Subscribe
Share
Share
Embed
In this episode of The Deep View Conversations, we sit down with Alex Rinke, co-founder and co-CEO of Celonis, to unpack one of the most overlooked truths in enterprise AI.
Rinke and his co-founders started Celonis 15 years ago in Munich with just $15,000. What followed was a grind, including thousands of handwritten letters to land early customers, and a steady evolution from process simulation to what is now known as process intelligence.
Today, Celonis works with roughly half of the world’s 200 largest companies. Its platform acts like an MRI for the enterprise, creating a digital twin of how work actually happens across fragmented systems.
Rinke’s core argument is simple and provocative: there is no enterprise AI without process intelligence. Companies that deploy agents without understanding their underlying processes risk automating inefficiency at scale.
We also cover:
+ How Celonis re-engineered itself for the AI era
+ What the co-CEO model works like in practice and why it can be a competitive advantage
+ How hiring is changing inside AI-native companies
+ The tools Rinke uses to run his own workflow
If you want to understand where AI actually delivers results inside large organizations, don't sleep on this conversation.
Watch now and subscribe for more conversations at the frontier of AI.
And don't forget to sign up for The Deep View daily newsletter. We don’t just cover AI, we decode it. In a world flooded with hype, we deliver sharp, no-nonsense insights to keep you ahead of the curve and help you put AI to work every day: subscribe.thedeepview.com
Creators and Guests
What is The Deep View: Conversations?
From frontier labs and enterprise platforms to emerging startups reshaping entire industries, The Deep View: Conversations podcast interviews the brightest minds and the most influential leaders in AI.
Jason Hiner (00:01.558)
In this episode, I talk to Alex Rinke, co-founder and co-CEO of Celonis, the world's leading process intelligence company. Alex and his co-founders started Celonis 15 years ago in Munich with just $15,000 and a belief that they could help the world's largest organizations run a lot better and lot smarter. Today, Celonis works with roughly half of the 200 biggest companies in the world, but the path there was anything but easy.
Alex shared the gritty early days of bootstrapping and sending thousands of handwritten letters to potential prospects. We also talked about how the company evolved from its original idea around simulating business processes to pioneering process mining and intelligence, and how Celonis has evolved into what Alex describes as an MRI for the enterprise, giving companies a digital twin of how work actually happens across thousands of disconnected systems.
Alex makes a compelling case that there's no enterprise AI without process intelligence. And he explains why companies that try to deploy agents without first understanding their processes are setting themselves up for costly mistakes. We also talked about how Celonis is re-engineering its own internal processes in the AI era, how he's hiring differently, what the co-CEO model looks like in practice. And of course,
the AI tools that are transforming how he works. So here it is, our conversation with Alex Rinkie of Celonis.
Jason Hiner (00:01.468)
All right, so Alex, for those who aren't familiar with Celonis, tell us a little bit about what Celonis does and what your role is with the company.
Alex Rinke (00:11.008)
I'm the co-founder and co-CEO of Celonis. Celonis is the world's leading process intelligence company. We've started 15 years ago and we've built the world's leading platform to understand and optimize business processes for our clients. Business processes could be customer service, could be HR, could be supply chain, lots of processes in supply chain, could be financial processes. We can go across any environment.
and understand how this process works today to base on it, and then use AI to optimize this process to make it faster, make it better for clients, make it more efficient. And this is really becoming very quickly a core platform for our customers to industrialize enterprise AI as they are trying to automate their process using agents.
Jason Hiner (01:04.05)
Yeah. So at the end of the day, a lot of that comes down to you. You're just helping companies run a lot better. And every company in the world is always trying to run better, whether you're one person or you're, you know, a hundred thousand people or half a million people. Like you're always trying to run better. And so the work that you all do is, helping companies like always sort of spin that wheel to get, to get better and more efficient and sort of more.
Alex Rinke (01:11.544)
Correct.
Correct. Yes.
Jason Hiner (01:33.522)
more optimized, yeah? Very good.
Alex Rinke (01:35.374)
Correct. Correct. Exactly. I think Jason, you summarized it very well.
Jason Hiner (01:40.561)
beautiful. So you've had an interesting journey, you said 15 years, and now sort of we're in this very different sort of moment, which we'll talk some more about. But as I understand it, you all started, you didn't have a very easy time finding people to believe in the mission and invest in you. So talk to us a little bit about that. How did Celonis get started? And then how did you get to where you had a hard time getting people?
to invest and now you have people investing billions in your company, that's quite a journey.
