. But I do think what is important is to really decouple a few things. First, I think decoupling the data aspects, structured data, unstructured data, and then decoupling it from the reasoning, and then decoupling it from the experience, the user experience. So when we think about the data aspect of it, structured data is also spread throughout the organizations. Finance team has its own databases, marketing has its own, et cetera. So the first place we start with is having a data fabric that is actionable. And I think that's actually a very good way to start. Unstructured data, for example, all the talk we're having, you can imagine this becomes a video or audio content that gets transcribed, but then what we do actually as part of data is to get the actions, get the next steps out of that transcript using the power of LLM, make sure the users are actually okay, those are the right takeaways because in an hour long conversation, you and I could walk away with two different set of actions, right? So the users and control is important. And then we turn that unstructured data to structure data and then we store that within a single source of truth, which is the system of record for us has been the Salesforce platform where we actually go back and really use that to activate business growth through the CRM, through the data fabric. So that's the data part of it. And then we have the reasoning, as I mentioned, I think we have invested in making sure we bring the best of our domain knowledge to work. And then the user experience gets exposed in different sort of forms, could be Certinia application, could be Salesforce, could be a agent that looks like in your workflow, could be a Slack sort of experience. So those user experiences are wherever the users are and decoupling and making that data more actionable is more of a journey that we give companies a very strong headstart, but we also expect them that this is going to be a very dynamic exercise.

Gemma Allen

>> Can we talk for a second about Agentforce in terms of the role of data alongside Agentforce? Because we were at Salesforce, Dreamforce in October last year, there was so much promise and excitement around this product, right? Talk a little bit about what you're actually seeing, especially in terms of business outcomes. And we also since then have heard the term SaaS apocalypse. I mean, it's hard not to avoid that, right? And I know Benioff has had a lot to say about it. What are your thoughts? What are you hearing and seeing? And how is Agentforce playing out in a synergistic way or talk to me about it?

Raju Malhotra

>> Yeah. So Certinia is one of the largest native ISVs on Salesforce platform. And what that gives us is actually the power of a very easy experience, but also the power to go back to a unified data store. A lot of organizations have their data within different, the Salesforce system, and we basically go to the shared organization. We don't create any copies. We don't integrate. We're basically native in that. So that is a big advantage. We've also been very early, one of the first Agentforce partners, ISB partners. So we actually have worked very closely with the Agentforce team. How we are using Agentforce is very important because we bring our domain knowledge, we bring the data, the metadata, the telemetry, all the reasoning that we have, et cetera, to make the power of Salesforce platform with the data, with the power of the AI stack that is Agentforce, really provide that native ISV experience at scale for our customers. And our experience has been fantastic working with the Salesforce team, and we're seeing tremendous progress as Agentforce team is kind of building more capabilities. We like to think we're actually feeding in a lot of frontier requirements for that and bringing in the power of 1,400 enterprise customers, two million users. So one of the most heavily used product experiences to really make sure I think enterprise are basically their requirements are fed into the roadmap for Agentforce and really it stays differentiated and competitive.

Scott Hebner

>> I think there's another angle to this too, which is your commitment to open standards and interoperability. And we were talking offline, I mentioned that our Agentic AI Futures Index, 59% said over the next 18 months, open standards, open protocols are going to become an RFP requirement. So there's definitely a growing need to be able to integrate this everywhere. So I think Agentforce is a big step forward, but you can actually access it through other platforms, right?

Raju Malhotra

>> Yeah. So the open standards, you're absolutely right. And I think it does come back to the trust and enterprises adopt a new technology when they know that they are actually adopting a transparent, open, scalable technology, because otherwise you might be putting all your eggs in one basket. However good that basket looks today, you want to be future-proof also. So I think the open standards are key and we absolutely support the open standards and really sort of work with our customers to make sure the solutions we have actually fit into their sort of fabric as they think about it. And that's the approach we've taken with the package agents, but also we have intelligent actions. The Veta intelligent actions, we're announcing, excuse me, 64 Veta intelligent actions that actually help you invoke those actions in your own workflow. So they have to be open to be invocable. They can be basically integrated in your own workflow.

Scott Hebner

>> MCP, right?

Raju Malhotra

>> Correct. MCP, A2A, you can actually use them in some prebuilt experiences like with Slack, definitely in Certinia app, or many of the use cases you might have using those open standards.

