Table 1.

Overview of primary IS capabilities for Industry 4.0

Primary IS capabilities for Industry 4.0
Capability	Definition	Explanation & Example(s)
Connect & Store
Product connectivity & system interoperability	A company’s ability to establish connectivity between systems, physical objects, and humans Also, the ability of products to exchange data via defined and standardised interfaces (Givehchi et al., 2017; Javaid et al., 2021; Schuh et al., 2020).	Explanation: Connectivity is one of Industry 4.0’s most essential building blocks. To enable the efficient sharing, exchange, and management of data (e.g. product lifecycle data), a high value must be placed on interoperability and communication standards (Givehchi et al., 2017; Schuh et al., 2020). Meanwhile, transparency and traceability can be increased by integrating sensors, actuators, and computing capacities into physical objects (Bienhaus & Haddud, 2018). Example(s): Product components store location, process, and manufacturing data (Man & Strandhagen, 2017). Implementation ranges from RFID technology and smart sensors (I1, I3) to machine-to-machine communication interfaces as integral product components (I4).
Seamless connection to customers	A company’s ability to remotely monitor product usage. Also, the ability to acquire and analyse data on customer needs (Beverungen et al., 2019; Siggelkow & Terwiesch, 2019)	Explanation: With digital technologies, connections to customers can be made continuous and seamless. They can even be reconfigured as individualised interactions, allowing companies to proactively address customer needs as they arise (Siggelkow & Terwiesch, 2019). To achieve this, data gathered through connected products can be used to understand better and predict customer needs (Beverungen et al., 2019). While this capability has not yet been explored at any great depth in the literature, it was emphasised by all of our interviewed industry experts. Example(s): Expert I1 drew our attention to a car wash installation that monitors various operating criteria 24/7, i.e., the car wash’s usage, customers, car visits, and required chemicals. This enables the manufacturer to adjust their services, such as seasonal wash programmes, to the demands of individual customers.
Continuous collection & storage of real-time data	A company’s ability to collect and store real-time data from manufacturing processes and connected products (Emmanouilidis et al., 2019; Javaid et al., 2021; Lasi et al., 2014; Schuh et al., 2020)	Explanation: Most Industry 4.0 services and benefits rely on product and process data, which is why the collection and storage of structured and unstructured data is a core capability (Alcácer & Cruz-Machado, 2019; Schuh et al., 2020). Due to new possibilities in exchanging real-time data (e.g. cloud solutions), the amount of stored data grows exponentially as single entities and data from entire manufacturing processes, plants, and connected products can be stored (Emmanouilidis et al., 2019; Lasi et al., 2014). Example(s) For example, the storage of traceable event timelines enables the verification of data evaluation and recovery of technical issues and evidence-based decision-making (I2).
Understand & Act
Automated processing of (un-)structured data	A company’s ability to derive meaningful information from (un-)structured data from different sources (Alcácer & Cruz-Machado, 2019; Prifti et al., 2017; Schuh et al., 2020).	Explanation: The transformation of data into information, which can require the processing of large amounts of data, includes the categorisation, characterisation, consolidation, and classification of data. On the one hand, structured data (e.g. ERP data) allows for quick aggregation of data from different machines, products, and systems into a unified format that facilitates more straightforward analysis. Such automation is necessary to efficiently process the ever-increasing amount of input without human intervention (Prifti et al., 2017). On the other hand, mainly processing unstructured data is an even more significant challenge with enormous potential (Alcácer & Cruz-Machado, 2019). Our industry experts especially emphasised the opportunity in analysing unstructured data. Example(s): To automate processes, some of our interviewed experts suggested the use of data from (technical) service reports that contain both structured and unstructured data (I1, I4).
Descriptive & diagnostic analysis	A company’s ability to analyse product and process data in such a way as to easily detect (unwanted) deviations and identify their root causes (Dai et al., 2020; Duan & Xu, 2021; Porter & Heppelmann, 2014; Shi-Nash & Hardoon, 2017).	Explanation: Using statistical methods and descriptive analysis of historical data (e.g. product lifecycle data, manufacturing processes), manufacturers can better detect unwanted deviations from a predefined standard (Dai et al., 2020; Shi-Nash & Hardoon, 2017). This involves the description of KPIs (e.g. for a manufacturing process), visualisation of machine states, and detecting machinery failures as well as any anomalies or patterns in the data. Once a malfunction or failure is detected, the diagnostic analysis attempts to identify the root cause (Dai et al., 2020). The diagnostic analysis is a reactive response (Porter & Heppelmann, 2014), much like descriptive analysis and greatly depends on human-expert reasoning. Example(s): A manufacturer of transmission technology (I2) performs diagnostic services for customers by detecting anomalies in the vibration patterns of their transmission systems (gained from sensor data). However, an engineer is always required to analyse and further evaluate the data in order to identify the root cause.