Alex Rinke (02:13.782)
Yeah, true. When we started out, I think I was 21, 22, we were just coming out of college and we wanted to build an enterprise software company targeting Siemens and Bayer Pharmaceuticals. Initially it was in Europe, big European companies, but ultimately the global enterprise which we operate in today. People didn't believe us back then.
Jason Hiner (02:33.862)
Yeah.
Alex Rinke (02:41.164)
the VC infrastructure in Europe was very immature. people didn't believe us. They sort of at that time believed that young founders can start consumer businesses, but they didn't believe that they can start enterprise businesses. And especially what we do, which deals with the largest organizations and some of the most complex challenges, was not something where people said, this team can actually do it, right? So we decided to bootstrap.
Jason Hiner (02:52.626)
Hmm.
Alex Rinke (03:10.99)
for the first years and it was a very scrappy time. We were sleeping in motels with eight other people that we didn't know. were really going from client to client. At one point, we sent out thousands of handwritten letters to people to get their attention. Because how often do you get a handwritten letter? Not a lot. So we really found our way to work with large enterprise customers. Today, we work with roughly half of the
200 biggest companies in the world, a good chunk of the global 2000. then even if you go to the global 10,000 companies, we have a good chunk of those. So the company has really been able to work with and convince the biggest, most complex organizations in the world. But obviously that takes time and it takes, especially if you don't have any capital, we really started with $15,000.
it takes a fair bit of grit as well.
Jason Hiner (04:12.05)
So that's like $15,000 is like the minimum in Germany you had to raise, In order to form a business, right? Because you couldn't find anybody to invest. So you sort of scraped together everything you could, you and your co-founders and sort of got it going on that.
Alex Rinke (04:15.832)
Correct. Yes. Yes. Yes.
Alex Rinke (04:24.748)
Yeah, I mean, many countries you can start a limited corporation, a limited liability corporation with a dollar. In Germany, you needed 12,500 euros, so $15,000 round about. I remember, so each founder needed to put up 5,000 bucks and I didn't have 5,000 bucks, so I bought it from my co-founder who was a little bit older already.
So he had worked for a few years, so he had a little bit more money saved up.
Jason Hiner (04:55.58)
Very good. And then, I remembering this correctly, that for the first five years, you still couldn't get anybody to invest in you. You were still just trying to find customers and get that thing started.
Alex Rinke (05:08.044)
Yeah, I would say so. mean, at one point then we had actually quite some substantial traction. And then we didn't need capital also in the first year. So it shifted a little bit. And then in 2016, so 10 years ago, we decided to really go after the US market and become a global company. And that's when we then decided to go back and really seriously raise money. And then we raised our first round and then some subsequent financing rounds.
But yeah, initially it was tough, you and then it wasn't necessary for what we were trying to achieve and then it was necessary as we tried to scale.
Jason Hiner (05:46.833)
good and when you started it was really this focus on process mining right that was the that was the main idea tell us a little bit like double click on that for a sec for us like tell us what in process mining look like you know as you started
Alex Rinke (05:58.114)
Yeah. How do three guys wake up in Munich one day and say, we want to start a process mining company right. So we were actually, so I was studying math and my co-founders were computer scientists. we back then wanted to have this idea that we can build simulation models, which actually ties back to where we are today.
Jason Hiner (06:05.17)
Right, right.
Jason Hiner (06:21.852)
Okay.
Hmm.
Alex Rinke (06:24.799)
So we wanted to build simulation models where you could take a customer service desk or a accounting team or a number of teams and simulate the process. And take the process and simulate. if you say, hey, what if we change our staffing? What if we automate something? How much faster are we going to get? How much faster can we serve clients? What is sort of the optimal way to do this? And we actually had some success.
building a prototype and then approaching first potential clients while we were still studying at university. The first time I was interested in said, if you could simulate my IT service processes, that would be great. We came there and then we sent them a questionnaire, Give us all your existing processes, the volumes, how many calls you get that get routed this way versus that way, how many touches you have on average for a ticket.
If someone has an issue with the network, what's the process path it follows? How often does it deviate from that path? So things that we wanted to feed into our simulation engine. And we thought it's pretty straightforward because every company knows exactly how their processes are running. And they looked at us and they're like, we don't have any of this information. We have some reports, but we don't know exactly how our processes are running. It's complicated. There's different people do it differently.
Jason Hiner (07:44.432)
you
Jason Hiner (07:49.062)
Yeah.
Alex Rinke (07:52.072)
There's so many different things we support that the process can look very different. There's a lot of variation. And it was really this key insight that organizations have too little visibility into how their processes are running. In large businesses, around 40 % of the work happens on desktops outside of systems, Excel sheets, emails. There's this tremendous amount of complexity. You look at them,
Jason Hiner (08:17.362)
Mm.
