The Candid POV of Jon Hall and Yoann Urruty on AI in Industrial Maintenance
/
read
Jon Hall and Yoann Urruty are no strangers to the industrial maintenance world. Both have spent their careers helping organizations navigate the challenges of reliability, digitalization, and technology adoption. And today, as artificial intelligence (AI) dominates headlines, they bring a perspective that many in the industry need to hear.
Hall and Urruty often compare AI to the “shiny object” that seems impressive at first glance but rarely addresses the core need.
In maintenance and reliability, AI has become that shiny object. Everyone is talking about it, but too few are asking the hard questions: Are we ready? Do we have the foundations in place to make AI work for us?
Their point of view is clear: while AI has potential to create the new top performers in maintenance, it cannot compensate for weak fundamentals. For most organizations, the real journey begins not with AI pilots, but with cleaning up data, strengthening daily workflows, and empowering people.
Why AI Isn’t the Silver Bullet in Maintenance
The Limits of Current AI
AI in its current form, what experts call Artificial Narrow Intelligence (ANI), can do impressive things. It can process sensor data, generate reports, flag anomalies, and even support administrative tasks. But it remains narrow. Unlike Artificial General Intelligence (AGI) or Artificial Superintelligence (ASI), concepts often hyped in popular media, today’s AI cannot make complex judgments in the field, replace operator intuition, or account for nuanced real-world conditions.
Hall and Urruty emphasize this distinction because many organizations mistakenly assume that AI will “solve” maintenance. In reality, it can only assist in specific domains, and even then, only if the underlying data and processes are reliable.
The Foundation Problem
This brings us to what both experts call the foundation problem. Too many plants operate in what can best be described as data chaos: incomplete Master Equipment Lists (MELs), missing Bills of Materials (BOMs), unclear criticality rankings, and maintenance strategies that exist more on paper than in practice.
Yoann often illustrates this with a simple but costly example: a motor failure that could have been prevented with a clear lubrication strategy and accurate asset data. Without those basics, deploying AI doesn’t create intelligence, it just accelerates failure. Or as Urruty puts it, “AI amplifies what’s already there, if the foundation is chaos, you just get faster chaos.”
The Path to Digital Maturity
The Six Levels of Maturity
According to Hall, digital maturity is not a switch, it’s a ladder. They describe six levels of progression:
- Pre-Digital – Paper-driven, reactive work.
- Digital Silos – Point solutions with limited integration.
- Connected – Systems start to talk to each other.
- Predictive – Analytics anticipate failures.
- Adaptive – Processes adjust dynamically based on insights.
- Autonomous – True self-optimizing systems.
The danger, they warn, is when organizations try to jump from level one or two directly to predictive or autonomous capabilities. “That’s putting the cart before the horse,” Hall explains. The result is wasted investment and frustrated teams.
The Role of the OT Data Fabric
Central to climbing this ladder is the OT data fabric: the system by which data is collected, contextualized, governed, and assured for quality. Without this, AI models are built on sand. Reliable, trustworthy data is not a nice-to-have; it’s the very prerequisite for analytics, predictions, and eventually autonomous operations.
Everyday Digital Work: Where ROI Really Appears
Digital Work Execution
Both Hall and Urruty argue that the greatest return on investment today comes not from advanced AI but from digitizing everyday work. Digital procedures, inspection routes, and collaboration tools reduce downtime, cut errors, and free technicians from repetitive paperwork. The value is immediate, measurable, and scalable.
Workforce Enablement
Another overlooked source of ROI is workforce enablement. When training is embedded into digital workflows, when technicians can access AR instructions on demand, and when Generative AI assistants provide contextual support, organizations see not only fewer mistakes but also faster skill development. In an era of talent shortages, this is where digital truly pays off.
Predictive and Proactive Domains
That said, predictive tools still matter. But Hall and Urruty stress that predictive maintenance should focus first on reducing failures and improving reliability not chasing full autonomy. Proactivity is the sweet spot where digital technology consistently delivers value without overpromising.
Building the Roadmap for AI Readiness
Four Enablers
A successful roadmap for AI readiness rests on four enablers:
- Data – Clean, complete, contextualized.
- Workflows – Streamlined and standardized.
- People & Culture – Skilled, engaged, and digitally fluent.
- Technology Stack – Integrated, scalable, and future-ready.
Neglect any of these, and the roadmap quickly derails.
Reliability Foundations
Equally essential are the reliability foundations: Asset Master Data, Criticality Ranking, and robust Maintenance Strategies. These are not glamorous, but they are the non-negotiables.
Pragmatic Progression
Finally, Hall and Urruty advocate for a pragmatic approach. Organizations should conduct a self-assessment to identify their place on the maturity curve, then focus on closing the most valuable gaps first. Progression is incremental, not revolutionary. The goal is not to “catch up” with hype cycles but to build sustainable readiness for AI.
Conclusion
AI is here, and it will shape the future of industrial maintenance. But Hall and Urruty’s candid point of view is a reminder that it is not a silver bullet. AI amplifies what already exists. On solid foundations, it can drive reliability and efficiency. On weak foundations, it simply produces faster chaos.
Their advice to maintenance leaders is simple: resist the temptation of the shiny object. Instead, focus first on strengthening data, workflows, and reliability practices. Assess your maturity, close the critical gaps, and then experiment with AI where you are ready.
The message is not to slow down, but to build wisely. Because in the end, readiness is what separates organizations that thrive with AI from those that waste their investment chasing it.
Raphael Tremblay,
Spartakus Technologies
[email protected]
