The Achilles heel of production digitalization: Why two decades of Industry 4.0 have failed in the face of reality

When PowerPoint visions meet the gym floor – A reckoning

Two decades have passed since the dawn of the so-called fourth industrial revolution, and the sobering assessment is disheartening. Nearly eighty percent of all digitalization initiatives in production fail—a success rate bordering on self-deception. While consultants and software companies promise a breakthrough to the digital enterprise, plant managers and production supervisors grapple with an uncomfortable truth: the digitalization of manufacturing, in its current form, is fundamentally flawed. Not because the technology is lacking, but because the implementation logic follows two fundamentally different paradigms, each doomed to failure.

The top-down approach, in which management selects a software solution after extensive presentations and tenders, regularly ends in the same debacle. What appears on glossy presentation slides as the perfect integration of all requirements turns out, in practice, to be a years-long adaptation project. Manufacturing Execution Systems with an average implementation time of fifteen to sixteen months are still the rule, not the exception. The systems are rigid, expensive to adapt, and require production to adapt to the software, not the other way around. Processes that have proven optimal over decades are forced to fit into pre-made templates. The result: implementations that never deliver the promised efficiency gains because they were planned without regard for operational reality.

The bottom-up approach fails for diametrically opposed reasons. Excel macros, Access databases, and custom-programmed tools emerge out of necessity when IT departments are overloaded and standard software doesn't meet specific requirements. Initially conceived as stopgap solutions, these isolated systems quickly become business-critical. Their developers, often skilled employees without formal programming training, create pragmatic tools that actually work. But with each additional feature, the technical debt grows exponentially. Faulty documentation, lack of version control, no audit trails, and insufficient scalability are just the most obvious problems. When the developer leaves the company, a black box remains that no one can maintain, but everyone is forced to continue using. The backlog grows while more and more resources are diverted to maintaining outdated solutions instead of tackling new challenges.

Both approaches fail not for technical reasons, but for structural ones. Top-down digitalization ignores the operational intelligence of those who actually produce. Bottom-up initiatives fail due to a lack of governance and technical expertise. The promise of Industry 4.0 – intelligent, networked, and flexible production – remains unattainable in this stalemate. Three out of four German companies lack a well-developed digitalization strategy, and eighty percent operate with largely manual or only partially automated processes. Data repositories are filling up, but insights remain elusive because the data is trapped in silos.

The hidden shadow IT: When Excel becomes business-critical infrastructure

In the production halls of German medium-sized companies and even large corporations, a parallel world of digital solutions exists that doesn't appear in any IT inventory. Excel spreadsheets with macros handle production planning. Access databases manage quality data. Custom-written Python scripts analyze machine data. This shadow IT has become the backbone of many production processes because official systems are too slow, too inflexible, or simply nonexistent.

The origin story is almost always the same: A problem arises, the IT department is overloaded, or the existing ERP system lacks the necessary functionality. A technically skilled employee creates a pragmatic solution using the available tools. The solution works, spreads, and is expanded. Within a short time, the tool becomes a business-critical application used daily by dozens of employees. This evolution occurs outside of any IT governance, without security audits, backup strategies, or professional maintenance.

The risks are considerable. Data changes are untraceable, there is no logging, and auditability is nonexistent. Authorization concepts are lacking, making fundamental control principles like the four-eyes principle impossible. Access across distributed locations and with multiple users is problematic, especially at a time when cloud-based, real-time access should be standard. Data security—whether integrity, consistency, or confidentiality—is not guaranteed. Release stability is nonexistent, meaning that an operating system update or a new Office version can cripple the entire solution. Documentation is poor or completely missing, and the knowledge is lost when the developer leaves the company.

Nevertheless, these solutions survive year after year because they have a crucial advantage: they solve real problems and were developed by people who understand the production process. A planning spreadsheet that a shift supervisor has refined over years often reflects the reality of manufacturing better than a standardized MES module costing millions of euros. This implicit recognition of their functionality is what makes replacing them so difficult. Everyone knows they are problematic, but no one dares to shut them down because production would grind to a halt without them.

The real tragedy lies not in the existence of these solutions, but in the fact that they are symptomatic of a fundamental failure. They prove that local, needs-based digitization works when developed by the right people with the right tools. At the same time, they demonstrate the IT industry's inability to provide flexible, adaptable tools that are both professionally maintainable and quickly adjustable to specific requirements. This gap between supply and demand is the real Achilles' heel of production digitization.

