Celosphere 2025: The place enterprise AI moved from experiment to execution

Offered by Celonis

---

After a yr of boardroom declarations about “AI transformation,” this was the week the place enterprise leaders got here collectively to speak about what truly works. Talking from the stage at Celosphere in Munich, Celonis co-founder and co-CEO Alexander Rinke set the tone early in his keynote:

“Solely 11 % of corporations are seeing measurable advantages from AI tasks at present,” he mentioned. “That’s not an adoption drawback. That’s a context drawback.”

It’s a sentiment acquainted to anybody who’s tried to deploy AI inside a big enterprise. You may’t automate what you don’t perceive — and most organizations nonetheless lack a unified image of how work of their corporations actually will get carried out.

Celonis’ reply, showcased throughout three days on the firm’s annual occasion, was much less about new tech acronyms and extra about connective tissue: easy methods to make AI match throughout the messy, dwelling processes that drive enterprise. The corporate framed it as attaining an actual “Return on AI (ROAI)” — measurable affect that comes solely when intelligence is grounded in course of context.

A dwelling mannequin of how the enterprise works

On the coronary heart of the keynote was what Rinke known as a “dwelling digital twin of your operations.” Celonis has been constructing towards this second for years — however this was the primary time the corporate made clear how far that idea has developed.

“We begin by liberating the method,” mentioned Rinke. “Liberating it from the restrictions of your present legacy programs.” Information Core, Celonis’ knowledge infrastructure, extracts uncooked knowledge from supply programs. It’s able to querying billions of information in close to actual time with sub-minute refresh — extending visibility past conventional programs of file.

Constructed on this basis, the Course of Intelligence Graph sits on the middle of the Celonis Platform. It’s a system-agnostic, graph-based mannequin that unifies knowledge throughout programs, apps, and even gadgets, together with task-mining knowledge that captures clicks, spreadsheets, and browser exercise. It combines this knowledge with enterprise context—enterprise guidelines, KPIs, benchmarks, and exceptions. Each transaction, rule, and course of interplay turns into a part of a repeatedly up to date duplicate that displays how the group truly operates.

On prime of the Graph, the corporate’s new Construct Expertise permits organizations to investigate, design, and function AI-driven, composable processes — integrating AI the place it delivers enterprise affect, not simply technical demos:

• Analyze the place processes stall or repeat

• Design the long run state, setting outcomes, guardrails, and AI touchpoints

• Function with people, programs, and AI brokers working in sync — now orchestrated by a usually accessible Orchestration Engine that may set off and monitor each step in a single stream

It’s a deliberate shift from discovery-driven AI pilots to outcome-driven AI operations — and a blueprint for orchestrating agentic AI, the place human groups, programs, and autonomous brokers work collectively by shared course of context somewhat than in silos.

Actual-world proof: Mercedes-Benz, Vinmar, and Uniper

The Celosphere stage supplied actual proof of theCelonisPlatform in motion, by dwell tales from prospects already constructing on it.

Mercedes-Benz shared how course of intelligence turned their “connective tissue” through the semiconductor disaster. “We had knowledge in every single place — vegetation, suppliers, logistics,” recalled Dr. Jörg Burzer, Member of the Board of Administration of Mercedes-Benz Group AG. “What we didn’t have was a option to see it collectively. Celonis helped us join these dots quick sufficient to behave.”

The partnership has since expanded throughout eight of the corporate’s ten most crucial processes, from provide chain to high quality to after-sales. However what impressed the viewers wasn’t simply the dimensions — it was the cultural shift.

“In the event you present knowledge in context, and let groups visualize processes, you additionally change the tradition,” Burzer mentioned. “It’s not simply course of transformation — it’s individuals transformation.”

At Vinmar, CEO Vishal Baid described Celonis as “the muse of our automation and AI technique.” His world plastics distribution enterprise has already automated its complete order-to-cash course of for a $3 B unit, attaining a 40 % productiveness carry. However Baid wasn’t there to simply rejoice completed work — he was trying forward.

“Now we’re tackling the non-algorithmic stuff,” he mentioned. “Matching buy and gross sales orders sounds easy till you’ve gotten 1000’s of edge instances. We’re constructing an AI agent that may do this allocation intelligently. That’s the subsequent frontier.”

