Artificial intelligence is fundamentally changing the role of Manufacturing Execution Systems. It not only identifies patterns and deviations in production data but also supports users in their daily work with the system – from data analysis to system operation and master data management.
In the following sections, we present different applications of artificial intelligence in Cosmino Panteo – ranging from functions that are already available to topics we are currently working on.
A key advantage of AI in data analysis is its ability to identify correlations and changes that are difficult or impossible to detect using conventional analyses or predefined rules. In our projects, we combine different types of process, machine, and production data from Cosmino Panteo and other data sources and make them available to AI for analysis. The key difference compared with conventional limit monitoring is that you do not need to know in advance what you are looking for. AI learns typical patterns and correlations from production data and independently detects deviations from them.
The benefits become particularly apparent when individual values appear normal when considered separately, but their combination or development over time indicates unusual process behavior.
In this way, we can reveal anomalies that would remain undetected using conventional analyses or predefined limits. Such anomalies can provide early indications of quality issues, process instability, increasing wear or potential equipment failures.
The specific benefits:
Fig.: A highly simplified example of anomaly detection across features and processes: An anomaly can also occur in a specific combination of various individual process data points or in their temporal evolution.
Monitoring process parameters and intervention limits and initiating appropriate measures are already important tasks of an MES. When deviations occur, Cosmino Panteo can, for example, automatically trigger additional inspections, place parts on hold, notify employees, initiate process adjustments or stop machines.
When artificial intelligence detects anomalies or unusual correlations in real-time data from Cosmino Panteo, feedback to the MES ensures that these insights can immediately trigger a response on the shop floor. Today, the rules for such responses are predefined – similar to conventional limit monitoring.
In the future, AI can also support the selection of appropriate measures. If the measures taken and their effectiveness are documented, this information can be used to provide increasingly targeted recommendations for suitable responses.
The specific benefits:
Our vision is a production environment in which impending downtime, quality deviations and other forms of waste in machines and related processes can be detected early and, in some cases, prevented automatically before they have any impact.
Another important area is reducing the effort required to operate and configure complex systems. MES applications often require extensive knowledge of master data, configurations and reporting. Many of these tasks are time-consuming and tie up employees whose expertise could be put to better use elsewhere.
We have already developed an AI assistant for Cosmino Panteo, which we are currently training and preparing for various use cases:
We are pursuing a similar approach for product knowledge and documentation, making information easier to access in multiple languages through an AI assistant without requiring users to search through manuals themselves.
The specific benefits:
We have already created the technical foundation in Cosmino Panteo for integrating different AI solutions, including Large Language Models. Companies are not tied to specific AI models and can also use their own models.
Cloud-based AI is not required – AI can be operated entirely on the customer’s own infrastructure.
This allows companies to pursue their own AI strategy while using the capabilities of Cosmino Panteo to integrate artificial intelligence into their production processes.
Artificial intelligence already supports our developers in the further development of MES functions and helps reduce development and testing effort.
One of our next goals is an AI assistant that allows customers to describe their use cases in natural language. A Large Language Model interprets and structures the requirements. Based on Cosmino Panteo’s modular, widget-based architecture and rule-based workflow control, the AI assistant then creates an initial functional prototype. This makes it possible to assess at an early stage whether and how an implementation concept will work in practice.
Artificial intelligence is therefore more than just a new function in Cosmino Panteo for us: it is becoming an integral part of our products and processes – with the goal of continuously advancing the possibilities of production digitalization together with our customers.