The new wave: When artificial intelligence democratizes software development

While traditional approaches to digitalization are languishing, a fundamental shift is underway. AI-powered low-code and no-code platforms promise nothing less than the democratization of software development. Tools like Lovable, Microsoft Power Platform, and Mendix enable employees without formal programming skills to create functional applications. The figures are impressive: Gartner predicts that by 2026, approximately 75 percent of all new enterprise applications will be built using low-code technologies, a dramatic increase from just 25 percent in 2020. Eighty percent of low-code users by 2026 will come from business departments outside of IT.

The technological foundation of this revolution lies in the fusion of low-code platforms with generative artificial intelligence. Instead of laboriously assembling components via drag-and-drop, users can describe their requirements in natural language, and the AI ​​generates executable code. Lovable, a platform that rapidly gained momentum after a $15 million funding round, enables the generation of complete web applications from text descriptions, including frontend, backend, and database logic. All code is synchronized to GitHub, allowing developers to take over and further develop the generated code as needed. Development time is reduced from months to days, and costs can decrease by up to 60 percent.

For the manufacturing sector, the timing of this development is hardly coincidental. The shortage of skilled workers is worsening dramatically, while the pressure to digitize is increasing. Six out of ten industrial companies in the DACH region complain about a lack of data analysts, and more than half of the companies fail to put the insights gained into practice. Waiting lists at IT departments are growing longer, while the reality of production tolerates no delays. Low-code offers a solution: Production managers, shift supervisors, and process engineers could develop the tools they actually need without having to wait for overburdened IT departments.

More than 800 employees of Munich's municipal utilities are now citizen developers, using low-code tools to develop their own applications. Porsche is rolling out a company-wide low-code platform that enables departments to independently digitize their processes. These success stories point to a fundamental shift: Digitization is moving to where the problems arise, instead of being mandated by central IT departments.

The vision of the autonomous company: When software disappears

The most radical implication of this development was formulated by none other than Satya Nadella, CEO of Microsoft, in a remarkable statement: Business apps as we know them will disappear. His argument is compellingly logical: Traditional SaaS applications are, at their core, CRUD databases with business logic layered on top. This business logic, Nadella argues, will increasingly be taken over by AI agents that are not bound to specific backends. Instead of each application implementing its own logic, autonomous AI agents will manage this logic in an overarching AI layer, accessing multiple databases and systems.

More about it here:

• YouTube: Satya Nadella On Evolution Of SaaS

This vision is not some distant dream. Gartner predicts that by 2028, one-third of all enterprise applications will have integrated agentic AI capabilities. IDC anticipates more than 1.3 billion deployed AI agents by 2028. McKinsey reports that 78 percent of companies already use generative AI in at least one business function, and 88 percent plan to increase their budgets for AI agents.

For Manufacturing Execution Systems (MES) and shop floor applications, this could mean the end of the current architecture. Instead of monolithic MES installations that require fifteen months of implementation and are then rigid, AI agents could orchestrate production processes, analyze quality data, predict maintenance needs, and optimize production plans—all configurable through natural language interaction. The line between user and developer blurs when a shift supervisor can simply describe to their AI agent what analysis they need, and the software then generates and provides it.

Excel, as an example of this transformation, illustrates its scope. With the integration of Python, Excel transforms from a spreadsheet program into a virtual analyst that generates scenarios, suggests solutions, and executes plans. This redefinition demonstrates how traditional tools, through AI integration, become autonomous assistants that not only execute commands but also solve problems independently.

Industry focus: B2B, digitalization (from AI to XR), mechanical engineering, logistics, renewable energies and industry

More about it here:

• Xpert Business Hub

A topic hub with insights and expertise:

• Knowledge platform on the global and regional economy, innovation and industry-specific trends
• Collection of analyses, impulses and background information from our focus areas
• A place for expertise and information on current developments in business and technology
• Topic hub for companies that want to learn about markets, digitalization and industry innovations

The coming paradigm shift: Local intelligence instead of central control

The convergence of AI-powered development tools and the need for flexible shop floor solutions points to a fundamental paradigm shift. The next generation of production systems may not be developed by IT departments or software companies, but directly on the production floor by those who understand the processes best. This change would resolve the top-down/bottom-up dilemma by opening up a third option: decentralized development with centralized governance.

The technical prerequisites are increasingly in place. Low-code platforms with AI integration enable the rapid development of prototype solutions and their iterative refinement. GitHub integration and version control ensure that the generated code doesn't disappear into a black box but can be professionally managed. Cloud-based architectures allow for immediate deployment and scaling without costly infrastructure projects. API-based integrations allow new applications to be seamlessly connected to existing systems without forcing monolithic reimplementations.