Scott Hebner

>> So you start checking boxes here. You've got open standards, right? You've got Agentforce, you've got a big focus on trust, right? And then you have the ability to help people make better decisions and do it as a group. So it's not just pure automation.

Raju Malhotra

>> That's correct.

Scott Hebner

>> It's intelligent automation.

Gemma Allen

>> Let's add another box there and let's finish out on this topic. Let's talk about ROI for a second and the business case for this, right? Especially in a market that feels noisy right now. You say that a project manager, for example, can save about 30, 20 hours per month and the cost approximately, correct me if I'm wrong here, is about $30 per user, correct? Or more?

Raju Malhotra

>> That's correct. Yeah. That's the add-on cost we're getting bid at. Yeah.

Gemma Allen

>> The add on cost. For sure. Okay. So give me the pitch on the ROI here. Give me the playback, the business case here.

Raju Malhotra

>> Yeah. So we've been working with many customers over the last few months, and what we've seen is extremely productive and high ROI that they're able to realize. So you're right. A project manager can save up to say 20 hours per month. Think about this. This is almost half a week for a given project manager. And when you scale that across your organization, that has a significant impact. Other thing I want to call out is it's not just a efficiency comparison. Because it's one way to think about efficiency, how many hours we are saving. Think about a Veta agent, an average Veta agent, a reflection of your best employee, because it's trained on your best practices. It's trained on your best policies, best data that you have. So it's actually not your average sort of employee. It's actually probably your best expression of your employees. The impact is quite significant there. So what can you do with the 20 hours a month, almost half a week? If you're a project manager, you can actually do some creative stuff. You can potentially be part of a growth story, not just a delivery story for your company. So I think the impact is quite huge. As we were talking earlier within the professional services, it's a very human bespoke business. So what we've seen is 1% change, 1% improvement in a utilization rate actually has 1.3 to 1.5% impact on the revenue and about 1.5% impact on the EBITDA, the profit margins for our customers. So we're talking quite significant impact by making the humans actually way more productive and more creative. And that translates into tens of millions of dollars or more in terms of revenue and profits for our customers. So the business case is quite staggering actually, and you would expect that given I think we're starting with a very human focused business anyway. In terms of, I think our commercial model, we try to keep it as simple as possible to make sure every single customer has access to using data in their production environment as soon as possible. So you're right, what we are announcing is it's a $30 per user per month add-on to what they're paying as part of being a Certinia customer already. That is sort of like a recurring subscription fee. It uses Agentforce credits as part of consumption. And assuming you have the right contracts on the Agentforce, you are spending your Agentforce consumption, but you're kind of managing that as part of your existing relationship with Salesforce and Agentforce.

Gemma Allen

>> Well, considering the promise, it's certainly a very compelling product and we're excited to see where this goes. And it's wonderful to have you, Raju and Scott, on theCUBE. Thanks so much for coming to the NYSC.

Raju Malhotra

>> Thanks for having me.

Gemma Allen

>> I'm Gemma Allen here at theCUBE Studio at the New York Stock Exchange. This is Mixture of Experts, part of our NYSC wired programming. Thanks for watching.

Raju Malhotra, Certinia & Scott Hebner

search

Gemma Allen

>> Welcome back to theCUBE studio here at the New York Stock Exchange. This is Mixture of Experts, one of our programs with NYSC Wired, and today we have a special segment. We're going to have a conversation on two topics that we talk about day in, day out here on theCUBE, and that is the AI factory thesis, creating AI production infrastructure outcomes at scale, and matching that with business outcomes that compel enterprise users and buyers. Joining me now is Scott Hebner, one of our key analysts at theCUBE Research, and Raju Malhotra, Chief Product and Technology Officer at Certinia. Welcome, folks.

Scott Hebner

>> Thank you.

Raju Malhotra

>> Thank you.

Gemma Allen

>> So Raju, I'm going to start with you because Certinia, talk to me a little about the value proposition of this company, especially from the perspective of really creating business outcomes in the world of AI and agentic.

Raju Malhotra

>> Yeah. So thanks for having Certinia on your show. Thanks for having me. Certinia, think of as a growth engine for services oriented businesses. Maybe one of the ways to think about what we do is to think from the perspective of our customers. On one hand, we work with customers like IBM, PWC, KPMG, companies that have a large pure place services business. And then we also work with companies where services is a big part of their overall business. So companies like Salesforce, Google, AWS, Philips, HPE, Cisco, et cetera, et cetera. And they have services, businesses as part of their overall solutioning. And for all of these customers, we help them sell, deliver, and manage their services business effectively.

