Predictive maintenance, accessible know-how, and faster troubleshooting. The panel discussion showed that the real value of artificial intelligence in industrial maintenance lies not only in predicting failures, but also in how companies manage data, documentation, and their employees’ experience.
Predictive maintenance is one of the most frequently discussed topics in industrial digitalization. However, implementing it across an entire operation is not as straightforward as it might seem. As Tomáš Krejčí, Managing Director of Better Industry, pointed out, truly effective solutions are often built around a specific machine or a group of similar machines. What works on one production line cannot automatically be applied to every operation.
Beyond predicting failures, another important question is coming to the forefront: How can companies use the data they already have and turn it into practical support for maintenance technicians?
When Maintenance Technicians’ Experience Is No Longer Locked Away in Their Heads
Bohumír Novák, Production Director at Tenneco, shared a practical example from the manufacturing industry. One of the reasons for implementing the BI Maintenance system was to gain better visibility into maintenance activities. While the company had detailed insight into operators’ work, similar information about maintenance technicians was missing. The company needed to know how many breakdowns were being resolved during each shift, how maintenance resources were being utilized, and what history lay behind individual interventions.
The system also made it possible to record breakdown histories and comments on how they were resolved. However, this is where a typical problem emerges: different employees may use different terms to describe the same fault and explain the repair process in different ways. As a result, the data becomes difficult to evaluate and identify patterns in.
Passing on experience presents another challenge. A significant amount of know-how remains in the minds of experienced employees who are gradually approaching retirement. At the same time, companies struggle to recruit new maintenance staff, particularly for multi-shift and continuous operations.
The situation is further complicated by the diversity of machinery and technical documentation available in different languages. Navigating manuals can be time-consuming even for an experienced maintenance technician.
AI as an Assistant in Identifying Causes and Resolving Failures
Světlana Margetová, Managing Director of Pulsar Solutions, sees the potential of artificial intelligence precisely in these situations. Her solution works with technical manuals and practical experience to help employees find relevant information for resolving specific problems.
For example, AI can identify recurring faults, suggest repair procedures, or help assess whether preventive maintenance intervals should be adjusted. The goal is to reduce time-consuming searches through documentation and make the necessary information directly accessible to the people who need it in their daily work.
An important feature of such a solution is its use of reliable sources. If the system cannot find relevant information, it should not make up an answer. Users should also be able to locate the specific passage in the documentation on which a recommendation is based. AI can therefore support maintenance technicians without replacing their professional judgment or their responsibility for carrying out interventions safely.
Integrating Systems Can Deliver the Greatest Benefits
A standalone tool for searching information is useful. However, its potential grows when it can work alongside the system a company uses to record breakdowns and repairs.
Krejčí pointed out that Better Industry helps reduce breakdown response times through its BI Maintenance system. Another possibility is to build on this data and connect it with Pulsar Solutions’ solution. This would make it possible to better link the initial fault report with the information needed to resolve the issue.
According to Margetová, companies do not need perfectly structured data to get started. The basics are identifying the machine and gradually collecting a history of breakdowns and repairs. The system can also handle different ways in which employees describe their maintenance interventions.
Both systems can subsequently benefit from this integration. AI can provide additional information about affected components, error codes, and repair procedures, while the maintenance system supplies further data for evaluating recurring faults and identifying opportunities to make the entire process more efficient.
Technology Must Fit the Company’s Requirements
When implementing AI, companies must also consider infrastructure, implementation costs, and data security. Solutions can be deployed locally, in the cloud, or in a hybrid environment. Each option has different requirements in terms of hardware, deployment speed, and data management.
Local deployment can be particularly suitable for companies that need to keep their technical documentation and internal know-how under their own control. A hybrid approach, on the other hand, allows local processing to be combined with selected online services. The choice should depend on the company’s requirements, infrastructure, and security policies.
Krejčí also emphasized that Better Industry’s approach includes providing data and making it accessible to AI systems through its AI connector service, AI CONNECT. This new service from Better Industry leaves the choice of language model and its security configuration in the hands of the customer and their IT department. Better Industry makes data available to language models, enabling them to perform better-informed, context-aware analysis.
This is another important aspect of the overall project: AI cannot be implemented separately from a company’s data management policies.
It Is Not Just About Technology, but About Making Use of What a Company Already Knows
The discussion showed that the benefits of AI in maintenance do not have to start with complex predictions of future breakdowns. Simply making documentation accessible, standardizing information about past repairs, and preserving the knowledge of experienced employees can already deliver significant value.
For companies, this means an opportunity to reduce the time spent searching for the right procedures, make better use of maintenance resources, and gradually build a knowledge base that does not depend solely on individual employees.
The key to success is not artificial intelligence alone, but quality data, integrated systems, and solutions that genuinely make maintenance technicians’ everyday work easier.