The organizational challenges, however, are considerable. Citizen development without governance inevitably leads to uncontrolled shadow IT with all its well-known risks. Security, data protection, compliance, and maintainability must be considered from the outset, not as an afterthought. This requires new organizational structures: Central IT departments must transform from gatekeepers to enablers, providing platforms, setting standards, and offering support, but leaving the actual development to the business units. Application lifecycle management is essential to control uncontrolled growth without stifling innovation.

These successful examples demonstrate how this balancing act can be achieved. Munich's municipal utility company employs software coaches who support citizen developers in using low-code tools, while central governance structures ensure compliance with security and quality standards. Porsche, in collaboration with MHP, has developed an implementation methodology that combines company-wide standardization with local flexibility. ZF utilizes a digital manufacturing platform that enables individual plants to independently onboard and develop their own use cases within a week, while the central organization provides standards, guidelines, and support.

The disruption of enterprise software architecture

If Nadella is right, nothing less than the end of the enterprise software architecture as it has existed for decades is imminent. The implications for the manufacturing industry would be dramatic. Manufacturing Execution Systems as they exist today could become obsolete, replaced by modular, AI-orchestrated agent systems. The rigid separation between ERP, MES, SCADA, and other production systems would be softened in favor of an intelligent middleware layer that flexibly accesses various data sources and combines them contextually.

This transformation won't happen overnight. Existing systems will continue to run for years, and hybrid scenarios, in which traditional software coexists with AI agents, will dominate the transition phase. But the direction seems clear: software will become increasingly invisible, while interaction will take place through natural language and intelligent assistants. The question is not if, but when and how quickly this change will reach production reality.

The winners of this transformation will be companies that experiment early and build expertise. Integrating low-code development, AI agents, and modern data architectures requires new skills that neither traditional IT departments nor classic manufacturing engineers possess. Successful organizations will need to build hybrid teams that combine technical understanding with process knowledge.

The limits of revolution: Governance as a critical success factor

Despite all the enthusiasm, the risks should not be underestimated. Low-code and no-code don't automatically solve the problems that plagued Excel solutions as well. Shadow IT can develop even with modern tools if clear governance is lacking. Security vulnerabilities, data quality issues, vendor lock-in, and a lack of scalability are real dangers that require strategic management.

The challenges begin with adaptability. While low-code works excellently for simple to medium-sized applications, the platforms reach their limits with highly complex business logic. Specific requirements of regulated industries or highly specialized manufacturing processes may not be achievable with visual editors. In such cases, traditional software development remains indispensable, requiring a clear strategy for determining when each approach is appropriate.

Security is a particularly critical issue. Low-code platforms themselves consist of complex code that can contain vulnerabilities. Because they offer development opportunities to many users, the attack surface potentially increases. Without effective testing methods such as static and dynamic application security testing, insecure applications can emerge that endanger production systems. In safety-critical manufacturing environments, this can have catastrophic consequences.

Vendor lock-in is another risk. Many low-code platforms are proprietary, which makes migrating to other systems difficult and incurs high switching costs. A company that has developed hundreds of applications on a specific platform is practically locked in. These lock-in effects must be considered when making strategic platform choices.

Most important, however, is a functioning governance structure. Without clear rules about who is allowed to develop which applications, how quality assurance is carried out, how security standards are enforced, and how lifecycle management works, chaos quickly threatens. Finding the balance between the freedom of innovation that low-code is meant to enable and the necessary control is difficult, but essential for success.

The future of shop floor digitalization: A decentralized ecosystem

The vision of a future in which production workers develop their own digital tools is neither pure utopia nor unconditionally desirable. It will become reality, but only under specific conditions. The key lies in creating a controlled ecosystem that enables innovation without descending into anarchy.

This ecosystem consists of several layers. The platform layer provides the technical infrastructure: low-code tools, AI agents, databases, APIs, and integration with existing systems. The governance layer defines standards, security policies, quality criteria, and release processes. The enablement layer offers training, templates, coaching, and support to help citizen developers succeed. The community layer fosters knowledge sharing, best-practice sharing, and collaborative development.

In such an ecosystem, applications are not developed in isolation, but within a structured framework. A team leader who needs a new analysis doesn't start from scratch, but uses templates and building blocks that have already been validated. The developed solution undergoes automated security checks and is only put into production after approval. The code is managed centrally, so that other systems can also benefit from it. Updates and maintenance are performed systematically, not ad hoc.