Alex Rinke (08:21.165)
typical ERP process like handling orders and invoices, for example. When we look at it now, we often see this tens of thousands of variations on how they do one process. It's crazy, right? It's very, very complicated. what we found out is that visibility was missing. So then we found out about process mining, which was back then a research concept, but there was no practical adoption in businesses and there was very little tooling available.
Jason Hiner (08:35.036)
Yeah.
Alex Rinke (08:51.359)
So we said, well, before we start this whole simulation thing, maybe we should just start the ultimate process X-ray company where everybody can get an X-ray of any process in their company and understand exactly how it works, how it can be improved, where the bottlenecks are, where they spend too much time, money, effort, to do things. And that was a real, turned out to become a real success. We've obviously built around that and we can talk about how our platform has evolved. But the initial...
problem was we wanted to power models, simulation models in our case, and didn't have the right context.
Jason Hiner (09:26.569)
Wow. Okay, so this makes me think, know, Alex, you know, in these enterprises, often these systems are islands unto themselves. They're very siloed and the systems become very siloed. And so because of the systems, then the sort of people end up being siloed, the people focused on those different systems. So I have to imagine that as you're doing this, one of the challenges you're doing process mining and process intelligence as you often refer to it now.
that that becomes one of the biggest challenges. What does that look like in the context of all of the work that you are trying to do to provide this visibility?
Alex Rinke (10:06.637)
Well, over the years, and with AI, this is even accelerating further. We got really, really good at that. So we can connect to any environment. I mean, we have customers connecting us up to mainframes. We can connect to the desktops and non-system work to find out everything that happens outside of the system. So we can really get a comprehensive digital twin of what's happening in the company. And now we can even join up. So we launched this latest.
evolution of our platform a couple of years ago called the Process Intelligence Graph. I would say the latest foundation, the platform has evolved a lot since then. It's evolving very rapidly right now, as you can imagine. But this Process Intelligence Graph is really the ability to represent multiple interconnected processes. So often you have a customer service process, but it's related to onboarding, right? You can't serve as a customer, have an onboarder. If there's an issue in onboarding that might create a service issue down the road.
product activation, customer complaints, for example, in the supply chain context, ordering, shipping, buying, making things, storing things, all these things are connected processes. So we can represent interconnected processes across systems and data domains today in a way that's really, really powerful and unique. the key thing is, I think you're right, it's very complex, it's very siloed. There's a lot of variations.
The reason there's so many variations is the systems are very rigid often, the way they were implemented. So they are like train tracks, it's hard to go outside of the train track. So some folks you want to get back on the train tracks, but sometimes you also need some tracks to do the last mile or maybe because something changed or you want to go a slightly different way. So people are building a lot of customization, they're building a lot of workarounds and we can really, really help our customers with that.
Jason Hiner (11:36.946)
Yeah.
Alex Rinke (12:01.516)
by in the first instance, like really switching the MRI on, right? It's not even an x-ray, it's an MRI where it shows everything that's going on and where the issues are. It's very powerful. And again, as you said, it's very complex. Our core strength is that we're able to deal with that complexity.
Jason Hiner (12:21.628)
think that that's one of the main reasons companies want to work with Celonis is that, you know, getting these things, making them talk to each other or seeing them or the visibility and sort of one, you know, vision is really challenging. mean, it's one of the reasons that this whole idea of like,
Alex Rinke (12:26.336)
Yes.
Jason Hiner (12:39.824)
you know, tech business alignment has been sort of hanging out there for so long because it becomes very difficult. And then when your data is locked in certain systems, when you know, your, you have systems that don't necessarily talk to each other or they contextualize even parts of your business in different ways, it gets really, it gets really challenging to have like,
I'm sure you hear this all the time, but I hear it from enterprises like one version of the truth. I look at one system and it tells me one thing. I look at another system. tells me something else. And so am I, am I understanding it correctly that part of what, you know, part of the founding thesis and the, and the thesis that you're still working on is like being able to, to be able to have that level of visibility and insight, into these systems, broadly.
sort of disconnected from whatever the sort of mental models that are associated with the software itself is what companies are are so excited about what sort of working with you guys.
Alex Rinke (13:40.077)
Yes, exactly. I think then, and then we also can now encode how you want your process to work. So the design, right? And we can actually help you put it into action. And how you do that depends a little bit on the scenario, Either you want to maybe put in a new system, we can help with that and accelerate that. Maybe you want to bring AI agents in if you're already on a modern system, but you still have this problem.
