Artificial intelligence is growing at accelerated rates in Italian companies, but the race to adopt technology has left behind a fundamental prerequisite: data governance. This article explores what industrial data governance really means and why its absence is today the primary cause of failure of AI initiatives in production.
We start from the Italian context, where the duality between rapid adoption and delayed governance is stark, to then analyze the European regulatory framework (Data Governance Act, Data Act and AI Act) that is transforming governance from best practice to progressively binding requirement. We examine the investment directions of the most advanced companies, the critical points still open, and a four-step operational path, inspired by Lean Thinking, to initiate governance in an incremental and measurable way. In conclusion, a reflection on how well-governed data is the strategic enabler of servitization and competitive advantage in the industry of the future.
In 2025, the artificial intelligence market in Italy reached 1.8 billion euros, with a growth of +50% in just one year. During the same period, 71% of the large companies had already launched at least one active AI project. Yet only 9% of those same companies can declare that they have structured AI governance processes — and only 24% are satisfied with the quality of their data.
It’s not a paradox: it’s the exact photograph of what happens when you run faster than your control capabilities allow. And in a manufacturing context — where data is generated by machines, sensors, ERP, MES, and SCADA systems with different standards and heterogeneous origins — this gap between adoption and governance becomes a concrete operational risk, before it even becomes a compliance issue.
In the book “25 Years of Lean Thinking in the Italian Way“, Michele Bonfiglioli writes a sentence that serves as a manifesto: “All these tools are worthless without a reliable database and centralized governance. Without consistent data, any system becomes a mere simulation.” This article develops that insight: what is industrial data governance, where do the most advanced companies invest, what are the organizational bottlenecks, and what European regulatory framework is reshaping the rules of the game for the manufacturing industry.
Industrial data governance is an organizational discipline, not a technology. It does not coincide with the purchase of a data catalog, the installation of a data lake, or the deployment of a master data management platform. It is a set of policies, roles, processes, and standards that define who owns the data, who ensures its quality, how it is classified, who can access it, and most importantly, who is responsible when something goes wrong.
In production, this leads to concrete questions: do the efficiency data that come out of MES and those that enter the ERP tell the same story? If an operator in the plant and an analyst in the office are looking at the same line indicator, are they seeing the same number? When a supplier sends the quality data for a batch, are they integrated coherently with the data from the assembly process, or do they end up in a parallel Excel sheet?
If the answer is “it depends,” “not always” or “I don’t know,” industrial data governance is an open problem. And as the experience of Bonfiglioli Consulting in digital transformation projects confirms, this problem is among the top causes of failure for AI and analytics initiatives: not because the algorithms are wrong, but because the data they work with is not governed.
One of the most widespread digital waste in manufacturing companies is data redundancy: the same information replicated in multiple systems with different definitions, asynchronous updates, and no single authoritative source. Michele Bonfiglioli explicitly identifies it in his book as one of the digital waste—the new waste in the age of connected industry: automating useless activity means digitizing waste; replicating inconsistent data across multiple platforms means amplifying the error.
The breaking point is organizational before it is technical. It requires the definition of three roles that are still lacking in most companies:
the data owner : the person responsible for the data at the domain or process level
The data steward: the operational manager of daily quality
The data product owner: responsible for the use of that data as a business asset.
According to research by the MIT Sloan Management Review, organizations that assign formal ownership of data achieve significantly better results in analytics and AI projects than those that treat governance as a secondary IT activity, and identify clarity of roles as the primary enabling factor, before even technological tools.
This alignment between organizational responsibility and data quality is exactly what the Bonfiglioli Consulting Lean World Class® model applies to the digital perimeter: first the operational responsibilities are structured, then they are digitalized, never the reverse.
Italy presents a significant dichotomy. On the technological adoption front, ISTAT data 2025 shows that 38.1% of companies with at least 10 employees have reached a high or very high level of digitalization, with the manufacturing sector among the most active in investing in IoT, automation, and robotics. On the governance front, the picture is more fragile.
The Artificial Intelligence Observatory of the Politecnico di Milano notes that, with an AI market growing by 50% in 2025, only 9% of the major Italian companies have structured AI governance processes. 54% are in the process of being structured, still without defined deadlines. 19% of the workers declare that they exclusively use AI tools approved by the company — which means that the vast majority use uncontrolled tools, with company data circulating outside any governance framework. In the manufacturing sector, where data concerns processes, design, and the supply chain, this dispersion is a real risk.
To this is added the data from the Data Management Observatory of the Politecnico di Milano: only 24% of the Italian companies declare themselves satisfied with the quality of their data, with many of them falling below the average in terms of governance and the ability to calculate the business value of their information assets. Investing in sensor technology, MES and IoT platforms without building a data quality governance system means, to use the metaphor of the book ‘25 Years of Lean Thinking in the Italian Way", building a sophisticated navigation system on a map full of errors.
Europe has built in a few years a regulatory framework that transforms data governance from best practice to progressively binding requirement. For manufacturing companies, three measures define the scope of action.
