Rodriques Johnpeter, global industry segment manager, Harting
Manufacturers are operating in a moment where disruption is the norm, not the exception. Supply chain volatility, inflationary pressure and labor constraints all collide with a parallel push to embed automation and AI into plants that were never designed for today’s technologies.
Most facilities – well over two-thirds globally – are still brownfield sites, built around legacy machines, heterogeneous controls and fragile integrations that have been patched together over decades.
These environments leave almost no room for extended downtime or wholesale rip-and-replace projects, which is where many digital transformation strategies quietly break down.
At the same time, there is ample proof that meaningful automation and AI gains are possible on real shop floors, not just in slides and lab pilots.
Forward-looking manufacturers are steering investment toward technologies, such as supply chain digital twins and modular, plug-in components, that can live alongside installed equipment without massive rewiring or extended shutdowns.
By putting business outcomes first, designing for interoperability from the outset and favoring systems that can scale in small, low-risk increments, they are converting chronic pain points into durable advantages.
In this climate, leadership is no longer defined by who touts the most ambitious Industry 4.0 roadmap, but by who can move the needle with the factory footprint they already own.
Why execution outperforms vision
Many industrial operations still run in buildings that predate the internet, built around a mix of proprietary controls, aging PLCs and stand-alone production islands that were never meant to talk to each other.
These plants are expected to run nearly 24/7, which means the tear it down and start over mindset found in many digital roadmaps is a nonstarter.
Today, a large share of factories are wrestling with fragmented OT/IT environments where every integration is a custom job and unified data access is a daily struggle for engineering and maintenance teams.
A major culprit is what’s often called the greenfield illusion: assuming transformation can start from a clean, sensor-rich slate, with uniform connectivity and standardized data pipelines.
In reality, execution-focused leaders accept that true greenfields are rare, so they design programs around incremental, interoperable upgrades that directly support specific targets such as a defined percentage reduction in unplanned downtime or a measurable increase in line throughput.
Execution in this context is about steadily improving visibility, reliability and flexibility, rather than betting everything on an eventual full overhaul that may never be funded or completed.
By working within constraints instead of waiting for them to disappear, these manufacturers demonstrate that practical progress beats hypothetical perfection.
Where automation and AI initiatives go off track
When robotics, automation platforms or AI-driven analytics are introduced into these legacy environments, interoperability usually becomes the first obstacle.
New robots, vision systems or edge controllers often cannot natively communicate with older PLCs, fieldbuses or proprietary sensor networks, creating integration “dead zones” that slow or halt deployments.
It is not surprising that a clear majority of manufacturers identify connectivity and interoperability as their primary barriers to scaling AI across multiple lines and plants, frequently resorting to expensive middleware or complete hardware replacements to bridge the gap.
Data maturity adds another layer of friction, with 54 percent of industrial leaders citing data quality and availability as the top challenge blocking scalable AI.
Plants generate huge volumes of operational data, but it is often scattered across different systems, stored in incompatible formats or locked inside vendor-specific platforms, making it difficult to feed into AI or advanced analytics without extensive pre-work.
Gaining consistent, trustworthy data streams without impacting production schedules is a recurring challenge, and leaders regularly cite siloed data as the top obstacle to enterprise-wide visibility.
The result is a familiar garbage in, garbage out problem: models that look promising in controlled tests fail to deliver reliable insights when exposed to noisy, incomplete or biased real-world inputs.
Finally, cultural and skills issues frequently determine whether automation and AI projects stall or scale.
Operations and maintenance teams are incentivized to protect uptime and often view experimental technologies as risk, while IT teams may champion cloud-first approaches that do not align with plant-floor latency, safety or resiliency requirements.
Without shared ownership between OT, IT and production leadership, initiatives can get stuck in an endless proof-of-concept loop.
At the same time, workforce development often lags the technology, leaving gaps in skills required to troubleshoot hybrid systems, interpret AI recommendations or maintain increasingly software-defined equipment.
These human dynamics are a major reason why a large percentage of industrial AI efforts never progress beyond pilot scale.
Technologies and practices that actually move the needle
Manufacturers that are breaking through brownfield limitations tend to focus on technologies that deliver quantifiable value without forcing complete system replacements.
Digital twins of supply chains, lines or critical assets are a prominent example, allowing teams to test changes, optimize flows and anticipate failures in a virtual environment before making adjustments in live production.
These models can unlock significant cost reductions and efficiency gains when used to guide scheduling, maintenance and capacity decisions.
Equally important is a shift toward plug-and-play infrastructure. Modular, standardized connectivity – from intelligent connectors to flexible cabling and I/O systems – lets engineers insert new robots, sensors or edge devices into existing architectures with minimal rework.
Instead of overhauling entire electrical and networking backbones, teams can extend what is already in place, shortening commissioning times and reducing risk.
This approach aligns naturally with a value-first mindset, where automation and AI are introduced to solve clearly defined problems such as faster changeovers, reduced scrap or improved energy management.
From a strategic standpoint, leaders are also using open standards and interoperable architectures as guardrails for every new deployment.
By prioritizing components and software that can integrate across vendor boundaries, they avoid future lock-in and keep capital expenditures more predictable over multi-year horizons.
To make these systems sustainable, they invest in the human side: giving plant teams timely access to actionable data, targeted training on new tools and collaborative platforms that make it easier to coordinate OT, IT and engineering work.
In many cases, modernization becomes less about ripping out legacy hardware and more about intelligently connecting and augmenting what already works.
A more pragmatic playbook
Strategic roadmaps for smart manufacturing and AI-driven production are now ubiquitous, but often collide with the downstream reality of aging equipment, patchwork data systems and a culture that cannot tolerate extended downtime.
The real differentiator is not the sophistication of the slide deck, but the discipline with which organizations execute: addressing interoperability early, building a robust data foundation and backing technologies that have a clear, demonstrable impact on uptime and efficiency.
For manufacturers competing in robotics- and automation-intensive markets, the path forward will not be a single leap to an ideal future state. Instead, it will be a series of deliberate, interconnected steps, each one reinforcing a more flexible, data-driven and resilient operation.
