theCUBE + NYSE Wired: Mixture of Experts Series | Alex Rinke, Celonis

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

Alex Rinke

CEO

Celonis

play_circle_outline Importance of re-engineering processes with new AI technologies for efficiency and customer benefit.

play_circle_outline Transforming Enterprise Operations: The Impact of Process Intelligence and Composable Frameworks on Modern Automation Solutions

play_circle_outline Notable use cases showcasing successful implementations of Celonis's process intelligence.

play_circle_outline Upcoming event, Celosphere, will focus on customer success stories and proof of concept.

Info

Alex Rinke, Celonis

Alex Rinke

CEO Celonis

search

>> Hello, I'm John Furrier, host of theCUBE, here at our NYSE CUBE studios on the East Coast. Of course, we have our Palo Alto Studio connecting Silicon Valley and Wall Street, as we expand our studio presence. Of course we cover the events. We're here to unpack the Mixture of Expert series and also preview Celonis' upcoming event. Alex Rinke's here, co-founder and CEO of Celonis. Alex, great to see you again. Welcome back to our new studio here. We were out there last time on the balcony, but this is our set studio.

Alex Rinke

>> I think this is so much cooler. There's actual trading. Thanks for having me, John. Thanks for having me back on the show.

>> It's very cool. And got the action behind us where they're actively trading, so it's kind of fun.

Alex Rinke

>> I didn't even know that existed anymore.

>> I thought it was all electronic-

Alex Rinke

>> Exactly....

>> but they soon will be electronic. But this is the theme of the whole world we're seeing. AI and physical AIs, the convergence of digital and physicals coming together. Last year, almost to a year, you were here. We discussed what you guys are doing for your business. Obviously, the growth has been phenomenal. We've been tracking that on SiliconANGLE and theCUBE and theCUBE Research, but we talked about the transformation of AI with agents. It was pretty obvious then. We talked about it. But you said, "We are at the beginning of the transformation." Now, a full year in, I'd say with the middle of the eye of the storm right now, if you look at what's happening at all the conferences, the agent hype is at an all-time high. And so, now you start to see the narrative, like okay, show me. Now, first of all, I like that because one, the proof is in the pudding, as they say. Two, with AI, things go faster. So, if you get something early, you better show something. And so, there is a new culture of, "I don't want to see that demo," some call it a parlor trick. Show me a real agentic system, probabilistic, that can pass to another agent that could be deterministic, go to a payment rail. I mean there are use cases of agents happening now, so give us the update. One year later you coined that term, transformation. We're in it. Where are we?

Alex Rinke

>> Oh, yeah, it's super exciting because it's happening every day. And the way I think about this is last year most companies were just using agents or AI for coding, for customer support and for internal knowledge retrieval tools, and that's still where many people are. But what we see on the cutting edge is that people are moving into agentifying, their core operations and processes. So, they are automating accounting, supply chain, customer order management, experience. And they are not just automating things, they're actually re-engineering their processes. You remember Michael Hammer?

>> Yeah.

Alex Rinke

>> Right. So, the new technology rolls around. Back then it was the P system client-server applications and people went, "Oh, we need to re-engineer all of our processes." And this is happening again because now people say, "AI is here, so let's take a hard look at our processes and re-engineer them to make them dramatically more efficient, better for customers, better for people." So, it's really exciting.

>> I want to ask you about the future of the enterprise because I think what I like about Celonis and what you guys are doing with the company is that you're attacking it in a modern way. You're not going back to the old-school process management. So, I think process management's had its day, and I want to get your thoughts on that. I know you mentioned process, but you go back to old process management. Okay, lay it out. But with generative AI, process management looks different. It almost shifts to a business model, not a mechanism, like an IT process. And so, you're starting to see the management books that were written, new ones are being written on the fly. So, I want to ask your thoughts on this. How do you re-engineer the process and what's the data tell you? Because you got to have some technology involved that are end-to-end. There are agents in there, but you've got domain experts, you've got people who know the business logic. So, the business model outcome-based solution is not started with process.

Alex Rinke

>> Oh, 100%.

>> So, talk about that dynamic of how the process generation's evolving because you still need process. End-to-end workflow is a process, but you don't just do that. So, what's your reaction to that?