You want to bring AI agents in to help you execute. You want to recompose a function. So we have many different ways on how we can also power the operationalization of this. And one is with AI agents. If you want to have an AI agent automate a process for you, you need to feed that agent the context of how the process is working. And you might need to also feed them how you want it to work, which might not be how it is working.
And then you need to monitor what the AI agent is doing to the process. So our process observability, which now is real time and cross process, becomes really, important as you have more more AI agents deployed to actually see how the process is evolving and make sure you have the guardrails in place. And you can actually explain what's going on.
Jason Hiner (14:53.276)
Very good, you're taking the questions right out of my mouth. was gonna get to there with AI agents and this observability, because agents go off and they do things and you're not always sure what they're doing or how. And so that's become one of the challenges that a number of companies are trying to face. So that's great that that's sort of already part of your thinking. So how did the company evolve from really focused on this idea of process mining, process intelligence,
Alex Rinke (14:56.001)
Ha
Jason Hiner (15:22.578)
to when the AI boom happened. Obviously, you all predated the sort of current AI boom by nearly a decade, over a decade. so when all of the current sort of generative AI explosion happened, how did you look at that and decide what that meant for you all and for the work that you do?
Alex Rinke (15:49.375)
Yeah, it's very clear from the beginning that it's very exciting. And the reason it's so exciting is because to implement AI in a company, you need to change all the processes. If you don't change your processes to actually leverage AI, you're not going to take the advantage and you're not going to get the benefit. So we have a customer, a big utility, Unipa, they've created millions and millions of savings.
Jason Hiner (15:59.805)
Okay.
Alex Rinke (16:17.822)
across 27 processes. One example is, they are utilities, so they maintain a lot of equipment. They have a maintenance agent coordinating and orchestrating all the maintenance that they do around their equipment. And we have process intelligence feeding that agent because you needed to really understand in the different regions, in the different settings, how is that maintenance being triggered, executed, what guardrails are around it. So we could provide that insight.
very, very quickly with a high degree of precision, had them adjust that process and then feed it into the AI and then also monitor the AI to make sure it's doing the right thing. So process intelligence was very clear that it is an essential context layer for artificial intelligence because you need to ground the artificial intelligence in the processes of that respective business. as sort of process nerds and process engineers, we understood that very quickly that there was a very significant amount of process transformation.
associated to actually making these models, you know, save customers money and work in an enterprise reality. And I think we're in the midst of that. I think we're in the early stages of that. A lot of companies haven't substantially automated their processes yet. Maybe they've automated code generation quite a bit. Maybe they've automated parts of customer service. But if you're talking back office, supply chain, front office, really complex operations, it's the very beginning.
Jason Hiner (17:32.658)
you
Jason Hiner (17:46.291)
I'm glad you mentioned supply chains, because that is one of the world's most complicated processes. And so I have to think that a number of your customers or the customers that are most maybe enthusiastic about what you do tend to be people that have very complicated supply chains. Because if you can optimize those processes, you can save tons of money, you can make it more efficient, you can make it faster, all of those kinds of things.
We've come through an era where there, you know, we had a lot of supply chain disruption during the COVID era. Now we've sort of hit another era, another gear with some of the geopolitical challenges of, of supply chain, disruptions happening again, and anticipations of others, that the AI boom itself is in some ways constrained by some supply chain challenges. So with all of that in mind, I have to think that supply chain is some of the
people that come to you wanting help with processes because that matters a lot for those folks.
Alex Rinke (18:47.307)
100%. I mean, we work, you mentioned oil and gas, we work already with both the top five oil and gas players as an example. We work with many of the top life science players, many of the top OEMs. Mercedes is going really, really big across their processes. They have over 30 different key production plans where they have to optimize the supply chain. And as I mentioned, our ability to one incorporate AI deeply into what we do to speed up
Jason Hiner (18:54.149)
Okay.
Alex Rinke (19:17.385)
understanding the processes even better to speed up bringing together disparate data sources to speed up understanding, really generating insights from all this information that we gather. So that really helps us. And the other thing, the ability to represent these interconnected processes really, really created a hockey stick of our growth in supply chain related use cases, because supply chain related use case, you always need those multiple processes in combination, a single process of you.
which is where we started, is not enough for a supply chain use case. You need to sync up procurement with demand, with auto management, with delivery, with inventory, with manufacturing, depending on the industry. You need to link all of those things up in order to generate meaningful outputs.
Jason Hiner (20:05.212)
Very good. So if I'm using your product, you know, in my business, which part of the, you know, who in the company is using it the most? it the tech team, you know, IT? Is it operations? You know, is it, yeah, who, are there certain parts of the company that tend to be most, yeah, big users of your team, of product?
