Digital twins don't fail because of sensors, but because of data

A recent analysis by the Proalpha Group shows that digital twins are not made possible by sensor technology, simulation, and connectivity alone. The key is a unified database that brings together ERP, production, CAD, and PLM information in an interoperable manner.

High Expectations Clash with a Lack of Data Infrastructure

German industry attaches great strategic importance to digital twins but is making only slow progress in their implementation. In a representative Bitkom survey of 555 industrial companies with at least 100 employees, 62 percent rate digital twins as important for future competitiveness.

At the same time, 57 percent state that they do not have the necessary data. 48 percent cite inadequate IT infrastructure as an obstacle. 60 percent view their own company as a laggard, while another 14 percent believe they have already missed the boat.

The Proalpha Group’s Media Insight classifies these results as a data and integration problem. Simulation software, industrial sensors, and connectivity are generally available. However, without well-maintained master data, clear semantics, and consistent information flows, a robust digital twin cannot be created.

ERP Data Complements Sensor Data and Design Information

A digital twin requires data from multiple levels. Sensors and wireless IoT systems provide real-time measurements from machines, equipment, and production environments. CAD and PLM systems provide design parameters and technical product information.

The ERP system supplements this data with material master records, bills of materials, batches, supplier information, certificates, orders, and sustainability metrics. Only by connecting these data sources can a digital twin be created that not only visualizes but also supports operational processes.

Data exchange occurs in both directions. ERP information feeds into the digital model. Conversely, status data from the twin can trigger, for example, service processes, material orders, or the provision of replacement parts in the ERP system.

Asset Administration Shell Enables Interoperability

For digital twins to function effectively across organizational boundaries, data must be provided in standardized structures. This task is handled by the Asset Administration Shell, or AAS for short.

The Asset Administration Shell structures information into submodels, such as those for technical documentation, operational histories, CO₂ values, or maintenance information. This allows proprietary data from various manufacturers to be converted into a uniform format and exchanged via standardized interfaces.

Within decentralized data spaces such as Catena-X or Manufacturing-X, access can be controlled in a differentiated manner. Publicly accessible information can be made available, while confidential design, pricing, or supplier data remains within the company.

Digital Product Passport Becomes Part of the Same Data Architecture

With the EU Ecodesign Regulation for sustainable products, the structured provision of product data is increasingly becoming a regulatory requirement. The Digital Product Passport is to be introduced gradually for additional product groups.

According to Proalpha, companies do not need to set up a completely separate software landscape to achieve this. Information on material origin, recyclability, or product carbon footprint can be integrated into an existing AAS structure as additional submodels.

This makes a consistent database available for both digital twins and the Digital Product Passport.

From Disruption to Automated Business Process

Together with data center operator Pfalzkom, Proalpha is demonstrating how ERP information can be standardized via the AAS and made available to supply chains in an interoperable manner.

As part of the Factory-X research initiative, the company is also collaborating with SmartFactory-KL to demonstrate how detected malfunctions can be automatically converted into service tickets and spare parts orders. This transforms condition and sensor data directly into executable business processes.

For machine builders, system integrators, and IoT solution providers, this connection is crucial: The value of a sensor does not stem solely from the measurement itself, but from the context in which the measured value is placed within the respective product, plant, and process.

Structured data becomes the foundation for AI agents

Data quality becomes even more important as soon as AI agents take over operational tasks. An AI system must understand the meaning of a data field, the process in which it is used, and the applicable access rights.

Proalpha cites Gartner’s prediction that by the end of 2026, approximately 40 percent of enterprise applications could include task-specific AI agents. However, without semantically unambiguous data, governance, and reliable process information, such systems cannot operate securely and autonomously.

Data Consistency Is Key to the Digital Factory

The digital twin does not fail primarily due to a lack of technology in industrial SMEs. The key factor is whether master data, transaction data, design data, and sensor data are consistently consolidated and made available in an interoperable manner.

Companies that structure their ERP data and make it interoperable via the Asset Administration Shell not only lay the foundation for digital twins and the Digital Product Passport. They also pave the way for predictive maintenance, automated service processes, resilient supply chains, and the deployment of operational AI agents.

The full Media Insight, “No Digital Twin Without Data: Why the Digital Factory Is Decided in the ERP System,” is available from the Proalpha Group.