The process intelligence company Celonis has taken advantage of its Celosphere 2025 event to communicate the update of its platformwith improvements designed to boost enterprise AI. Celonis Process Intelligence Platform Pulls raw data from systems, applications and devices within and across businesses. It then enriches this data with customers’ business context to generate a real-time digital twin of operations through the Process Intelligence Graph.
In addition, the platform offers capabilities for analysis, design and operation of automated processes and agents. Its new features now offer a multimodal digital twin that drives AI-based operations, allowing for the generation of scalable intelligent solutions.
Now, the Process Intelligence Graph lets you gather more types of data, from more sources, faster. Integrate data lakes without duplication using Celonis Data Core. Plus, pull data and feed process intelligence into data lakes with two-way copy-free integrations. And apart from Microsoft, it is also available for Databricks.
The new features make it possible to generate a richer and more complete digital twin. The platform connects desktop actions to business processes through enhanced Task Mining and AI-powered Task Discovery capabilities. Integrates unstructured data, such as PDFs, semi-structured data, such as emails; and data from outside the organization thanks to Celonis Networks.
On the other hand, it integrates business architecture plans with operational reality in real time. This way, IT and AI departments can understand which systems are used for each activity and improve decision-making, as they can do so based on actual performance.
The platform has more than 60 predefined objects and events to enrich the digital twin, plus new AI assistants to extract and model data. Celonis also offers capabilities for the analysis, design and continuous operation of modular processes powered by AI.
The company has new features in object-centric process mining (OCPM), Performance Spectrum and Instance Explore, as well as the Object.Centric Performance application, which helps customers identify problems at the intersection points of processes.
Among the new features is also the expansion of the orchestration engine to coordinate AI agents with people and systems, as it executes end-to-end processes. This is generally available as a core capability of the Celonis platform.
The company has also launched the first Model Context Protocol (MCP) server designed for process intelligence. With it, AI agents will have the dynamic operational context they need for decision making.
Daniel Brown, Product Manager at Celonishas highlighted the need to adopt a structured approach to agentic AI, and recalled that “For AI to really work, you need to identify the right use cases, redesign business processes, and coordinate agents, along with existing people and systems.”.