Alex Rinke (20:27.945)
No. You know, in the past, I would say it's been very line of business driven. What we see more more though is that, it's really, we actually just have a big meeting going on with a lot of CIOs that leave us alone. The CIO has really embraced it in the last few years because they understand it's a horizontal platform capability. You have processes everywhere. You need process intelligence everywhere.
Jason Hiner (20:34.311)
Okay.
Jason Hiner (20:47.376)
Okay.
Alex Rinke (20:54.411)
to enable artificial intelligence everywhere. Or do you want to gentrify everything in your business? And in order to do that, you need end-to-end process intelligence. It's necessary to have the right context and to do it in a high quality and right way. Because the quality and precision and guardrails in how effectively you can deploy AI into your business is going to be the competitive advantage. It's too slow right now still for many companies. And our customers are...
Jason Hiner (20:58.662)
Hmm.
Alex Rinke (21:21.201)
accelerating. Our customers are doing it, putting it in production. We're not talking about pilots. We're talking about production rollers, and they're transforming very substantial parts of their operations at a speed that is unrecognizable to what we would have assumed a few years ago. Maybe even a few months ago, even with the recent model evolution, we've already incorporated that into our product to speed up implementation dramatically as an example. You really see that AI
is affecting every single process and process intelligence is a foundation to industrialize it.
Jason Hiner (21:58.013)
Very good. So when the AI boom happened, was that when you changed the name of what you do from process mining to process intelligence to help people understand that this is a layer of contact, know, AI is very data hungry. AI is very context hungry. And what I hear you saying is like, you realize very quickly, like, we have very valuable data that can help guide AI and now AI agents.
in ways that are going to be incredibly useful and precise in working with these things.
Alex Rinke (22:34.539)
100 % yes. I think it was correct, yes. And we changed the process intelligence, but also because it expanded a lot. We do process predictions, we do cross process, we do task, combined task and process. So it goes beyond process mining and it has for a while. The other thing is we explain to people there's no AI without PI. There's no AI without process intelligence.
Jason Hiner (22:44.293)
Okay.
Jason Hiner (23:00.06)
Sure.
Alex Rinke (23:00.381)
at least in the enterprise, right? You could say there's no enterprise AI without PI. And that's really resonated with people. They understand that. They understand that they need to understand, know, base on their processes, change them and create this repository that AI agents can reason off of to make sure they are precise and they're doing the right things. And they also need this cross process element because you don't want an AI agent to optimize one thing, but then screw up something else, right? Like imagine you had an AI agent automate all your orders.
Jason Hiner (23:05.776)
Okay.
Alex Rinke (23:27.263)
But then when you send those customers invoices, it did something wrong and you don't get your money. That would be a problem. So you need cross process intelligence to also follow and compare what the agents are doing.
Jason Hiner (23:32.914)
Yeah
Jason Hiner (23:41.011)
So is what you're doing is process intelligence, putting this process intelligence layer into a company when they implement Celonis. Is it reconceptualizing or is it like a new mental model for thinking about your business through the lens of processes rather than through the lens of systems or lines of action or that kind of thing? Is it another lens or is it even sort of re-imagining sort of your business from the ground up through the...
Alex Rinke (23:48.063)
Yes. Yes.
Jason Hiner (24:10.31)
the lens of process and processes.
Alex Rinke (24:11.882)
Yeah, I think the unit that really powers agentic automation is the process, right? Because that's what ultimately drives the outcome. So I think it is a different lens. I think we are unique in the sense that we look at a business as a collection of interacting processes. And I think it's the right mental model for AI because the atomic unit of an agent is really a process, right?
Jason Hiner (24:17.211)
Okay.
Jason Hiner (24:35.878)
Yeah, want, so you're the agents, you want them to accomplish things for you. Ultimately, at the end of the day, you know, a chat bot is trying to provide you with information or context and an agent is trying to carry out tasks and processes.
Alex Rinke (24:50.352)
And they need to be chained together in a logical way so then you have a processor.
Jason Hiner (24:54.502)
Very good, very good. Okay, well, Alex, I'd like to ask you little bit about how has your role evolved from founding through the journey of these sort of 15 years, how has your role evolved in the company? What does it look like at the beginning compared to what it looks like now?
Alex Rinke (25:13.598)
Yeah, so I think it actually evolves quite substantially. In the beginning, you are doing everything, right? Because there's only three people. Then I think as the company grows, you have to learn how to take on managerial responsibility. And by the way, lot of founders at one point, it's challenging, right? And not everybody likes that evolution or embraces that. So you need to go from running
Jason Hiner (25:20.912)
Yeah, yeah.