Automated monitoring & reasoning:	A company’s ability to automatically monitor manufacturing processes and products to optimise their reasoning and act in a way that positively impacts production as well as decision-making processes (Babiceanu & Seker, 2016; Wagire et al., 2020; Waschull et al., 2020)	Explanation: Advanced and embedded analytics tools enable automatic monitoring of products and manufacturing processes, which in turn enables real-time decision-making in dealing with unwanted deviations (Waschull et al., 2020). Such systems aim to minimise human involvement by combining detection and reasoning. If a problem occurs for a specific entity, the monitoring system detects this critical event automatically. Based on predefined constraints and the identified reason, the system then triggers predefined processes (Babiceanu & Seker, 2016). Example(s): A car wash manufacturer uses the continuous condition monitoring of a car wash system in conjunction with data analytics to detect customer-specific sources of component failures (I1). This lays the foundation for novel services, such as the automated ordering of spare parts. This provides an ‘added value’ (I8) for the customer and higher profit margins for manufacturers.
Predict & Self-optimise
Predictive & prescriptive analysis	A company’s ability to automatically predict events and time-series based on data-driven methods, then suggest optional decisions based on the results (Baptista et al., 2018; Duan & Xu, 2021; Frank et al., 2019; Shi-Nash & Hardoon, 2017)	Explanation: Predictive and prescriptive analysis seeks to identify future patterns and anomalies, based on CPS approaches and advanced data analytics. The objective is to improve efficiency and address customer demands by improving reactions to unpredictable events (Baptista et al., 2018; Shi-Nash & Hardoon, 2017). By combining domain knowledge with data-based approaches (e.g. machine learning models), the aim is to foresee trends, behavioural patterns, and correlations to predict critical events (Duan & Xu, 2021; Frank et al., 2019). Important decisions and actions are then taken (i.e. prescriptive knowledge) to avoid downtimes or minimise the personal costs of the respective services (e.g. number of contacts needed). Example(s): Recently, predictive maintenance services for production robots grew in importance. Especially during the COVID-19-pandemic, access to technical service experts was limited or prohibited at many organisations (I8).
Self-optimising products & production systems	A company’s ability to develop self-optimising products and cyber-physical infrastructures, which can make and deploy automated decisions (Kang et al., 2016; Shi-Nash & Hardoon, 2017; Vaidya et al., 2018; Waschull et al., 2020)	Explanation: This capability is about converting traditional machines to self-aware and self-learning machines to improve their overall performance and maintenance management (Vaidya et al., 2018). This builds on most of the technical capabilities described above. For automated production infrastructures, production involves several centralised or decentralised workstations that independently make decisions, dynamically allocate tasks, and seamlessly negotiate appropriate reactions to overcome problems. This also includes a degree of freedom in decision-making and the ability to learn from events or previously made decisions (Waschull et al., 2020). Most of the interviewed industry experts confirmed that they have not yet developed corresponding capabilities. Example(s): The use of autonomous and collaborative robots to automate handling, welding, and painting activities has already begun on several production sites (Wagire et al., 2020), as well as the implementation of self-managing smart factories (e.g. Bellini et al., 2021).
Autonomous provision of products & services	A company’s ability to automate the provision of products and services, including pricing mechanisms, proactive product-triggered communication, and real-time adaptation to exogenous events (Oztemel & Gursev, 2020; Shihundla et al., 2019)	Explanation: From a technical point of view, this capability is more complex than those discussed above, because it is derived from (almost) all of those capabilities. Interviewees often related it to concepts such as ‘smart factory’ or ‘dark factory’ because it establishes automatic solutions that will execute versatile operations. They will do so independently from location and provide the ability to react context-specifically to fast-changing customer needs (Lasi et al., 2014; Oztemel & Gursev, 2020; Shihundla et al., 2019). Example(s): When it comes to the autonomous provision of products and services, smart factories are known to implement this capability already. Manufacturing environments are equipped with automated systems to the extent that they do not require the presence of humans for anything more complex than simple tasks like removing parts (Oztemel & Gursev, 2020).