Alex Rinke

>> Yeah, so it's very different. Obviously, the idea that you have to re-engineer process is similar, but obviously the mechanisms which if you do it are completely different. So, we call it process intelligence. And it's really a layer that can sit across all of your different systems and provide the foundation for this modernization. And we believe the future of the enterprise will be AI-driven and composable. You'll be able to put together processes very quickly. You'll be able to use different Lego blocks, like agents and workflows and copilots, and you'll be able to have a much, much higher level of automation. But you need to fuel it with the context of how your business operates and how you want your business to operate, what your goals are and that's what process intelligence provides. When you look at it, the reality of adoption is that still... There was this MIT study, we talked about this earlier. One out of 20 AI pilots fail to deliver P&L impact. I mean you can look at it differently. 19 of 20 are not quite there yet. They're not failures, they're experiments. But how do you get that success rate higher? And how do you get people to really transform their business operations? And that's where AI is lacking the business context. AI is trained a lot of public data, but it doesn't have the context about your particular business, and that's where process intelligence is.

>> So, Alex, talk about the modernization methodology because this failure thing, I called that out publicly because I looked at the data, I'm not sure their sources were accurate, but they're not wrong. And all experimentation things fail. The nuance to that study that MIT put out about the failure of projects, I think that's a feature, not a bug, actually.

Alex Rinke

>> 100%.

>> But a lot of those were not in production. So, there's the pre-production, then there's actually in production. So, talk about what you're seeing in both production workloads with agents today and what's pre-production that's being worked on. Because people are doing a lot more in pre-production to get ready for production, but there are some production agents out there-

Alex Rinke

>> Oh, 100%. And this is the exciting part. We see production agents, whether it's clinical trials operations, accounting, supply chain, front office. We see them everywhere across our most advanced customers. And the way you put things into production. First, AI is very context-sensitive. So, you need to have the business context in a layer. You need to have this process intelligence foundation that tells the AI how the business flows. That's why we say there's no AI without PI. But then also you need the orchestration capability to basically embed and fuel this data to keep a consistent state of the process underneath. And then, finally, you need the observability. Now, you have a process that's part AI and part human and part traditional automation. You need an observability of that process across. So, that's why we call it foundation for this modernization and that's by PI enables AI.

>> So, talk about the management approach. I'm really honing on what you guys are calling on the business outcome side, the modernization approach. Because it used to be, "Hey, we want to do this. IT, go get some money, go deploy it, test it, put it in production." Now, the businesses are driving it, it's more business model disruption. I've heard people say things like, I call it the at, on and in AI. So, we have agents at the company, looking at how we run our company, the processes and the workflows, on top, the product and then in the product. So, there seems to be three views that people are looking at AI, and that changes the scope. Because if you have it in our business logic, that's running the business. Then, you've got products that aren't AI-native. Okay, let's get AI in there. Vector search is a great example. And then in, AI in that is AI native. What's your reaction to that? How do you see customers tackling how to frame their build-outs and their planning? Because I think a lot of the failure that does happen in pre-production is because they're just jumping in the deep end.

Alex Rinke

>> Yeah, exactly. And that's where process mining and also task mining, we provide both at Celonis, is really, really important. We can go cross-system, cross-modality. So, tasks that on system and task the people do off system, emails, Excel, all that stuff. We can get full visibility in that, the real data, what people are doing, how the organization works. So, you get that x-ray picture across 1,000 people, 2,000 people, 10,000 people and all the backend processes, and you light up the room. And you're like, "This is where we are going to re-engineer. This is where we're going to agentify. This is we actually can avoid the whole workload altogether." So, then you modernize your entire operation. You can do it so much faster and you're not restricted by any system. It's not like, "Oh, let's put this system in and let's be restricted by whatever the system can do and retrofit our processes in new systems." We believe in freeing the process, freeing it from the constraints, and we believe in an open ecosystem. And this idea of composability.

>> Yeah, I mean I like the re-engineering of the processes, that's a key opportunity. All right. so it's been a year. Can you give me an update on what's happened? Take me through some of the use cases that you're seeing? What's going on in the business right now? Obviously, the event's coming up for all your customers, what's happened? Give us some highlights. What's the highlight reel look like?

Alex Rinke

>> Last year we talked a lot about this idea of the process intelligence graph, our capability to bring together process data and context, and also, process designs from across systems and domains across the enterprise. And that's really proven to be an essential dataset and data asset for this AI-driven and composable enterprise. So, we see a lot of validation for that, both in terms of new logos, expansions with our existing customers, a lot of growth. We are going to show some great cases at Celosphere, right? We are going to have Vinma there, right? Vinma is a distributor in the oil and gas space here in the US, and they are really reinventing their whole business, how they operate, how they execute, how they transform the organization. So, that's going to be a really, really exciting story. But even the public sector, the State of Oklahoma is at the forefront of using Celonis and AI to optimize their operations for their citizens to put money back in their citizens' pockets, which is really, really exciting. You look at the NHS, the , there's so many examples. thyssenkrupp Rasselstein, this is one of my favorite ones, they're going to be there. They had 300 systems running their supply chain, think about applying AI to that. Good luck. But they put in the process intelligence layer and they're kicking. So, it's-

>> Give us some examples of some anecdotal commentary. How did that change their world? What was the impact?