Gemma Allen

>> And Scott, you've been doing a lot of research on this company, but also this space, right? Talk to us about the macro-economic moment where there's a lot of noise, a lot of promise, and sometimes people feel not a whole lot of actual outcome. What are your thoughts?

Scott Hebner

>> Yeah, I think the macroeconomic part is what I would call digital labor transformation. It's more than just a technology trend that we see developing now. It's about digital labor. And what I love about what these guys are doing is they're taking the agentic technologies and they're applying it to a very labor-intensive segment of the marketplace, which is professional services. And it's always been the fact that there's been an effort to automate all this. And with agentic AI now, I think they're taking it to a whole new level because it's not just automating repeatable tasks, but it's also helping knowledge workers make better decisions. It's decision intelligence. And I think that's what you need to really bring to life the notion of digital labor, right?

Raju Malhotra

>> Yeah.

Gemma Allen

>> And when we think about knowledge workers and we think about the world of agentic, where you have agents working alongside you, agents working for you. And I mean, if we go with the fear-based thesis, agents potentially replacing you, right? What are your thoughts? What are you both seeing and hearing in the market, especially from your customers directly and your actual user base?

Raju Malhotra

>> Yeah. So first of all, I think this is a very cliched statement to make, but I'll just start something very foundational. AI is transforming every business. Interestingly, AI is transforming businesses where you have a lot of manual work, a lot of bespoke, custom, human focused sort of work. And immediately you kind of think about professional services is the most bespoke business that can be. So the disruption from AI on professional services is probably way more profound than it actually is feeling anywhere else. So you're right. I think there's a thesis of a little bit of a doom and gloom. Is AI taking my job? What's going to happen? We actually did this study and we hired a third party researcher to work with us last year. And what we found, there are a couple of very interesting things. One is that many of our customers and companies in the professional services are already using hybrid teams of humans and agents. Now what that means is that it actually opens up the aperture on the demand side to get the addressable sales pipeline up to six to seven times of what is otherwise possible. And you may wonder, why is that? It comes down to the whole nature of professional services it has been. You can't really sell something in professional services unless you have capacity to deliver it. And who's delivering it? It's the human beings that are delivering traditionally. And that comes down to, do you have the skills? Are you available in the right time zone? Do you have availability, et cetera, et cetera. And that constraint goes away to a large extent when you think about this as a hybrid team. So that's where I think our thesis is, yes, it's a disruption, but it's a very positive opportunity for customers who really take that disruption in the right way and use it to the advantage. So I think coming back to this, it is a fantastic opportunity for anyone in the industry that is working with the professional services, and particularly if you really not only adapt, but really lead that change.

Gemma Allen

>> Wow. And Scott, what are your thoughts, especially in terms of we think about things from the perspective of being an orchestrator in this world, right? Like the build versus buy thesis. There was a time when it was very straightforward, you bought SaaS tools, they created siloed, but very effective based outcomes for your business and particular practices worked really well in silo, right? But business grew and everyone was happy. We know that's not necessarily the future. What are your thoughts?

Scott Hebner

>> Yeah. Again, I go back to the digital labor thought, which is by definition, almost every business you can think of is built upon people working together and collaborating and orchestrating their work. And I think this is what technologies like Certinia is bringing to market here with beta is going to give these knowledge workers superpowers. And this theme that's been over sort of the cloud over AI for over a decade has been, it's going to take everyone's jobs. And our research shows that it's actually going to be a net job growth. In fact, I think there was just a poll out by Fox News where 59% of people thought it was going to eliminate jobs and only 7% thought it was going to grow. And I think what this is going to do is give more people, more knowledge workers, more superpowers. They can do more. They'll be able to do more jobs and people are going to invest those savings into strategic parts of the business. And in the end, I think it'll be a net job growth, right? And what you're doing is not only helping give these people superpowers, but getting everyone working together within a workflow.

Gemma Allen

>> Can we talk a bit about Veta? I was going to call it Veta. So I'm glad you came in there, Scott. But talk a little bit about what this announcement exactly entails. Give me the lowdown on what it is that you guys are building and growing on here.

