The letters were handwritten because the software was hard to explain. In the early 2010s, Alexander Rinke and his two co-founders were trying to sell large companies a view of themselves that those companies had never seen. Their product could reconstruct the path of an invoice, a service ticket or an order from the digital crumbs left in corporate systems. The promise was simple once it appeared on a screen. Getting invited into the room was the difficult part.
So the young founders reached for pen and paper. They sent thousands of letters to prospective executives, traveled from client to client and slept in motel rooms shared with strangers. Rinke says he drove well over 100,000 kilometers a year during each of the first two or three years, crossing Germany to explain a category called process mining. There was no polished demand engine. There was repetition, curiosity and the useful stubbornness required to teach a market.
The image is almost too neat: a company built to read digital traces beginning with ink, envelopes and a lot of road. Yet it captures Rinke’s operating style. The technology matters. The customer’s reaction matters more. Years later, when asked where new Celonis products begin, his answer stayed plain: with customers from day one.
A simulation idea meets messy reality
Rinke was born in Munich in 1989 and went to school in Berlin. Entrepreneurship arrived early in practical form: as a student he started a tutoring agency. He later studied mathematics at the Technical University of Munich, spent an exchange semester at École Polytechnique in Paris, worked on algorithmic trading systems at KK Invest and took roles with Munich Re in Munich and Hong Kong. The path supplied a blend of abstract models and institutional complexity.
At TUM, Rinke met Bastian Nominacher and Martin Klenk through Academy Consult, a student consultancy. A project for the broadcaster Bayerischer Rundfunk asked the team to improve an IT service process. Conventional methods did not reveal enough. The students used event data to reconstruct what was actually happening and reduced a service process that had taken days. The exercise exposed a wider opportunity hiding inside ordinary software logs.
Rinke’s original instinct was mathematical. He thought process data could generate simulations of how a business worked. Then came the awkward discovery: before companies could simulate their operations, many first needed an accurate picture of those operations. Systems had been customized. Workarounds had multiplied. A process described in a manual could differ sharply from the route taken by the work itself.
That prerequisite became the product. Celonis would make a process visible, attach statistics to the visual map and let a user move from a strange branch to the invoices, suppliers or bottlenecks underneath it. The founders turned the university project into a company in 2011. Their first office was an apartment. Their starting capital was the €12,500 minimum needed to form the German entity. Rinke has recalled that he did not have his portion, so an older co-founder covered it.
The broadcaster project supplied more than an idea. Bayerischer Rundfunk recommended the team to Siemens, creating the kind of warm passage into a major account that a new category needed. Even the company name carried the founders’ posture. Celonis drew inspiration from Celos, a figure associated with aspiration and striving in Greek mythology. Rinke said they had a grand vision from the beginning, but its delivery depended on a chain of very small proofs.
The strange advantage of having no money
European investors at the time were more comfortable with young founders building consumer products than selling software into Siemens or Bayer. The Celonis team decided to bootstrap. The absence of capital forced a useful sequence: win customers, earn revenue, improve the product, repeat. Within weeks they had an early customer. For five years they expanded without outside financing, which meant that by the time they raised institutional money in 2016, they were funding a working engine rather than a theory.
Bootstrapping also put Rinke in a role that mathematics alone could not prepare him for. He had to translate a technical capability into economic language. A process map was interesting. A delayed payment, a compliance exception or a blocked order had a cost. The sale became easier when the picture could point to the number beneath it.
The company’s rise brought visible markers. TUM gave the team its Presidential Entrepreneurship Award in 2015. Rinke appeared in Forbes Europe’s 30 Under 30. In 2019, German president Frank-Walter Steinmeier presented Rinke, Nominacher and Klenk with the Deutscher Zukunftspreis for their work on process mining. Investment rounds followed, including a $1 billion Series D in 2021 and an extension in 2022.
Those numbers can flatten a founder story into a valuation chart. The more revealing thread is the continuity. Rinke still describes his work through three subjects: customers, products and people. The nouns survived the shift from an apartment to offices in Munich and New York. What changed was how he could act on them.
When the founder microphone cuts out
Rinke has a playful analogy for company scale. In the beginning, a founder speaks into an imaginary microphone and something happens. Everyone hears the instruction because everyone is close. As the organization grows, the microphone seems to stop working. Saying the same thing louder cannot restore the old intimacy. The founder needs an executive team, common objectives and a system that carries decisions without requiring personal transmission.
He divides the evolution into three identities: founder, leader and executive. Each asks for a different craft. The founder creates. The leader takes managerial responsibility. The executive aligns leaders across a company that no longer runs on hallway knowledge. Rinke’s point is not to discard the earlier identities. The creative conviction of the founder remains useful, especially when technology forces a mature company to question its assumptions again.
That has become relevant in the AI cycle. Coding agents can produce software faster, but specification, review and product-launch decisions must accelerate too. Giving employees new tools while leaving the surrounding workflow untouched creates a faster section inside a slow system. Rinke says Celonis uses its own product internally to find these bottlenecks. The company is applying the premise it sells: technology pays off when the process around it is redesigned.
When one step becomes dramatically faster, inspect every decision and handoff around it. Local speed often exposes the next organizational bottleneck.
The next map is for machines
The newest version of Rinke’s argument begins with a limitation of enterprise AI. A general model may know public facts, but it does not automatically know which approval route a particular company uses, why an invoice is blocked, how a customer-service exception affects an order or which person should make a judgment. That operational context is scattered across systems, rules, desktops and institutional habit.
Rinke understands a company as a collection of interacting processes. That view has pushed Celonis from individual process maps toward a graph connecting service, onboarding, procurement, finance and supply-chain flows. In May 2026, the company introduced the Celonis Context Model, described as a living digital twin that combines process data, business knowledge and decision intelligence. Celonis also agreed to acquire Ikigai Labs, adding simulation and forecasting work to that model.
His aspiration has widened with the product. Rinke talks about the top line, the bottom line and a green line, arguing that operational improvement should account for environmental impact alongside revenue and cost. He has also promoted wider access to education through free Celonis software and training for students and universities. The connective idea is practical: if more people can see where work, materials or time are being lost, they have a better chance of changing the outcome.
The ambition loops back to the student idea. Rinke wanted to build simulations, then discovered that the underlying context was missing. Fifteen years later, context has become the central product question again, this time because AI agents need a reliable account of how work moves before they can act on it. The original detour now looks like preparation.
Rinke calls the moment a refounding. It is an apt word for an executive trying to recover the alertness of the founder without pretending the company is still three people in an apartment. The tools are different. The sales letters are gone. The discipline underneath remains recognizable: observe the real workflow, find the blockage, sit with the customer and improve what happens next.
There is a quiet personal symmetry here. The young mathematician wanted a model precise enough to simulate a business. The operator learned that the model first had to survive messy data, skeptical buyers and thousands of kilometers of real roads. The executive now wants machines to understand that same mess well enough to be useful inside it. Alexander Rinke’s career has been an extended argument that visibility comes before optimization - and that seeing the work clearly is itself a form of progress.