Alex Rinke

>> Well, I mean you think about order cycle times, right? Getting back to customers faster, getting quotes out, not reworking everything in the back-end when you put invoices out, making sure that the agent knows or the process knows what actually you committed to on the front-end across systems, so that you get the precision up, you accelerate cash flows. At the NHS, it's about making sure people don't need to wait for appointments that long, so they accelerate the waiting times of people to get their-

>> So, The pipes are clogged and you just free up all the flow.

Alex Rinke

>> Exactly.

>> The money flow. The process flow.

Alex Rinke

>> We look at every business as a collection of interacting processes. And if we can accelerate that, we make the business go faster, we free up capacity for innovation, we make it more efficient, we make it better for customers, better for employees. It's just really-

>> All right. So, I'm interested. Just say I'm a customer... Well, I'll play devil's advocate. I'm a customer. "Man, my processes all screwed up. I got to fix this. I got all these legacy systems. I got silos everywhere. What do I do? Give me the playbook."

Alex Rinke

>> 100%. We have a playbook for that. This is good news. So, you put the process intelligence in, you connect it to your system, the process intelligence graph. First of all, you're going to get a lot of quick wins. You're going to already have paid for the projects just with the quick wins. And then, you can say, "Well, let's rethink these processes." So, then you get into our design component and you design the future state and you do it free of all the limitations that you're used to because now you have AI, now you have intelligent automation. And then, you use this composable framework to then compose it. And you want to drop Microsoft in there for agentic and Copilot? Great, put it in there. Let it run the process. You want to put workflow in there? Whatever you want to put in there, you just have a canvas to put it together and completely modernize your operations across the front office, the back office, the middle office, and you're going to do it at a speed that's unheard of.

>> Yeah, talk about the ecosystem because I was talking recently with the head of KPMG and I've talked to all the GSIs, Accenture, you name, I know them, PwC, they're all out there. But this KPMG executive said, "The way we used to do management consulting and advisory has changed significantly," one. The ecosystem of services are broader than what they can offer, but two, that the services demand was on the business model. So, what's happened is he said, "The process management mindset that was in the middle layer of the management ranks, go out and do process, shifted to the C-suite, where they say, 'Let's re-architect the business.'" Then it's got to come back down. This is where you guys come in. This is where the process intelligence layer graph comes in. But you guys have almost a system to say, "Hey, here's some optimization." So, you are essentially business model.

Alex Rinke

>> Yeah, I understand. You re-engineer-

>> Did I get that right?

Alex Rinke

>> Yeah, you have the , all the data, the operational state of the system. And then, you put in the target state, here's the design and how it should operate. And then, you can make it happen also. So, we think about analyze, design, operate. Celonis supports all three.

>> So, that's kind of like the holy grail, so business must be booming?

Alex Rinke

>> It's really exciting, what we see.

>> Give us a taste of the growth.

Alex Rinke

>> Yeah, I mean I had eight customer meetings in the last 10 days and everyone is like, "How can we move as fast as possible?"

>> "Can we go faster?"

Alex Rinke

>> "Can we go faster?" It's really, really exciting. People are also starting to really see with those capabilities that are now available in this ecosystem, how can we completely reinvent how we do things? And that's exciting because when you're talking, not just about an agent here or there. You're talking about true business transformation.

>> I was talking to Brian Baumann of NYSE Wired, and we were talking about CFOs because we had an event last year with CFOs and I wanted to test the thesis out, how well do they plugged into AI? Okay, we're going to have our friends come in. Turns out we had 30 CFOs, roughly, and they all were actually interested and knew a bit about AI. And I discovered that they're the architects of the business model, they're instrumenting the business model. So, the finance group, not obvious, they come in because money's involved, one. Savings, top-line growth, the efficiencies kick in. What's your reaction to that? Man, I'm not asking if you're calling on CFOs, but maybe you do. I'm sure you do. But the role of the CFO becomes now a stakeholder.

Alex Rinke

>> I think it's the CFO. I think it's just an example of the fact that organizations are moving from very siloed and functional. I have my HR application, HR process, and you have your manufacturing-

>> Stay in your lane....