The Data Governance Act (DGA), in force since September 2023, establishes the conditions for the secure and trusted sharing of data between businesses, sectors, and EU countries. It introduces the role of certified data intermediaries and creates the framework within which sectoral Data Spaces are built — controlled sharing spaces that allow industrial sectors to collaborate on data without losing informational sovereignty.
The Data Act, in effect from September 2025, introduces a fundamental principle for connected factories: anyone who generates data through IoT devices and connected machinery has the right to access that data and share it with third parties on fair terms. This redefines the relationship between machine manufacturers and users: a manufacturing company that uses connected equipment can claim access to operational data, even when the machine manufacturer prefers to retain it for commercial reasons. This is a turning point that, when properly exploited, opens up new opportunities for optimization and service provision.
The AI Act, the world’s first law on artificial intelligence, provides in Article 10 specific data governance requirements for high-risk AI systems: the datasets must be documented, verified, representative, and free of critical biases, with management practices that ensure integrity and relevance. Full applicability for high-risk systems has been enacted in August 2026. For companies using AI in critical manufacturing contexts — predictive maintenance, automated quality control, production planning — compliance is no longer delayed.
The investment directions in industrial data governance in the 2025-2026 biennium focus on three main axes.
Governance embedded in operational workflows. The traditional model — governance as a separate layer, often managed by IT as a post-audit activity — is giving way to an approach where quality, classification, and data access rules are embedded directly into the workflows. The principle is that governance that is not continuous is not governance: quality controls must occur at the moment the data is generated or transformed, not downstream. The global data governance market, worth $4.60 billion in 2026, is growing at a CAGR of 16% through 2031, with Europe recording the highest rate — estimated at 20% annually — and Germany as the dominant market in the continental context. These numbers not only reflect the growth of software; they also signal a growing awareness that industrial data is a strategic asset that requires the same level of protection as physical facilities.
Data Space for Manufacturing. The European Commission’s Digital Europe program has funded the construction of industrial data spaces dedicated to manufacturing, with funding of up to 3 million euros per project, designed to allow companies to share data along the supply chain — between OEMs, suppliers, service providers, and customers — while maintaining control over the sovereignty of their information. The Gaia-X framework provides the technical architecture to build these spaces in a federated and interoperable way, based on principles of self-description, sovereignty, and verifiable trust. For Italian manufacturing SMEs, participation in these data sharing ecosystems is not just a technological opportunity: it is a competitive positioning in the European supply chain.
AI Data Governance. With the increase in AI projects in production, specific data governance for the training and validation of models becomes an autonomous priority. This includes managing the provenance of training data, documenting biases, tracking the versions of datasets, and ensuring certified quality of input. In a manufacturing company that uses predictive models for quality or maintenance, an ungoverned training data can lead to incorrect automated decisions — and difficult responsibilities to assign ex post.
System silos. ERP, MES, PLM, CRM and SCADA speak different languages, with inconsistent definitions of the same entities: order, product, machine downtime, defect, batch. The result is digital islands — that is, systems that manage data linked to the same process but without a shared model — which force manual reconciliations, with time loss and increased errors.
Lack of clear ownership. When no one is formally responsible for a given data, that data gradually degrades. Blank fields, outdated values, and unstandardized coding accumulate without anyone having the authority or incentive to correct them. Data governance requires a formal assignment of responsibility, with clear roles and measurable KPIs.
Reactive governance instead of preventive. Most companies discover data quality problems when launching an AI or advanced analytics project: that is, when it is too late to perform remediation in a timely manner without compromising the project. Integrating quality controls into daily operational processes is still rare but crucial for the sustainability of any advanced digital initiative.
Shadow AI as a new governance risk. A critical front that is becoming increasingly relevant is Shadow AI: the use of artificial intelligence tools not approved by the company, often through personal accounts or external services, which introduces non-traceable data flows and renders traditional governance ineffective. This is not just a matter of cybersecurity: when technical documents, process data, business information, or industrial know-how are inserted into off-the-shelf tools, the company loses control over where those data end up, who processes them, and with what guarantees of confidentiality. In this sense, Shadow AI is a form of silent erosion of the single source of truth and forces industrial data governance to oversee not only systems and platforms but also daily usage behaviors.
Italian regulatory news. The recent implementing decrees in the field of AI confirm that the government of artificial intelligence is now a matter not only technological but also organizational and of responsibility. The message coming from the Italian framework is clear: AI must be governed with an integrated system of compliance, cybersecurity, data protection, supplier control, training, documentation and traceability. For manufacturing companies, this means that data governance is no longer an accessory support to digital projects, but a condition of legitimacy and operational continuity.
Difficulties in demonstrating value. Data governance is an investment with non-immediate returns, which makes it vulnerable to budget cuts. Building a narrative of value — how much a wrong data costs in a productive decision, how much a reliable data is worth in a servitude contract — is a managerial skill that is still underdeveloped.
