Presented by Celonis
After a year of boardroom declarations about the “AI transformation,” this was the week for enterprise leaders to come together to talk about what actually works. He speaks from the stage at Celosphere in MunichCelonis co-founder and co-founder Alexander Rinke set the tone at the beginning of his speech:
“Only 11% of companies currently see tangible benefits from AI projects,” he said. “It’s not an adoption problem. It’s a context problem.”
It’s a feeling familiar to anyone who has tried to implement artificial intelligence in a gigantic enterprise. You can’t automate what you don’t understand – and most organizations still lack a unified view of what it’s really like to work in their companies.
Celonis’ answer, presented over three days at the company’s annual event, was less about recent technology acronyms and more about connective tissue: how to fit artificial intelligence into the confused, living processes that drive business. The company described this as achieving true “return on AI (ROAI)” – a measurable impact that can only be achieved when intelligence is embedded in the context of the process.
A living model of enterprise operation
At the center of the talk was what Rinke called “the living digital twin of your operations.” Celonis has been preparing for this moment for years, but this is the first time the company has made clear how far the concept has evolved.
“We start by liberating the process,” Rinke said. “Freeing it from the constraints of current legacy systems.” Data CoreCelonis data infrastructure, extracts raw data from source systems. It enables viewing of billions of records in near real-time with refresh times of less than a minute, providing visibility beyond established logging systems.
It was built on this foundation Process intelligence chart is located at the center of the Celonis platform. It is a system-agnostic, graph-based model that unifies data across systems, applications, and even devices, including task mining data that records clicks, spreadsheets, and browser activity. It connects this data to business context – business rules, KPIs, patterns and exceptions. Every transaction, rule and process interaction becomes part of a constantly updated replica that reflects how the organization actually operates.
In addition to Graph, the recent Build Experience tool enables organizations to analyze, design and operate composable AI-powered processes – integrating AI where it delivers business impact, not just technical demonstrations:
-
Analyze where processes stop or repeat
-
Design the future state by establishing AI outcomes, barriers, and touchpoints
-
Collaborate with people, systems and AI agents working in sync – now managed through a publicly available tool Orchestration engine that can trigger and monitor each step in a single flow
It is an intentional shift from discovery-driven pilots to results-driven AI operations—and a blueprint for agentic AI orchestration in which human teams, systems, and autonomous agents work together within a shared process context rather than in silos.
Real-world proof: Mercedes-Benz, Vinmar and Uniper
The Celosphere stage provided real proof of the Celonis platform in the form of live reports from customers who were already using it.
Mercedes-Benz shared how process intelligence became their “connective tissue” during the semiconductor crisis. “We had data everywhere – in plants, suppliers, logistics,” recalls Dr. Jörg Burzer, member of the management board of Mercedes-Benz Group AG. “We had no way to see it together. Celonis helped us connect the dots quickly enough that we could act.”
Since then, the partnership has expanded to eight of the company’s ten most crucial processes, from supply chain to quality to sales. But what impressed viewers wasn’t just the scale – it was a cultural shift.
“If you show data in context and allow teams to visualize processes, you will also change the culture,” Burzer said. “It’s not just a transformation of processes – it’s a transformation of people.”
On VinmarCEO Vishal Baid described Celonis as “the cornerstone of our automation and artificial intelligence strategy.” His global plastics distribution company has already automated the entire order-to-cash process for a $3 billion unit, achieving a 40% boost in productivity. But Baid wasn’t just there to celebrate the completion of the job – he was looking to the future.
“Now we are dealing with non-algorithmic issues,” he said. “Matching purchase and sales orders seems simple until you have thousands of edge cases. We are building an AI agent that can intelligently perform this allocation. This is the next frontier.”
And in the energy sector Unipertogether with partner Microsoft, they demonstrated how process-aware second-hand AI pilots are already changing the shape of operations. Using Celonis and Microsoft’s artificial intelligence stack, Uniper can predict when hydropower plants will require maintenance and cluster those tasks to reduce downtime and emissions.
“Every technician, every part, every system plays a role in a living process,” said Hans Berg, CIO of Uniper. “Humans can’t see everything. But process intelligence can do that and push the system towards the best outcome.”
Agnes Heftberger, CVP and general manager of Microsoft Germany and Austria, who joined Berg on stage, summed it up succinctly:
“The hardest part isn’t building AI features – it’s scaling them responsibly,” she explained. “You have to connect intelligence with the beating heart of the company: its processes.”
According to Celonis, throughout the global community realized business value exceeds $8 billionand over 120 certified value champions – proof that process intelligence has measurable impact far beyond pilot activities. Rinke called it “early evidence of a real turnaround in artificial intelligence.”
From closed systems to composable intelligence
Cellosphere 2025 marked the transition from architecture to interoperability – from defining AI for enterprises to enabling it to operate across borders.
Rinke’s vision for the future is uncompromisingly open: “Good things grow from open ecosystems,” he said. This philosophy is taking shape through deeper platform integration – including Microsoft fabric, Data cubesAND Bloom filter — with zero-copy, bi-directional access to Lakehouse that enables customers to query process data in-place with minimal latency. The company also announced support for the MCP server to embed the process intelligence graph directly into agent-based AI platforms such as Amazon Bedrock and Microsoft Copilot Studio.
These updates will make “composable AI for enterprises” physical – organizations can now assemble and manage AI solutions across ecosystems, rather than being restricted to a single vendor.
Rather than competing over who has the “best agent,” the message was that enterprise AI would advance as agents collaborated within a shared context and models that reflected how enterprises actually operated.
“Each seller hires their own agent,” Rinke said. “But each of them is limited to that vendor’s world. If they can’t work together, they can’t work for you. That’s what process intelligence fixes.”
The idea was met with constant applause. For companies combining multiple cloud platforms, ERP systems, and data tools, composition isn’t just elegant; it’s survival.
Beyond operations: data, democracy and direction
The final moments of the speech took an unexpected turn, from corporate architecture to human courage. Venezuelan opposition leader and Nobel Peace Prize winner María Corina Machado went live via satellite to share how her movement is using data, encrypted apps and civic coordination to expose voter fraud and mobilize millions.
It was a powerful contrast: the same principles – transparency, accountability, context – operated in both business and democracy.
“Technology can be a weapon or a liberator,” Machado said. “It depends on who holds the context.”
Her words had an impact on a room full of people used to talking about data, systems and management, a reminder that the context is not just technical, it is human.
Why this year was crucial
Celosphere 2025 marked a shift in enterprises’ approach to AI, from experimentation to process intelligence-driven outcomes. The change was evident in both tone and technology, with a stronger data core, an improved process intelligence graph, and a recent build environment. But the deeper takeaway was philosophical: AI only scales when it’s embedded in how people and systems actually work together.
Celonis CEO Carsten Thoma candidly admitted that early process mining projects often “created a storm of discovery” before they understood the organizational value – a lesson that now defines the company’s measured, pragmatic approach to enterprise AI.
Rinke said it best at the end of his speech:
“We don’t just automate steps,” he said. “We build companies that can adapt rapidly, innovate freely and continually improve.”
Did you miss it? Here you will find all the most interesting events of Celosphere 2025.
Sponsored articles are content created by a company that pays to publish or has a business relationship with VentureBeat and is always clearly marked. For more information, please contact us sales@venturebeat.com.