Raju Malhotra

>> Yeah. So Veta is a suite of packaged agents and intelligent actions that we are making available to our customers. And think about, I think the transformation we've had as an industry, the traditional workflow automation-based SaaS was fantastic to work with human users, but that is sort of in the past because the age that we're moving requires us to use humans and agents working together. As Scott was saying, the digital labor working with the human teams. And that's where Veta comes in. It's really our thesis is that it's going to be a collaboration between human and digital workers, and that means we're going to need the workflow automation from software as a service, a lot of those capabilities, but we're also going to need agentic capabilities. We're going to need package agents that are role-based, that are specialists, that are trusted, they have the guardrails, they have the right capabilities, but we also have intelligent actions that can be invoked anywhere in the enterprise based on the workflows that they have. So it's going to be a combination of sort of the software as a service and the agentic and the way the launch that we have really calls out that differentiation that we are driving for our customers. And it's that agentic suite of those actions and the package agents.

Gemma Allen

>> Wow.

Scott Hebner

>> Yeah. I mean, I would add to that. One of the things I really love about what you're doing is, and I'm going to tee you on the word you used, which was trust. In all the work I've been doing, all the research, all the conversations, if I was to describe enterprise AI in one word, it would be the word trust. I think that's going to be the defining word going forward. And at the same time, an LLM only architecture, when you're building agentic workflows and agents, I think becomes a liability trap on its own. You have to build value around that LLM. And what you've done with context and logic and you've extended that architecture to make it more trustworthy. Just talk a little bit about that.

Raju Malhotra

>> Yeah. So you're, by the way, spot on. I think trust is not only a differentiation, it's a foundational requirement for anything to work, not just in enterprise, anywhere, even in a personal consumer way. I want to work with companies that are, I can trust, and I know my data and everything else is actually trusted. And you're right, we moved very fast from experimentation to being in production for AI. And we started out as LLM, as being able to solve a lot of those problems. And we very soon found out that's actually a very baseline. That's a very foundational requirement, but you really need some more capabilities that drive that trust. So first of all, trust in LLM is key, and we want to make sure I think the LLM providers we work with are trusted, are very transparent, auditable. But on the top of it, I think how we generate trust is really think about this sort of like a five layer cake. Who doesn't like cake, right? So it's about sort of the cake. And before I get into a little bit more specific deep dive, I think if you think about number one is the customer data, could be structured data, unstructured data. Second layer is about metadata and telemetry. That's the data about data. Third is really about a business logic layer, all the code, the actions that are needed. The fourth layer is about reasoning, which basically says which actions you should pick in what context. And then the fifth layer is about user experience. And the user experience could be in your workflow, could be in the Certinia app, could be Slack, could be a lot of different places. But if we think from that perspective, the bottom two layers really become context graph or context engineering because that's based on data and metadata. The next two layers become your business logic actioning and reasoning layer. And then the third part is about the actual, how the agents look like, how the actions surface, et cetera. That's the user experience part of it, which is very important. But it's really, I think the trust is built into all of them and the trust is built on the LLM that we really think about. And that's why we've taken the approach of having role-based agents, having specialist agents, having the actions, but all of this within the trust and governance guardrails that we have baked in.

Scott Hebner

>> Yeah. And I would say that is what is really unique here. LLMs are good at automating repeatable tasks.

Gemma Allen

>> For sure.

Scott Hebner

>> And they're as good as you want your future to be the same as your past, right?

Gemma Allen

>> Yeah.

Scott Hebner

>> When you move into the world of knowledge work where people are making judgments and they have to understand context and they have to know what other people are doing, that's where you need this extended architecture. And you're setting a model for where I think the industry needs to go because the vast majority of agents today are built on LLM only architectures.

Gemma Allen

>> And let's talk about like the LLM comparative for a second, right? In terms of pointing an LLM at your services data in a PWC or any of these large organizations, right? There are risks that, yeah, we hear that it hallucinates. It doesn't have domain context, right? That's another risk. And then we know that there's certainly guardrail risks broadly, okay? You talk about grounded AI. How do you build a model that has domain context, can talk silo to silo and connect that layer of technology stacks that has not really ever been connected historically and still have a level of agency, what is secure?

Raju Malhotra

>> Yes.

Gemma Allen

>> How do you technically meet that challenge?