The role of professional developers changes fundamentally in this model. Instead of programming every application themselves, they become architects of the ecosystem, providing platforms, developing complex integrations, ensuring security, and setting standards. They become mentors for citizen developers and curators of the emerging application landscape. This shift is not a devaluation, but rather an enhancement of their role, as they can multiply the impact of their work.

The promise and the reality: A realistic assessment

Twenty years after the proclamation of Industry 4.0, manufacturing digitalization stands at a crossroads. The old approach – either top-down implementation of expensive standard software or a bottom-up patchwork of Excel and Access – has failed. The success rate of around twenty percent speaks volumes. At the same time, the challenges are more acute than ever: skills shortages, global competitive pressure, sustainability requirements, and the need for flexible, resilient production leave no alternative to successful digitalization.

The new wave of AI-powered low-code tools offers a potential solution. Technical prerequisites are improving rapidly, success stories are multiplying, and the economic incentives are compelling. Reducing development costs by sixty percent, shortening time-to-market from months to days, and simultaneously creating solutions that truly fit existing processes – these are convincing promises.

However, caution is advised against excessive optimism. Democratizing software development does not automatically solve all problems; it merely shifts some of them. Instead of overburdened IT departments, we may end up with uncontrolled application sprawl. Instead of rigid, standardized software, we risk incompatible, isolated solutions. Instead of lengthy implementation times, we risk unsafe, rushed projects.

Success will depend on whether companies can create the right framework. Governance without bureaucracy, standards without rigidity, control without paralysis – finding this balance is the real challenge. Technology alone does not determine success or failure. Organizational maturity, cultural change, and strategic management are crucial.

The coming decade: Transformation or disruption?

The next ten years will show whether AI-driven decentralization of software development fundamentally transforms the digitalization of manufacturing or whether it goes down in history as yet another failed panacea. The course is being set now. Companies that experiment early, build platforms, develop expertise, and establish governance structures will reap the benefits. Those who wait or allow the new tools to spread unchecked risk either falling behind or creating chaos.

The provocative thesis that the next generation of shop floor systems will be built locally by the people who actually control production is neither far-fetched nor guaranteed. It will become reality in some areas, but not completely and not everywhere. Hybrid models, in which professional core systems coexist with locally developed extensions, are more likely than a complete disruption.

However, what is very likely is that the role of specialist departments in digitalization will increase massively. The strict separation between IT development and business departments will soften. New competency profiles will emerge that combine technical understanding with process knowledge. The speed of innovation cycles will accelerate because the path from idea to implementation will be drastically shortened.

If Nadella's vision proves correct and business apps are indeed replaced by AI agents, an even more fundamental transformation is on the horizon. The entire architecture of enterprise software as it has existed for decades would dissolve. Manufacturing Execution Systems would no longer exist as monolithic installations, but rather as an orchestration of intelligent agents that flexibly combine data and control processes. This future may still be a decade away, but the development is already well underway.

Regardless of which scenario prevails, one thing is certain: the digitalization of manufacturing as practiced over the past twenty years is coming to an end. The old order, in which IT departments or software companies alone decide the digital future of production, is crumbling. A new era is dawning, in which the boundaries between developers and users, between centralized and decentralized systems, and between standard software and customized solutions are being renegotiated. Whether this new era finally delivers on the promises of Industry 4.0 or merely creates new problems will be decided in the coming years. In any case, the tools for success are, for the first time, truly available.

☑️ Our business language is English or German

☑️ NEW: Correspondence in your national language!

 

I would be happy to serve you and my team as a personal advisor.

You can contact me by filling out the contact form or simply call me on +49 89 89 674 804 (Munich) . My email address is: wolfenstein ∂ xpert.digital

I'm looking forward to our joint project.

 

 

☑️ SME support in strategy, consulting, planning and implementation

☑️ Creation or realignment of the digital strategy and digitalization

☑️ Expansion and optimization of international sales processes

☑️ Global & Digital B2B trading platforms

☑️ Pioneer Business Development / Marketing / PR / Trade Fairs

Xpert.Digital has in-depth knowledge of various industries. This allows us to develop tailor-made strategies that are tailored precisely to the requirements and challenges of your specific market segment. By continually analyzing market trends and following industry developments, we can act with foresight and offer innovative solutions. Through the combination of experience and knowledge, we generate added value and give our customers a decisive competitive advantage.

More about it here:

• Use the 5x expertise of Xpert.Digital in one package - starting at just €500/month