Alex Rinke (25:42.869)
a three-person organization to a few hundred people, and then ultimately a few thousand people. And then you really need to become an executive. So I always say there's founder, there's leader, and then there's executive. And you always want to keep that founder mindset, right? That doesn't go away, but you also need to have that other persona that you need to learn. And for example, an executive needs to build and align an executive team. And it's very different from running a few hundred people where everybody knows each other and...
sort of through gossip, you know, the word gets out. I like to talk about this analogy of the founder microphone. So when the company is small, you have this microphone and you say something into the microphone and it happens, someone picks it up and it works. And then when the company gets bigger, you're like, like mic test, test, test, know, like your microphone, you know, it doesn't work anymore. And then you need to really learn how to create a management system and an executive team.
Jason Hiner (26:29.906)
Mike's not working.
Alex Rinke (26:38.418)
and align everybody around common objectives. Now I think it's changing again a little bit where I think that original founder persona is more important because we are looking at such a significant transformation of how organizations work internally, of how your products work, of how you're building your products, how you think about engineering.
So I would say what founders are usually good at is they're creative, they have strong convictions, right? They don't easily get sort of thrown off the path that they want to be on. And I think some of this comes back and obviously you apply differently to an organization of a few thousand versus an organization of a few people, but I think it's changing again.
Jason Hiner (27:33.116)
What do you spend your time on? What do you spend your days on these days? What stuff takes up the most, yeah, bandwidth for you, or mental tokens, as people like to call it these days.
Alex Rinke (27:43.323)
Mental tough.
But I'm thinking that this hasn't changed. I'm always thinking about three things. Our customers, our products and our people, our team, right? Those three things have to, you know, are the key sort of things that make the company better, right? If you have a great team, they're all aligned, they're marching in the right direction, it's a great organization. If you have great products that really resonate with your customers, your customers are happy with the service and the value they're getting, right? If you do those three things well, you know,
That's sort of always been my focus. And I think each one of those dimensions changes in the AI era, right? Our products change significantly. The way we approach our customers and the message and the way we engage changes quite dramatically and also internal organization evolves very dramatically. So I'm thinking of those three things, but really in an AI first way.
and leveraging our unique strength to amplify the impact of AI on our products and our customers and on our organization is really, really exciting. So it hasn't been as exciting. It's probably the most exciting time maybe since we started the company right now. It feels almost like a refounding moment.
Jason Hiner (29:06.418)
I'm hearing that a lot from folks. So many things are evolving the way that they run the company and the way that they operate. How about your team internally using AI? How are you all using AI internally to transform what you do to rethink things and to help you take the next step?
Alex Rinke (29:31.466)
And it's very interesting because you this same challenge that our customers have, that you give broad access to AI to your company, but then you have to speed up all the processes so that people actually get faster and things actually get done better. So I'll give you an example. One focus is obviously coding and engineering. So everybody has access. Pretty much everybody at Celonis is using AI in one shape or form every day.
So we have broad adoption, but now we have to re-engineer the processes. So for example, in engineering, you used to have a very different model. The way we do engineering now is that we get an empowered decision maker with a few people in a room working on complex problems, producing outcomes every single day and launching products much, much faster. Because what we found is with coding agents, you can produce code much faster.
But now you also need to create the infrastructure and the process to make decisions much quicker, to launch products much quicker, to speed everything up. So it's really been, and we use Celonis a lot internally to understand where those bottlenecks are and how we can tackle them. So for example, in software engineering, was speed of specification decision-making needed to be dramatically accelerated and the organizational model needs to evolve to be able to do that at scale. We are looking at how we deliver our product.
we can automate implementing Celonis with AI, which means that our customer, when they used to be able to do three processes in a time frame, they can now do 30. They can go across all of their systems. They can get insights across everything that they do. They can get end-to-end process intelligence, which one of our customers, spoke to them two weeks ago, they're a large bank. They have 4,000 systems just in their markets. 4,000.
Jason Hiner (31:26.866)
Wow.
Alex Rinke (31:27.975)
So how do you get all this process intelligence across 4,000 systems? Well, we used to be, in the past we would have said, hey, show us which ones are your most important ones and we can focus on those. Now we can say, let's go. So it's a completely different way in how we implement, how we build products, but then also all the other operational processes in the company. So I think in those three buckets, how we build products, how we implement those products. And then the third is how we,
roll it out in the organization across HR, across finance, go to market. How do we in AI enable and identify everything? And I think the first step is you give everybody access and you get that sort of momentum going across knowledge work. In parallel, you then need to really work on your end-to-end processes and speed them up and re-engineer them so they can actually take advantage of this AI speed.