Alex Rinke

>> procurement. Stay in your lane. You have these silos. But really now, where the processes are free to interact and you can get intelligence across, you now work in a much more collaborative model and that's much more fun. That's what people want to do anyway. So, I think it's all about teamwork. It's all about putting people in the position and empower them to actually make those changes because this change is not just driven from the top down, you need the top. I talk to the CIO and they'd be like, "Can you come meet our CFO or CEO because they want to learn about this?" So, there's a lot of interest in this.

>> Yeah, the money involved, I have to count it.

Alex Rinke

>> Exactly. And there's a lot of C-level push. "How are we agentifying? How are we using AI in our business?" And I think part of that is because it's so front-setting in the consumer world, it's like, "If it can do my daughter's or son's homework, why aren't we using it in the business?" So, I think you have this huge interest and sponsorship, but then you have every single person-

>> Yeah, it touches everyone....

Alex Rinke

>> collaborating and changing how they work and how their business works, and I think it's very empowering technology.

>> Alex, thanks for coming in and I am sure the show will be phenomenal, like last year. We'll be there with theCUBE, covering it.

Alex Rinke

>> We look forward to having you.

>> I have to ask you about agent, the current trends in agentic infrastructure. I'm hearing fleets of agents because we've talked in the past about data quality, data feeds AI. I've been having a debate with some other analysts about large language models, small language models. But if you look at the agent conversation, you're in the middle of it. You can have multiple, multiple agents talking to each other, so context becomes huge. And data quality used to be the conversation, in search and any conversation around AI or any kind of system, but you don't need to have a lot of data to get a lot of quality. You got to have the right context. And I want to talk about that role of in context to how agents will interact with each other. If I want to book a trip or I want to do a process, I might not know what I want to do yet. I'm not really deterministic yet. I'm like, "Okay, now I generate what I want. I'm going on vacation. Where do I want to go?" I search around and then it's like, "Boom, I'm going to Hawaii." Now, I got to book a hotel, now I got to go to a payment rail. That's a different agent. So, agents are going to have tasks. How do you think about that? How are your customers thinking about the role of agents, the delegation, the trust? Because one could be doing collection, one could be doing payments, one could be doing this, so you have roles of the agents.

Alex Rinke

>> You need to have this process data, that's why we call it process intelligence graph. It can associate all those different process flows, and not just the process, but the tasks, what people do manually and the desired approach as well, and even other types of information. So, it becomes this ultimate context layer where you say, "Hey, this is what's happening in other process over here. By the way, the last three cases that look similar, it went like this and it actually went into a roadblock." The more you feed into this agent, these context windows are getting bigger and bigger, the better the agent will be at coming up with the right decision, making the right recommendation, putting the right automation. So, that's why it's saying it's a very input-sensitive technology. The better the input, the better the output.

>> So, good context window helps, but also task specifically-

Alex Rinke

>> You need the foundation, and there's also some things that you want to be deterministic.

>> Yeah, exactly. Yeah, like a payment rail.

Alex Rinke

>> So, that's why we say there's no AI without PI. You need this process intelligence layer to fuel the AI. Also, you want to give the AI not just what's going on in your business, your data. You also want to tell them what's supposed to, what's the design of the process, right? What's our guardrails? How are our approval rules? When does a human need to get involved? We can provide all of this with a single API.

>> So, the agents talking to agents, sub-agents, you see that as a feature?

Alex Rinke

>> Oh, 100%.

>> 100%?

Alex Rinke

>> And you can put orchestration, you can have part deterministic. That's why we call it composable. You have these Lego blocks and you put them together to completely agentify and automate your business.

>> And that changes the process management. All right, final question. What can we expect at the event? Give us a little teaser without revealing too much. Saving for the main stage.

Alex Rinke

>> So, it's super exciting because process intelligence is really getting into the forefront because there's so much process change as a result of those agentic initiatives. So, we're going to talk about how help. The progress we made on our platform. Last year we released AgentC, this agent enablement and AI, and we talked about that last year. We've made a lot of progress. We're not going to spill all the beans right now, but it's really exciting. And we're going to put a really, really strong focus on customer examples, customers on stage.

>> Proof?

Alex Rinke

>> Proof points, because everybody wants to learn from each other. "What are you seeing? How are you doing it?" Something very new that we wrapping our heads around, so there's going to be very strong emphasis and there's no place where you can learn more about process intelligence and how people are doing it than Celosphere.

>> Yeah, and that's the marker right now. They want to see the proof. They want to see the hype cycle convert-

Alex Rinke

>> Its the show me.