Compliance and governance as separate silos. In many organizations, data management is entrusted to legal/compliance for GDPR reasons, separate from the operational governance of industrial data. With the AI Act linking directly the quality of data and the legal responsibility of the AI system that uses it, the two dimensions can no longer coexist in separate compartments: they must converge into a single integrated oversight.
Lean Thinking — which Bonfiglioli Consulting has been applying in manufacturing operations for over 25 years — also offers a fundamental operating principle for data governance: first eliminate waste, then standardize, then continuously improve. Applied to data, this means starting with an 18-month enterprise program, not an incremental approach focused on measurable value from the very first cycle.
The process is divided into four sequential steps. The first is the mapping of critical domains: identifying which data are essential for the most relevant decisions — product quality, line efficiency, supplier reliability, planning accuracy. The second step is defining minimum quality rules for those domains: completeness, accuracy, timeliness, consistency, calibrated to the level sufficient for the decisions those data must support, not to an ideal of statistical perfection.
The third step is the formal assignment of ownership: for each critical domain, who is the data owner? Who is the data steward? The answer must be nominative, with measurable quality KPIs and visibility in reporting. The fourth step is the construction of a minimum data catalog: an inventory of data assets with metadata, source, frequency of update, owner, and certified quality level. Discipline is more important than the tool: a well-governed shared registry produces more value than a sophisticated platform abandoned.
Operational metrics to monitor maturity over time include: the data quality score for each domain (percentage of records that comply with the defined rules), the average time to resolve an anomaly in data from detection to correction at the source level, and the percentage of critical decisions supported by certified data. These indicators transform governance from an abstract activity into a measurable process, with visibility in operational reporting.
Industrial data governance is not only a factor of internal efficiency. It is also the strategic enabler of the business model that the book "25 Years of Lean Thinking in the Italian Way" identifies as the competitive frontier of Industry 5.0: the serviceification. Companies that no longer sell only products, but "performance as a service" — guaranteed availability of a plant, certified quality of an output, energy efficiency as an SLA contract — need governed data not only within the factory, but along the entire value chain, including the end customer.
This broadens the scope of governance: from internal data to data flowing between enterprises and customers through connected machines, predictive platforms, and service level agreements. The value of data increases when it is shared securely and governed — and that is the promise of European industrial Data Spaces, where trust in the quality and provenance of the shared data is the prerequisite for building competitive collaboration ecosystems.
Walter Caiumi, an entrepreneur at Voilàp Group, summarizes this direction with precision: "We have learned to look beyond our direct customer, to understand the behaviors and needs of the end user." Understanding the end customer means having reliable data about what happens after the sale. Data that, without governance, never arrives — or arrives too late and too dirty to be useful.
Companies that build organizational data governance before it becomes a regulatory requirement are ahead in several dimensions: faster and more reliable decision-making processes, AI that is more robust and compliant with the AI Act, stronger digital partnerships along the supply chain thanks to the Data Act, and the ability to offer business models based on the certified quality of information.
At Bonfiglioli Consulting, we help manufacturing companies build this organizational foundation — processes, data, people — as a concrete prerequisite for any digital and AI initiative. Because as the Lean Thinking principle teaches: first you eliminate waste, then you build value. In the digital world, the first waste to eliminate is the data that no one governs.
Do you want to measure the level of maturity of data governance in your company? Contact the Bonfiglioli Consulting team for a personalized assessment.
It is an organizational discipline — not a technology — that defines who owns the data, who ensures its quality, how it is classified, and who can access it. It does not coincide with the purchase of software or the installation of a data lake: it is a set of policies, roles, and processes that make the data reliable and governed throughout the entire operational chain.
Because an algorithm, even the most sophisticated one, produces incorrect results if the data it works with is not governed. The first cause of failure of AI projects in production is not technological tools, but the quality and consistency of the input data. Without governance, AI remains a promise.
Three figures are fundamental and still missing in most companies: the data owner, responsible for the data at the domain or process level; the data steward, responsible for the operational quality of daily work; and the data product owner, responsible for the use of data as a business asset.
Europe has built a regulatory framework that transforms governance from best practice to mandatory requirement. The Data Governance Act regulates the secure sharing of data between businesses and sectors. The Data Act guarantees manufacturing companies the right to access data generated by their connected machinery. The AI Act imposes specific data governance requirements for high-risk AI systems, with full applicability from August 2026.
It is the use of artificial intelligence tools not approved by the company, often through personal accounts or external services, that introduces untraceable data flows beyond any control perimeter. When process data, technical documents, or industrial know-how are inserted into tools outside the perimeter, the company loses control over where that data ends up and with what guarantees of confidentiality.
With an incremental approach in four steps: mapping the critical data domains for the most relevant decisions, defining minimum quality rules for those domains, formally assigning ownership with measurable KPIs, and building a minimum data catalog with certified metadata, source, and quality level. Discipline is worth more than the tool: a well-governed shared registry is better than a sophisticated platform abandoned.
Did it help you? Add it as your preferred source on Google
ai translated