As AI continues to gain momentum, many industrial companies are facing an uncomfortable truth: their ability to innovate is being held back by decades of IT complexity. The issue is not a lack of ambition, but fragile foundations. Fragmented data, siloed systems, disconnected product models… Before even considering automation, one priority stands out: restoring reliable digital continuity. In this op-ed, Cédric Kalifa from PTC explains how this can be achieved.
According to a Frost & Sullivan study on the Connected Product Lifecycle model, 93% of industrial companies still report suffering from data silos. This figure highlights a paradox: while companies are eager to accelerate their use of AI, their technology environments still struggle to circulate information effectively. Data exists, but it remains fragmented, difficult to leverage, and sometimes unreliable.
This situation naturally creates frustration. In practice, the most effective response is a phased approach. Rather than multiplying POCs (proofs of concept) or adding more AI tools, leaders should focus on the basics: producing reliable, contextualized, and current data. In short, they need to build the digital continuity that is still too often missing.
What kind of debt are we talking about?
Industrial technological debt stems from several well-identified dynamics:
- The rapid evolution of equipment and products: Machines now integrate electronics, software, automation, and digital services. This added complexity leads to an explosion of data, which often accumulates in heterogeneous environments with little overall consistency;
- Organizations that remain too siloed: While collaboration between engineering, production, and logistics is increasing, it is still not fully structured. In practice, this results in gaps between engineering-defined bills of materials and actual production conditions. Shop floors frequently rely on their own systems, limiting feedback loops and preventing immediate alignment across teams when changes occur;
- Excessive customization of IT systems: This reflects an underlying IT architecture debt. Over time, many industrial companies have built their information systems through successive layers including local tools, parallel data repositories, ERP customizations, etc. While these choices address short-term operational needs, they ultimately lock the architecture, making it harder to evolve and less suitable for optimization or AI-driven initiatives.
Overcoming technological debt: a three-step approach
Addressing technological debt does not mean replacing everything or launching a large-scale transformation program. In an industrial context, such approaches are often unrealistic. Existing systems support critical processes, historical datasets, and business rules that ensure daily operations. The real challenge is not to start from scratch, but to gradually regain control over three key pillars: data, processes, and people.
The first step is to identify the areas of debt that genuinely impact performance, as not all debts carry the same weight. For instance, a misaligned product bill of materials across engineering, production, and procurement leads to cascading errors, delays, and additional costs. These critical areas must be tackled first.
The second step focuses on information governance. While many industrial companies possess large volumes of data, only a small portion is reliable, contextualized, and usable. Establishing clear “sources of truth” is therefore essential: who owns product data? Who validates changes? How are updates shared across functions? Without this framework, companies fall into a vicious cycle of rework, inconsistencies between systems, and decisions based on incomplete information.
The third step is more organizational than technological. It involves reconnecting teams such as engineering, industrial methods, shop floor, procurement, quality, maintenance, logistics, around a shared information backbone. This digital continuity depends on the ability to circulate reliable, up-to-date, and accessible information throughout the product lifecycle. As a result, field feedback can inform product evolution, while regulatory and operational constraints are seamlessly integrated into workflows.
The illusion of a “shortcut” through Artificial Intelligence
Some industrial companies may be tempted to rely on AI to compensate for or mask this technological debt. This is a risky assumption. While AI can help identify inconsistencies or automate document searches, algorithms applied to fragmented or biased data cannot deliver reliable outcomes. On the contrary, they risk reproducing or even amplifying existing dysfunctions.
Ultimately, without digital continuity, industrial companies will continue to multiply technological initiatives without ever fully capturing their value. The priority is not to inject intelligence into siloed systems, but to ensure these systems can first produce and share trusted data.