>> Yeah, it's a hype cycle converting into reality. Alex, thank you for coming on and congratulations again. theCUBE bringing all the action here at the NYSE studio for theCUBE. And of course, the velocity of the company's growths are going to be depending on how well they can get agents... Celonis is really on the escape velocity. Check out their event. Of course, theCUBE will be there. I'm John Furrier. Thanks for watching.

Alex Rinke

>> Thanks, John.

Alex Rinke, Celonis

search

>> Hello, I'm John Furrier, host of theCUBE, here at our NYSE CUBE studios on the East Coast. Of course, we have our Palo Alto Studio connecting Silicon Valley and Wall Street, as we expand our studio presence. Of course we cover the events. We're here to unpack the Mixture of Expert series and also preview Celonis' upcoming event. Alex Rinke's here, co-founder and CEO of Celonis. Alex, great to see you again. Welcome back to our new studio here. We were out there last time on the balcony, but this is our set studio.

Alex Rinke

>> I think this is so much cooler. There's actual trading. Thanks for having me, John. Thanks for having me back on the show.

>> It's very cool. And got the action behind us where they're actively trading, so it's kind of fun.

Alex Rinke

>> I didn't even know that existed anymore.

>> I thought it was all electronic-

Alex Rinke

>> Exactly....

>> but they soon will be electronic. But this is the theme of the whole world we're seeing. AI and physical AIs, the convergence of digital and physicals coming together. Last year, almost to a year, you were here. We discussed what you guys are doing for your business. Obviously, the growth has been phenomenal. We've been tracking that on SiliconANGLE and theCUBE and theCUBE Research, but we talked about the transformation of AI with agents. It was pretty obvious then. We talked about it. But you said, "We are at the beginning of the transformation." Now, a full year in, I'd say with the middle of the eye of the storm right now, if you look at what's happening at all the conferences, the agent hype is at an all-time high. And so, now you start to see the narrative, like okay, show me. Now, first of all, I like that because one, the proof is in the pudding, as they say. Two, with AI, things go faster. So, if you get something early, you better show something. And so, there is a new culture of, "I don't want to see that demo," some call it a parlor trick. Show me a real agentic system, probabilistic, that can pass to another agent that could be deterministic, go to a payment rail. I mean there are use cases of agents happening now, so give us the update. One year later you coined that term, transformation. We're in it. Where are we?

Alex Rinke

>> Oh, yeah, it's super exciting because it's happening every day. And the way I think about this is last year most companies were just using agents or AI for coding, for customer support and for internal knowledge retrieval tools, and that's still where many people are. But what we see on the cutting edge is that people are moving into agentifying, their core operations and processes. So, they are automating accounting, supply chain, customer order management, experience. And they are not just automating things, they're actually re-engineering their processes. You remember Michael Hammer?

>> Yeah.

Alex Rinke

>> Right. So, the new technology rolls around. Back then it was the P system client-server applications and people went, "Oh, we need to re-engineer all of our processes." And this is happening again because now people say, "AI is here, so let's take a hard look at our processes and re-engineer them to make them dramatically more efficient, better for customers, better for people." So, it's really exciting.

>> I want to ask you about the future of the enterprise because I think what I like about Celonis and what you guys are doing with the company is that you're attacking it in a modern way. You're not going back to the old-school process management. So, I think process management's had its day, and I want to get your thoughts on that. I know you mentioned process, but you go back to old process management. Okay, lay it out. But with generative AI, process management looks different. It almost shifts to a business model, not a mechanism, like an IT process. And so, you're starting to see the management books that were written, new ones are being written on the fly. So, I want to ask your thoughts on this. How do you re-engineer the process and what's the data tell you? Because you got to have some technology involved that are end-to-end. There are agents in there, but you've got domain experts, you've got people who know the business logic. So, the business model outcome-based solution is not started with process.

Alex Rinke

>> Oh, 100%.

>> So, talk about that dynamic of how the process generation's evolving because you still need process. End-to-end workflow is a process, but you don't just do that. So, what's your reaction to that?

Alex Rinke

>> Yeah, so it's very different. Obviously, the idea that you have to re-engineer process is similar, but obviously the mechanisms which if you do it are completely different. So, we call it process intelligence. And it's really a layer that can sit across all of your different systems and provide the foundation for this modernization. And we believe the future of the enterprise will be AI-driven and composable. You'll be able to put together processes very quickly. You'll be able to use different Lego blocks, like agents and workflows and copilots, and you'll be able to have a much, much higher level of automation. But you need to fuel it with the context of how your business operates and how you want your business to operate, what your goals are and that's what process intelligence provides. When you look at it, the reality of adoption is that still... There was this MIT study, we talked about this earlier. One out of 20 AI pilots fail to deliver P&L impact. I mean you can look at it differently. 19 of 20 are not quite there yet. They're not failures, they're experiments. But how do you get that success rate higher? And how do you get people to really transform their business operations? And that's where AI is lacking the business context. AI is trained a lot of public data, but it doesn't have the context about your particular business, and that's where process intelligence is.