And within the vitality sector, Uniper, with associate Microsoft, demonstrated how process-aware AI copilots are already reshaping operations. Utilizing Celonis and Microsoft’s AI stack, Uniper can predict when hydropower vegetation will want upkeep — and cluster these jobs to cut back downtime and emissions.

“Every technician, every half, every system performs a job in a dwelling course of,” mentioned Hans Berg, Uniper’s CIO. “The human can’t see all of it. However course of intelligence can — and it could possibly nudge the system towards one of the best end result.”

Agnes Heftberger, CVP & CEO, Microsoft Germany & Austria, who joined Berg on stage, summed it up crisply:

“The onerous half isn’t constructing AI options — it’s scaling them responsibly,” she defined. “It’s worthwhile to marry intelligence with the beating coronary heart of the corporate: its processes.”

Throughout the worldwide group, Celonis reviews greater than $8 billion in realized enterprise worthand over 120 licensed worth champions — proof that course of intelligence is driving measurable affect far past pilots. Rinke known as it “the early proof factors of a real return on AI.”

From closed programs to composable intelligence

Celosphere 2025 marked a shift from structure to interoperability — from defining enterprise AI to creating it work throughout boundaries.

Rinke’s imaginative and prescient for the long run is unapologetically open: “Good issues develop from open ecosystems,” he mentioned. That philosophy is taking form by deeper platform integrations — together with Microsoft Cloth, Databricks, and Bloomfilter — with zero-copy, bidirectional lakehouse entry that lets prospects question course of knowledge in place with minimal latency. The corporate additionally introduced MCP Server assist for embedding the Course of Intelligence Graph instantly into agentic AI platforms like Amazon Bedrock and Microsoft Copilot Studio.

These updates make “composable enterprise AI” tangible — organizations can now assemble and govern AI options throughout ecosystems somewhat than being locked into any single vendor.

Quite than competing on who has the “finest agent,” the message was that enterprise AI will thrive when brokers work collectively by shared context and fashions that mirror how companies truly run.

“Each vendor is bringing out their very own agent,” Rinke mentioned. “However every one is proscribed to that vendor’s world. If they will’t work collectively, they will’t be just right for you. That’s what course of intelligence fixes.”

The thought drew sustained applause. For corporations juggling a number of cloud platforms, ERPs, and knowledge instruments, composability isn’t simply elegant; it’s survival.

Past operations: knowledge, democracy, and route

The closing moments of the keynote took an surprising flip — from enterprise structure to human braveness. Venezuelan opposition chief and Nobel Peace Prize winner María Corina Machado joined dwell through satellite tv for pc to share how her motion used knowledge, encrypted apps, and civic coordination to show election fraud and mobilize tens of millions.

It was a strong distinction: the identical rules — transparency, accountability, context — at work in each enterprise and democracy.

“Expertise is usually a weapon or a liberator,” Machado mentioned. “It relies on who holds the context.”

Her phrases landed with weight in a room full of individuals used to speaking about knowledge, programs, and governance — a reminder that context isn’t simply technical, it’s human.

Why this yr mattered

Celosphere 2025 marked a shift in how enterprises method AI — from experimentation to outcomes grounded in course of intelligence. The shift was evident in each tone and expertise, with a extra highly effective Information Core, enhanced Course of Intelligence Graph, and new Construct Expertise. However the deeper takeaway was philosophical: AI solely scales when it’s grounded in how individuals and programs truly work collectively.

Celonis president Carsten Thoma was candid in acknowledging that early process-mining tasks usually “stormed in with discovery” earlier than understanding organizational worth — a lesson that now defines the corporate’s measured, pragmatic method to enterprise AI.

Rinke put it finest close to the tip of his keynote:

“We’re not simply automating steps,” he mentioned. “We’re constructing enterprises that may adapt immediately, innovate freely, and enhance repeatedly.”

Missed it? Meet up with all of the highlights from Celosphere 2025 right here.

---

Sponsored articles are content material produced by an organization that’s both paying for the submit or has a enterprise relationship with VentureBeat, and so they’re all the time clearly marked. For extra data, contact gross sales@venturebeat.com.