Raju Malhotra

>> Yeah. So I think first of all, what you just described, absolutely key. And just to give you an example, in the services business, when we think about selling, delivering and managing, there's a lot of financial information, there's a lot of compliance, there's a lot of government regulations that come into play. And in those cases, very simple thing I would point out is it's not okay to be 99.9999% right because that's when somebody gets fired, you have to be 100% right. And by definition, LLMs are probabilistic. They're fantastic. They're probabilistic. So how we really technically build that in, we have a reasoning layer within our architecture, which brings the determinism to the inherent probabilistic nature of the LLMs. And that's where we bring in 15 plus years of experience that we have as Certinia to work for our customers. When we think about a lot of policies, a lot of objects that we work for our customers, there are certain policies about how a timecard should look like, how a P&L should look like, how account payables, how account receivables should look like. And that's the policies, actually the determinism combined with the probabilistic power of LLM, especially when you used in a trusted context, becomes very powerful for customers.

Gemma Allen

>> Wow. So when we think about the kind of front of house, back of house model of like large professional services firms, right? I know from my own experience, you will see scenarios whereby you could be selling a fantastic product and positioning and rolling out and deploying a great product, but your own backend can sometimes be running cobalt or it can be doing all sorts of things that are based on legacy systems, right? And a lot of the challenge there is around structured data. It's around actually having the right data to feed an agentic model in a way that adds value. How have we suddenly solved for this? I think as somebody who's been in tech for 20 years, it's hard to understand how we've had a magical solution overnight. What are your thoughts?

Raju Malhotra

>> So first of all, I don't think we've solved this problem yet. And I also think it's a journey. It's not quite a destination. And I think sometimes it becomes important to point out that as we un-silo the siloed fractured data within enterprise, the quality of recommendations, the quality of impact, the quality of tasks completed, the quality of, frankly, autonomy becomes better and better. So it's really a journey from that perspective. But I do think what is important is to really decouple a few things. First, I think decoupling the data aspects, structured data, unstructured data, and then decoupling it from the reasoning, and then decoupling it from the experience, the user experience. So when we think about the data aspect of it, structured data is also spread throughout the organizations. Finance team has its own databases, marketing has its own, et cetera. So the first place we start with is having a data fabric that is actionable. And I think that's actually a very good way to start. Unstructured data, for example, all the talk we're having, you can imagine this becomes a video or audio content that gets transcribed, but then what we do actually as part of data is to get the actions, get the next steps out of that transcript using the power of LLM, make sure the users are actually okay, those are the right takeaways because in an hour long conversation, you and I could walk away with two different set of actions, right? So the users and control is important. And then we turn that unstructured data to structure data and then we store that within a single source of truth, which is the system of record for us has been the Salesforce platform where we actually go back and really use that to activate business growth through the CRM, through the data fabric. So that's the data part of it. And then we have the reasoning, as I mentioned, I think we have invested in making sure we bring the best of our domain knowledge to work. And then the user experience gets exposed in different sort of forms, could be Certinia application, could be Salesforce, could be a agent that looks like in your workflow, could be a Slack sort of experience. So those user experiences are wherever the users are and decoupling and making that data more actionable is more of a journey that we give companies a very strong headstart, but we also expect them that this is going to be a very dynamic exercise.

Gemma Allen

>> Can we talk for a second about Agentforce in terms of the role of data alongside Agentforce? Because we were at Salesforce, Dreamforce in October last year, there was so much promise and excitement around this product, right? Talk a little bit about what you're actually seeing, especially in terms of business outcomes. And we also since then have heard the term SaaS apocalypse. I mean, it's hard not to avoid that, right? And I know Benioff has had a lot to say about it. What are your thoughts? What are you hearing and seeing? And how is Agentforce playing out in a synergistic way or talk to me about it?

Raju Malhotra

>> Yeah. So Certinia is one of the largest native ISVs on Salesforce platform. And what that gives us is actually the power of a very easy experience, but also the power to go back to a unified data store. A lot of organizations have their data within different, the Salesforce system, and we basically go to the shared organization. We don't create any copies. We don't integrate. We're basically native in that. So that is a big advantage. We've also been very early, one of the first Agentforce partners, ISB partners. So we actually have worked very closely with the Agentforce team. How we are using Agentforce is very important because we bring our domain knowledge, we bring the data, the metadata, the telemetry, all the reasoning that we have, et cetera, to make the power of Salesforce platform with the data, with the power of the AI stack that is Agentforce, really provide that native ISV experience at scale for our customers. And our experience has been fantastic working with the Salesforce team, and we're seeing tremendous progress as Agentforce team is kind of building more capabilities. We like to think we're actually feeding in a lot of frontier requirements for that and bringing in the power of 1,400 enterprise customers, two million users. So one of the most heavily used product experiences to really make sure I think enterprise are basically their requirements are fed into the roadmap for Agentforce and really it stays differentiated and competitive.