>> So, Alex, talk about the modernization methodology because this failure thing, I called that out publicly because I looked at the data, I'm not sure their sources were accurate, but they're not wrong. And all experimentation things fail. The nuance to that study that MIT put out about the failure of projects, I think that's a feature, not a bug, actually.

Alex Rinke

>> 100%.

>> But a lot of those were not in production. So, there's the pre-production, then there's actually in production. So, talk about what you're seeing in both production workloads with agents today and what's pre-production that's being worked on. Because people are doing a lot more in pre-production to get ready for production, but there are some production agents out there-

Alex Rinke

>> Oh, 100%. And this is the exciting part. We see production agents, whether it's clinical trials operations, accounting, supply chain, front office. We see them everywhere across our most advanced customers. And the way you put things into production. First, AI is very context-sensitive. So, you need to have the business context in a layer. You need to have this process intelligence foundation that tells the AI how the business flows. That's why we say there's no AI without PI. But then also you need the orchestration capability to basically embed and fuel this data to keep a consistent state of the process underneath. And then, finally, you need the observability. Now, you have a process that's part AI and part human and part traditional automation. You need an observability of that process across. So, that's why we call it foundation for this modernization and that's by PI enables AI.

>> So, talk about the management approach. I'm really honing on what you guys are calling on the business outcome side, the modernization approach. Because it used to be, "Hey, we want to do this. IT, go get some money, go deploy it, test it, put it in production." Now, the businesses are driving it, it's more business model disruption. I've heard people say things like, I call it the at, on and in AI. So, we have agents at the company, looking at how we run our company, the processes and the workflows, on top, the product and then in the product. So, there seems to be three views that people are looking at AI, and that changes the scope. Because if you have it in our business logic, that's running the business. Then, you've got products that aren't AI-native. Okay, let's get AI in there. Vector search is a great example. And then in, AI in that is AI native. What's your reaction to that? How do you see customers tackling how to frame their build-outs and their planning? Because I think a lot of the failure that does happen in pre-production is because they're just jumping in the deep end.

Alex Rinke

>> Yeah, exactly. And that's where process mining and also task mining, we provide both at Celonis, is really, really important. We can go cross-system, cross-modality. So, tasks that on system and task the people do off system, emails, Excel, all that stuff. We can get full visibility in that, the real data, what people are doing, how the organization works. So, you get that x-ray picture across 1,000 people, 2,000 people, 10,000 people and all the backend processes, and you light up the room. And you're like, "This is where we are going to re-engineer. This is where we're going to agentify. This is we actually can avoid the whole workload altogether." So, then you modernize your entire operation. You can do it so much faster and you're not restricted by any system. It's not like, "Oh, let's put this system in and let's be restricted by whatever the system can do and retrofit our processes in new systems." We believe in freeing the process, freeing it from the constraints, and we believe in an open ecosystem. And this idea of composability.

>> Yeah, I mean I like the re-engineering of the processes, that's a key opportunity. All right. so it's been a year. Can you give me an update on what's happened? Take me through some of the use cases that you're seeing? What's going on in the business right now? Obviously, the event's coming up for all your customers, what's happened? Give us some highlights. What's the highlight reel look like?

Alex Rinke

>> Last year we talked a lot about this idea of the process intelligence graph, our capability to bring together process data and context, and also, process designs from across systems and domains across the enterprise. And that's really proven to be an essential dataset and data asset for this AI-driven and composable enterprise. So, we see a lot of validation for that, both in terms of new logos, expansions with our existing customers, a lot of growth. We are going to show some great cases at Celosphere, right? We are going to have Vinma there, right? Vinma is a distributor in the oil and gas space here in the US, and they are really reinventing their whole business, how they operate, how they execute, how they transform the organization. So, that's going to be a really, really exciting story. But even the public sector, the State of Oklahoma is at the forefront of using Celonis and AI to optimize their operations for their citizens to put money back in their citizens' pockets, which is really, really exciting. You look at the NHS, the , there's so many examples. thyssenkrupp Rasselstein, this is one of my favorite ones, they're going to be there. They had 300 systems running their supply chain, think about applying AI to that. Good luck. But they put in the process intelligence layer and they're kicking. So, it's-

>> Give us some examples of some anecdotal commentary. How did that change their world? What was the impact?