Jason Hiner (32:20.37)
What's the hardest thing about implementing AI within the company today? And what's the thing that sort of you wish could be moving a lot faster?
Alex Rinke (32:29.673)
Well, I think that process transformation and cultural transformation are the biggest bottlenecks. So process transformation is because of all the things we talked about, context is spread. You have to make sure it's inserted in the right parts of the process. You have to make sure you remove other bottlenecks. You have to make sure that you really reinvent, How do we want to do customer service? How do we want to do sales? How do we want to go to market? So there is a significant part in process transformation. And then secondly, you need to bring people along.
on this mission. Obviously, there's going to be some resistance. mean, good thing is Ceylon is we're still a pretty young company and it's a technology native company. So our team is very receptive to it and excited about it. But still, you need to bring people along. Some organizational structures have to evolve. You have different organizational models. And some of this is also something that's changing because AI is progressing pretty rapidly.
So you thought it couldn't do something three weeks ago and today you're like, it can actually do it. So you're just one model iteration away from it being able to do it. I mean, people are still tremendously valuable. Don't get me wrong, right? I'm not one of those folks that say, hey, you don't need people anymore. You can just have AI run your company. I think the question is how do you leverage, get much, much more leverage out of the people.
Jason Hiner (33:34.098)
Now it does. Yeah.
Alex Rinke (33:57.616)
and how do you get much, much more leverage out of the organization? How do you get much, much more leverage out of the process? And that requires significant change.
Jason Hiner (34:07.986)
How are you hiring differently based on all the changes that have happened and are happening? What does that look like for you? Are you still hiring entry-level jobs? is one of the big things that's hanging out there. The entry-level jobs are one of the first things that have been automated. How do you all think about it? I think you have about, if I'm remembering right, about 4,000 employees today. that? Okay.
Alex Rinke (34:29.523)
Yeah, close to for that.
Jason Hiner (34:31.506)
And so you have a fairly large team. Yeah, what does it look like as you're thinking about the people that you bring on board from this point forward?
Alex Rinke (34:41.501)
Well, I think we obviously incorporating AI into our hiring. We are testing a lot for how versatile people are. You really need people that are versatile. Secondly, you need people that are able to go deep into certain areas because that depth is ultimately if you're paired with AI, what gets the real acceleration. Like we see, for example, now engineering organization, some of our deepest
engineers that are the most creative, most knowledgeable, get the largest lift, right? Because then you turn a 10x engineer into a 50x engineer, a 100x engineer. So we really test for those abilities and we're changing the way we're approaching interviewing. We always do a challenge in our interviewing process. So we've changed all those challenges to incorporate AI and we
ask a lot of questions around the cultural attitude of people, because that matters a lot when you're in a time where things change so rapidly. Are they intellectually honest? Are they curious? Curiosity is super, super important. Are they willing to change, to evolve? Are they not stuck in their ways? Anything that can come both in entry-level and more experienced positions. Clearly, we are not just thinking about, hey, which people do we not need anymore? We are thinking a lot more about
with there is so much growth for us and so much opportunity ahead. How can we get 10x leverage out of people, right? Because there's a lot of processes there where process intelligence is very meaningful. We see a tremendous amount of demand because everybody wants to gentrify their operations and their processes. So we are quite excited to use this time to play offense.
Jason Hiner (36:31.73)
Very good. How about as you think about the future, if you think about sort of the next stage of things, we're in this very much agentic moment in 2026 and there's a lot that's happening there. we know that and we see the progress, the pace of change and new models and new tools. As you think forward, you know, for Celonis and for the industry, you know, what are you most excited about?
What are you most looking forward to? What are the things that you, know, or maybe even a little bit anxious about in terms of the future of AI and tech?
Alex Rinke (37:09.927)
Well, I think process intelligence is something very hard and transforming processes is very hard, right? It's not something that's easy. And I think with AI it can get much, much faster and much, much better because we have those reasoning capabilities available, right? That help us in every step of what process intelligence and process transformation does. And it really amplifies the category because
where it used to be for people that really wanted to be excellent at processes and that had a focus on this. Now everybody has a focus. You hear everybody talking about processes and we're like, yeah, we agree. We have for a while. So I think it's a tremendous opportunity for the space and it's a tremendous opportunity for Saloners. I think what I'm worried about is what every CEO is worried about is how do we make sure we move fast? And we are moving very fast right now. We are moving faster than we've ever moved.
I think that's really the key. You need to adapt, you need to move fast and you need take advantage of the opportunities out there.