Scott Hebner

>> I think there's another angle to this too, which is your commitment to open standards and interoperability. And we were talking offline, I mentioned that our Agentic AI Futures Index, 59% said over the next 18 months, open standards, open protocols are going to become an RFP requirement. So there's definitely a growing need to be able to integrate this everywhere. So I think Agentforce is a big step forward, but you can actually access it through other platforms, right?

Raju Malhotra

>> Yeah. So the open standards, you're absolutely right. And I think it does come back to the trust and enterprises adopt a new technology when they know that they are actually adopting a transparent, open, scalable technology, because otherwise you might be putting all your eggs in one basket. However good that basket looks today, you want to be future-proof also. So I think the open standards are key and we absolutely support the open standards and really sort of work with our customers to make sure the solutions we have actually fit into their sort of fabric as they think about it. And that's the approach we've taken with the package agents, but also we have intelligent actions. The Veta intelligent actions, we're announcing, excuse me, 64 Veta intelligent actions that actually help you invoke those actions in your own workflow. So they have to be open to be invocable. They can be basically integrated in your own workflow.

Scott Hebner

>> MCP, right?

Raju Malhotra

>> Correct. MCP, A2A, you can actually use them in some prebuilt experiences like with Slack, definitely in Certinia app, or many of the use cases you might have using those open standards.

Scott Hebner

>> So you start checking boxes here. You've got open standards, right? You've got Agentforce, you've got a big focus on trust, right? And then you have the ability to help people make better decisions and do it as a group. So it's not just pure automation.

Raju Malhotra

>> That's correct.

Scott Hebner

>> It's intelligent automation.

Gemma Allen

>> Let's add another box there and let's finish out on this topic. Let's talk about ROI for a second and the business case for this, right? Especially in a market that feels noisy right now. You say that a project manager, for example, can save about 30, 20 hours per month and the cost approximately, correct me if I'm wrong here, is about $30 per user, correct? Or more?

Raju Malhotra

>> That's correct. Yeah. That's the add-on cost we're getting bid at. Yeah.

Gemma Allen

>> The add on cost. For sure. Okay. So give me the pitch on the ROI here. Give me the playback, the business case here.

Raju Malhotra

>> Yeah. So we've been working with many customers over the last few months, and what we've seen is extremely productive and high ROI that they're able to realize. So you're right. A project manager can save up to say 20 hours per month. Think about this. This is almost half a week for a given project manager. And when you scale that across your organization, that has a significant impact. Other thing I want to call out is it's not just a efficiency comparison. Because it's one way to think about efficiency, how many hours we are saving. Think about a Veta agent, an average Veta agent, a reflection of your best employee, because it's trained on your best practices. It's trained on your best policies, best data that you have. So it's actually not your average sort of employee. It's actually probably your best expression of your employees. The impact is quite significant there. So what can you do with the 20 hours a month, almost half a week? If you're a project manager, you can actually do some creative stuff. You can potentially be part of a growth story, not just a delivery story for your company. So I think the impact is quite huge. As we were talking earlier within the professional services, it's a very human bespoke business. So what we've seen is 1% change, 1% improvement in a utilization rate actually has 1.3 to 1.5% impact on the revenue and about 1.5% impact on the EBITDA, the profit margins for our customers. So we're talking quite significant impact by making the humans actually way more productive and more creative. And that translates into tens of millions of dollars or more in terms of revenue and profits for our customers. So the business case is quite staggering actually, and you would expect that given I think we're starting with a very human focused business anyway. In terms of, I think our commercial model, we try to keep it as simple as possible to make sure every single customer has access to using data in their production environment as soon as possible. So you're right, what we are announcing is it's a $30 per user per month add-on to what they're paying as part of being a Certinia customer already. That is sort of like a recurring subscription fee. It uses Agentforce credits as part of consumption. And assuming you have the right contracts on the Agentforce, you are spending your Agentforce consumption, but you're kind of managing that as part of your existing relationship with Salesforce and Agentforce.

Gemma Allen

>> Well, considering the promise, it's certainly a very compelling product and we're excited to see where this goes. And it's wonderful to have you, Raju and Scott, on theCUBE. Thanks so much for coming to the NYSC.

Raju Malhotra

>> Thanks for having me.

Gemma Allen

>> I'm Gemma Allen here at theCUBE Studio at the New York Stock Exchange. This is Mixture of Experts, part of our NYSC wired programming. Thanks for watching.