Alex Rinke

>> Well, I mean you think about order cycle times, right? Getting back to customers faster, getting quotes out, not reworking everything in the back-end when you put invoices out, making sure that the agent knows or the process knows what actually you committed to on the front-end across systems, so that you get the precision up, you accelerate cash flows. At the NHS, it's about making sure people don't need to wait for appointments that long, so they accelerate the waiting times of people to get their-

>> So, The pipes are clogged and you just free up all the flow.

Alex Rinke

>> Exactly.

>> The money flow. The process flow.

Alex Rinke

>> We look at every business as a collection of interacting processes. And if we can accelerate that, we make the business go faster, we free up capacity for innovation, we make it more efficient, we make it better for customers, better for employees. It's just really-

>> All right. So, I'm interested. Just say I'm a customer... Well, I'll play devil's advocate. I'm a customer. "Man, my processes all screwed up. I got to fix this. I got all these legacy systems. I got silos everywhere. What do I do? Give me the playbook."

Alex Rinke

>> 100%. We have a playbook for that. This is good news. So, you put the process intelligence in, you connect it to your system, the process intelligence graph. First of all, you're going to get a lot of quick wins. You're going to already have paid for the projects just with the quick wins. And then, you can say, "Well, let's rethink these processes." So, then you get into our design component and you design the future state and you do it free of all the limitations that you're used to because now you have AI, now you have intelligent automation. And then, you use this composable framework to then compose it. And you want to drop Microsoft in there for agentic and Copilot? Great, put it in there. Let it run the process. You want to put workflow in there? Whatever you want to put in there, you just have a canvas to put it together and completely modernize your operations across the front office, the back office, the middle office, and you're going to do it at a speed that's unheard of.

>> Yeah, talk about the ecosystem because I was talking recently with the head of KPMG and I've talked to all the GSIs, Accenture, you name, I know them, PwC, they're all out there. But this KPMG executive said, "The way we used to do management consulting and advisory has changed significantly," one. The ecosystem of services are broader than what they can offer, but two, that the services demand was on the business model. So, what's happened is he said, "The process management mindset that was in the middle layer of the management ranks, go out and do process, shifted to the C-suite, where they say, 'Let's re-architect the business.'" Then it's got to come back down. This is where you guys come in. This is where the process intelligence layer graph comes in. But you guys have almost a system to say, "Hey, here's some optimization." So, you are essentially business model.

Alex Rinke

>> Yeah, I understand. You re-engineer-

>> Did I get that right?

Alex Rinke

>> Yeah, you have the , all the data, the operational state of the system. And then, you put in the target state, here's the design and how it should operate. And then, you can make it happen also. So, we think about analyze, design, operate. Celonis supports all three.

>> So, that's kind of like the holy grail, so business must be booming?

Alex Rinke

>> It's really exciting, what we see.

>> Give us a taste of the growth.

Alex Rinke

>> Yeah, I mean I had eight customer meetings in the last 10 days and everyone is like, "How can we move as fast as possible?"

>> "Can we go faster?"

Alex Rinke

>> "Can we go faster?" It's really, really exciting. People are also starting to really see with those capabilities that are now available in this ecosystem, how can we completely reinvent how we do things? And that's exciting because when you're talking, not just about an agent here or there. You're talking about true business transformation.

>> I was talking to Brian Baumann of NYSE Wired, and we were talking about CFOs because we had an event last year with CFOs and I wanted to test the thesis out, how well do they plugged into AI? Okay, we're going to have our friends come in. Turns out we had 30 CFOs, roughly, and they all were actually interested and knew a bit about AI. And I discovered that they're the architects of the business model, they're instrumenting the business model. So, the finance group, not obvious, they come in because money's involved, one. Savings, top-line growth, the efficiencies kick in. What's your reaction to that? Man, I'm not asking if you're calling on CFOs, but maybe you do. I'm sure you do. But the role of the CFO becomes now a stakeholder.

Alex Rinke

>> I think it's the CFO. I think it's just an example of the fact that organizations are moving from very siloed and functional. I have my HR application, HR process, and you have your manufacturing-

>> Stay in your lane....

Alex Rinke

>> procurement. Stay in your lane. You have these silos. But really now, where the processes are free to interact and you can get intelligence across, you now work in a much more collaborative model and that's much more fun. That's what people want to do anyway. So, I think it's all about teamwork. It's all about putting people in the position and empower them to actually make those changes because this change is not just driven from the top down, you need the top. I talk to the CIO and they'd be like, "Can you come meet our CFO or CEO because they want to learn about this?" So, there's a lot of interest in this.