Jason Hiner (38:17.938)
co-CEO role. You are the co-CEO. It's not a very common arrangement, right, to have multiple CEOs. How does it work for you all? How do you, you know, divvy up some of the work? Decide who focuses on what and how well has it worked for you all?
Alex Rinke (38:34.44)
Yeah. So I've always been more focused on the US market and product and Bastia is more focused on the other markets and also more on the operations of the company. Obviously, as we've evolved, we've grown our executive team that's always shifted around a little bit. Now it's a huge blessing.
Jason Hiner (38:56.903)
Mm.
Alex Rinke (38:57.416)
because we can divide and conquer so well. We've known each other for 15 years. It's not one of those things where, oh, actually longer than 15, it's almost 20. It's not one of these things where you have two professional executives that didn't know each other that come into a company and everybody wants to be sole CEO, right? That's not our model. Our model is we've been friends for a long time. We have a president as well who's very, very good. And we really divide and conquer and it's a blessing now because...
So many things are changing that you have that ability to divide and conquer. And for example, I have the ability to work much closer with our product teams on the next evolution, which is going to be very exciting. I mean, we deeply, obviously deeply incorporating AI and combining what we do with the model capabilities out there.
even better than we do it today. mean, those things are extremely important and having multiple founders to be able to divide and conquer is, think, is a blessing. And you know, our third founder, CTO, he's fully involved. all three of us have been fully involved for the whole time.
Jason Hiner (40:08.018)
So it's like double or triple the bandwidth of just what one executive might have. Yeah, very good. No, it's impressive that you've made it work so well, for sure. All right, I have question for you. Staying on the topic of leadership, one of the things in the AI era that I hear again and again from leaders is having to really be thoughtful about their time.
Alex Rinke (40:11.26)
Yeah, that's how I look at it for sure.
Jason Hiner (40:34.034)
and having to get leaders often express it as like, how do I get maximum leverage? How do I spend my time on the things that have the biggest impact? And so I love to ask leaders sort of what your tip is for that. How do you find the things that you can get maximum leverage spending your time on?
Alex Rinke (40:54.15)
I always think about where do we want to be in six months from now, in 12 months from now, in two years from now, and what needs to happen for that to be true, and how can I lean into the things where I can really provide a unique perspective, a unique difference. So right now, for example, know, again, I'm very, very closely with, probably closer than in a long time with our product and engineering teams, and it's a lot of fun.
And you can move things so much faster because of coding agents. I'm using coding agents every day and it's exciting, right? It's so much fun. You can move things better. And I think it's what used to be true is even more true that you really need to focus and pick your battles and make sure that you have a strong organization to cover all the bases, right? And you empower the right people to get the collective job done and accomplish the collective mission. I think you have to be even more alert.
and even more focused than ever because things are moving faster than ever and the opportunity is huge.
Jason Hiner (41:53.042)
Can I ask you what coding agents, what coding tools do you use?
Alex Rinke (41:56.873)
All of the above, mean, recently Claude has really made a big progress, but Codex is great. I mean, we use all of the above. again, I'm also experimenting with lots of things, right? I always want to stay at the cutting edge. The tool that recently is not a coding agent, but that's really cool is Whisper Flow. don't know if you've tried it, but it's a transcription. It's super cool. that is, you know, I type pretty fast, but now just...
whisper flowing it is even faster.
Jason Hiner (42:29.33)
That's amazing. was going to ask you last thing I usually ask is what's the coding tool, sorry not coding tool, but the AI tool you've been using lately that really has made a difference for you. Whisper Flow, are there any others that you've been using lately that you really like?
Alex Rinke (42:41.576)
I've been using Celonis a lot more with our latest evolution, which is really, really cool to help drive this process transformation internally. I mentioned the coding agents. Yeah, I think the ability for me, which I think I'm effectively like a very product oriented founder and CEO.
The ability for me to sketch out new ideas and architectures and product capabilities and just actually make them work for then obviously the engineers to incorporate them into our platform and really be specific about the ideas and iterate has been amazing. I mean, it's game changing. So that's been the biggest accelerator for me. Where we used to spend more time on alignment and then spend a bunch more time in actually producing the features, we can be so much faster.
Jason Hiner (43:25.97)
So like vibe coding, prototyping.
Jason Hiner (43:35.26)
Very cool, so you use some of the Vibe Coding tools to sort of make prototypes. Yeah, yeah, very cool, very cool. Well, Alex, thank you so much for your time. Appreciate you being here. This has been a great conversation and yeah, good luck with all the things that you're working on. Thank you.
Alex Rinke (43:38.042)
A lot, yes. Yes.
Alex Rinke (43:49.651)
Jason, this was fun. Appreciate it. Thank you. Talk soon. Bye bye.