
Unplanned downtime costs the 500 largest manufacturers in the world approximately $1.4 trillion annually, which is equivalent to about 11 percent of their revenue, according to the 2025 Global Industry Report (OxMaint). One of the most promising ways to reduce this problem is predictive maintenance through the use of AI.
Manufacturing is one of the industries that has the most data today – from machines, ERP and MES systems, sensors, quality processes, warehouses, suppliers, or technical documentation. At the same time, the sheer volume of data does not mean that companies know how to use it effectively. In many enterprises, the same information is often scattered across different systems, spreadsheets, and documents, which hinders the implementation of new technologies. This is also confirmed by Gartner data: in a study published in 2026, 38 percent of infrastructure and operations leaders who experienced problems with AI projects cited poor data quality or limited data availability as the direct cause of project failure[1]. As a result, the industry’s problem today is not a lack of data, but a lack of trust in data – and this can limit both AI implementations and the preparation of companies for new market requirements.
Ambiscale, a company implementing Data Governance for the industrial sector, explains, however, that organizing data in manufacturing does not have to start with a major transformation. Companies can start with product, material, supplier, batch, and index data, and by identifying data owners. Such a first step allows for building a reliable foundation for AI, new regulatory requirements, and further business scaling.
AI in production starts with data
Unplanned downtime costs the 500 largest manufacturers in the world approximately $1.4 trillion annually, which is equivalent to about 11 percent of their revenue, according to the 2025 Global Industry Report (OxMaint)[2]. One of the most promising ways to reduce this problem is predictive maintenance through the use of AI. Models analyze data from sensors and the operating history of equipment to detect symptoms of failure and allow for a response before the production line stops. This is particularly important in a situation where the cost of downtime can reach millions of dollars. Predictive technologies can reduce the scale of this problem, but their effectiveness is directly dependent on data quality. If data from machines is incomplete, inconsistent between plants, or cannot be linked to specific equipment, production batches, or service events, even the most advanced AI model will operate on a weak foundation.
This means a change in the approach to the digitalization of production. Implementing artificial intelligence does not start with choosing a model or a tool, but with answering the question of whether the company knows what data it has, where it comes from, and whether it can be trusted.
The industry is losing huge amounts of money due to unplanned downtime. Today, the problem is no longer a lack of technology, but the quality of data and how we manage it, although simply improving data quality is no longer enough. If a company does not establish data owners, identification standards, rules for maintaining reference data, and accountability for errors, problems will quickly return. This is how we move from Data Quality to Data Governance. It is a way to build trust in data, which becomes part of the production process, the product, customer relationships, and regulatory compliance – says Krzysztof Surowiecki, Senior Data & Analytics Advisor at Ambiscale.
Digital product passport enters the industry
From February 18, 2027, a digital battery passport will be mandatory for automotive and industrial batteries above 2 kWh placed on the EU market[3]. For manufacturers, importers, and suppliers, this means the need to prepare, among other things, product identifiers, supply chain information, environmental data, and secure mechanisms for their exchange – before the requirements begin to apply to subsequent product categories[4].
Companies will therefore need to have structured data on, for example, materials or the origin of components. This is one of the first clear signals that consistent product data is ceasing to be solely a matter of efficiency and is becoming a factor determining market viability. In the future, they may also form the foundation for digital twins – digital replicas that allow for simulating product behavior and reducing testing costs.
A similar direction is already being dictated by the automotive sector. Catena-X[5], an open data exchange ecosystem, brings together companies such as BMW, Volkswagen, Ford, Renault, Volvo, Bosch, Continental, Siemens, and ZF. Since April 2025, registration in Catena-X has been an integral part of BMW’s procurement process, and suppliers must be ready to exchange data according to this standard.
Pressure is also being increased by AI regulations. The AI Act provides for maximum fines of up to 35 million euros or 7 percent of global annual turnover, and for high-risk systems, requirements explicitly include, among other things, appropriate data management and quality[6].
In practice, this means that a manufacturer will increasingly have to not only know what they have produced, but also be able to prove what the product was made of, where the materials came from, and what data this information is based on. For companies whose data is currently scattered between ERP, Excel, and production systems, Data Governance is therefore becoming the foundation for both regulatory compliance and cooperation with large partners – emphasizes Jerzy Kozyra, CEO of Ambiscale.
You don’t have to start with a revolution
Implementing Data Governance does not have to mean a multi-year program covering all data and all enterprise processes at once. Especially in medium-sized manufacturing companies, the first step can be to organize areas that have a direct impact on operational activities.
In the first instance, these can include: product data and material indices, machine and equipment data, product configurations and spare parts, technical documentation, service data, information about suppliers and materials, production batch data, or defining owners of specific data and responsibility for its quality. Such an approach allows for building Data Governance around specific business problems instead of treating it as an abstract technological project.
Not every company has to immediately build an extensive Data Governance program covering the entire organization. You can start with one specific business need, such as mapping and preparing data for a digital product passport. The key is to create rules so that it is known who is responsible for the data, where its correct version is located, and what standards it must meet – notes Krzysztof Surowiecki.
Systematized data then facilitates the scaling of processes between plants, system integration, reporting, cooperation with large clients, and the implementation of subsequent AI projects. In the case of manufacturing companies, this can also influence how the enterprise is perceived by potential investors or buyers. As a result, Data Governance ceases to be just a way to “clean up” data and becomes an element of preparing the company for further growth and building a competitive advantage.
_____________________________________________
Ambiscale – an IT company belonging to the HEXE Capital group, implementing Data Governance solutions, supporting manufacturing and industrial enterprises in transitioning from scattered data located in ERP, spreadsheets, and production systems to an organized data management model. The goal is to create a foundation for scaling operations, implementing AI, reporting, meeting regulatory requirements, and cooperating with clients and partners in increasingly digital supply chains. Multiple award-winner in the IT industry, including Top Web Developers Clutch and Digital Impact Awards.
[1] https://www.gartner.com/en/newsroom/press-releases/2026-04-07-gartner-says-artificial-intelligence-projects-in-infrastructure-and-operations-stall-ahead-of-meaningful-roi-returns
[2] https://oxmaint.com/industries/manufacturing-plant/manufacturing-maintenance-global-industry-report-2025
[3] https://single-market-economy.ec.europa.eu/single-market/digital-product-passport/batteries_en
[4] https://digiprodpass.com/blogs/digital-product-passport-guide