>> Yeah, the money involved, I have to count it.

Alex Rinke

>> Exactly. And there's a lot of C-level push. "How are we agentifying? How are we using AI in our business?" And I think part of that is because it's so front-setting in the consumer world, it's like, "If it can do my daughter's or son's homework, why aren't we using it in the business?" So, I think you have this huge interest and sponsorship, but then you have every single person-

>> Yeah, it touches everyone....

Alex Rinke

>> collaborating and changing how they work and how their business works, and I think it's very empowering technology.

>> Alex, thanks for coming in and I am sure the show will be phenomenal, like last year. We'll be there with theCUBE, covering it.

Alex Rinke

>> We look forward to having you.

>> I have to ask you about agent, the current trends in agentic infrastructure. I'm hearing fleets of agents because we've talked in the past about data quality, data feeds AI. I've been having a debate with some other analysts about large language models, small language models. But if you look at the agent conversation, you're in the middle of it. You can have multiple, multiple agents talking to each other, so context becomes huge. And data quality used to be the conversation, in search and any conversation around AI or any kind of system, but you don't need to have a lot of data to get a lot of quality. You got to have the right context. And I want to talk about that role of in context to how agents will interact with each other. If I want to book a trip or I want to do a process, I might not know what I want to do yet. I'm not really deterministic yet. I'm like, "Okay, now I generate what I want. I'm going on vacation. Where do I want to go?" I search around and then it's like, "Boom, I'm going to Hawaii." Now, I got to book a hotel, now I got to go to a payment rail. That's a different agent. So, agents are going to have tasks. How do you think about that? How are your customers thinking about the role of agents, the delegation, the trust? Because one could be doing collection, one could be doing payments, one could be doing this, so you have roles of the agents.

Alex Rinke

>> You need to have this process data, that's why we call it process intelligence graph. It can associate all those different process flows, and not just the process, but the tasks, what people do manually and the desired approach as well, and even other types of information. So, it becomes this ultimate context layer where you say, "Hey, this is what's happening in other process over here. By the way, the last three cases that look similar, it went like this and it actually went into a roadblock." The more you feed into this agent, these context windows are getting bigger and bigger, the better the agent will be at coming up with the right decision, making the right recommendation, putting the right automation. So, that's why it's saying it's a very input-sensitive technology. The better the input, the better the output.

>> So, good context window helps, but also task specifically-

Alex Rinke

>> You need the foundation, and there's also some things that you want to be deterministic.

>> Yeah, exactly. Yeah, like a payment rail.

Alex Rinke

>> So, that's why we say there's no AI without PI. You need this process intelligence layer to fuel the AI. Also, you want to give the AI not just what's going on in your business, your data. You also want to tell them what's supposed to, what's the design of the process, right? What's our guardrails? How are our approval rules? When does a human need to get involved? We can provide all of this with a single API.

>> So, the agents talking to agents, sub-agents, you see that as a feature?

Alex Rinke

>> Oh, 100%.

>> 100%?

Alex Rinke

>> And you can put orchestration, you can have part deterministic. That's why we call it composable. You have these Lego blocks and you put them together to completely agentify and automate your business.

>> And that changes the process management. All right, final question. What can we expect at the event? Give us a little teaser without revealing too much. Saving for the main stage.

Alex Rinke

>> So, it's super exciting because process intelligence is really getting into the forefront because there's so much process change as a result of those agentic initiatives. So, we're going to talk about how help. The progress we made on our platform. Last year we released AgentC, this agent enablement and AI, and we talked about that last year. We've made a lot of progress. We're not going to spill all the beans right now, but it's really exciting. And we're going to put a really, really strong focus on customer examples, customers on stage.

>> Proof?

Alex Rinke

>> Proof points, because everybody wants to learn from each other. "What are you seeing? How are you doing it?" Something very new that we wrapping our heads around, so there's going to be very strong emphasis and there's no place where you can learn more about process intelligence and how people are doing it than Celosphere.

>> Yeah, and that's the marker right now. They want to see the proof. They want to see the hype cycle convert-

Alex Rinke

>> Its the show me.

>> Yeah, it's a hype cycle converting into reality. Alex, thank you for coming on and congratulations again. theCUBE bringing all the action here at the NYSE studio for theCUBE. And of course, the velocity of the company's growths are going to be depending on how well they can get agents... Celonis is really on the escape velocity. Check out their event. Of course, theCUBE will be there. I'm John Furrier. Thanks for watching.

Alex Rinke

>> Thanks, John.