Optimizing inventory for maintenance, repair and operations (MRO) is one of the hardest challenges for managers involved in asset care. Stocking too little or the wrong items can increase machine downtime, but carrying too many ties up capital unnecessarily, requires more storage space and risks waste due to obsolescence.
Artificial intelligence can create the most value in MRO when it connects asset criticality, failure risk, spare parts planning, predictive maintenance, and maintenance execution. However, simply implementing AI into MRO workflows doesn’t guarantee success. It needs accurate data, clean records, defined workflows, and expert interpretation.
Introduction to AI in industrial MRO
Computers have long been used in MRO, but conventional MRO software struggles to find patterns and optimize multiple variables in large datasets. AI provides an alternative because, rather than following rigid algorithms and formulas, it learns to spot patterns and identify anomalies.
AI systems learn by being trained on a set of data. This can be supervised, where the data is labeled, or unsupervised. In unsupervised learning, the AI software is left to find patterns and trends by itself.
Once trained, an AI program can be deployed to handle the task for which it was trained. In industrial MRO, applications include inventory optimization, predictive maintenance and demand forecasting, task optimization and human practice augmentation.
Looking forward, AI technology will help the MRO organization extract higher levels of productivity and efficiency from their activities while reducing downtime and driving down inventories. This will flow from the ability to work with ambiguous records and data while simultaneously identifying trends and relationships that provide new insights into maintenance activities.
Traditional MRO software | AI-enabled MRO |
Relies on defined rules and manual inputs | Learns from patterns in large datasets |
Tracks inventory and work orders | Identifies trends, anomalies and risk |
Supports reporting | Supports prediction and optimization |
Depends heavily on clean manual categorization | Can help normalize inconsistent descriptions |
The role of AI in enhancing MRO efficiency
Three activities in particular stand to benefit from MRO AI:
- Maintenance strategies
- Inventory management
- Data analytics
Maintenance strategies
Scheduled preventive maintenance reduces unplanned downtime but carries risks. It’s possible to do too much maintenance, which increases costs and consumption of MRO items. It’s also possible to do the wrong type of maintenance, raising costs but not preventing breakdowns.
By providing insights into breakdown history for similar types of equipment, AI can anticipate the type and frequency of likely failures. This can improve preventive maintenance planning. It can also support implementation of predictive maintenance, where machine health monitoring data is used to identify rising failure probabilities. AI-enabled maintenance strategy is most valuable for assets where failure history, downtime cost, or spare parts risk shows that the current preventive maintenance schedules are not enough.
Inventory management
MRO inventory is more diverse and complex than that used for manufacturing. Some items are critical yet seldom used, others are hard to obtain and a few become obsolete as equipment is updated and replaced.
ERP, while sometimes used, is not designed for this type of inventory management.
AI can help optimize MRO inventory in many ways. It can understand and relate the differing descriptions that are often used, especially in multisite maintenance operations. It can spot consumption trends and patterns (seasonality, for example), reduce obsolescence by improving linkages between machinery and spare parts, and address supply chain challenges like extended lead times. One of the biggest benefits of AI is how it can help prioritize critical spares based on asset criticality, supplier lead time and downtime impact.
Data analytics
On large industrial sites and in multisite maintenance scenarios, there’s often too much data coming from the machines and MRO stores to analyze manually. Computerized Maintenance Management Systems (CMMS) help, but with adoption of predictive maintenance the problem is only growing bigger.
AI excels at identifying patterns, trends and connections that would otherwise remain hidden in MRO data. This enhanced visibility gives managers new opportunities to eliminate duplication and waste while improving equipment availability. Not only can AI compare reliability performance across plants, but it also can help identify duplicate parts, inconsistent naming, and recurring failure patterns.
Looking to reduce stockouts, excess inventory or emergency purchases?
Integrating robotics and automation in MRO
MRO operations are seeing significant investment in advanced technology. Examples include:
- Robotic stores: Automated storage systems increase volumetric space utilization while improving record accuracy and eliminating the manual work of putting away and subsequently retrieving MRO items.
- Robotic welding: Repair work can often involve complex welding operations. Robotic welding cells, programmed offline, are faster and more consistent than performing the same job by hand many times.
- Automated inspections: Robots and even drones can access areas of a plant that are difficult, dangerous or expensive to reach, letting technicians inspect the images in safety.
A sometimes overlooked aspect of robotics and other automation in MRO is the potential to enhance sustainability. This results from reduced waste, coupled with opportunities to save on energy as used for HVAC and lighting.
Leveraging cloud computing and IoT in MRO
Industry 4.0 technologies—low-cost sensors, communication technologies and powerful analytics tools—are transforming manufacturing operations. They can do the same in MRO. Three areas are:
- The Industrial Internet of Things (IIoT): Sensors with computing and communication capabilities can transmit data on machine health in real time to the CMMS or other analytics packages. This data can inform maintenance timing as well as spare parts planning.
- Cloud computing: Instead of restricting access to a few users with limited computing capabilities, upload machine data to the cloud for processing with powerful AI algorithms to obtain valuable insights into machine performance. With cloud-based analytics, manufacturers also gain support for multi-site MRO visibility.
- Digital twins: Virtual representations of complex manufacturing systems, with direct data links to the physical equipment, provide ways of commissioning, testing and training without disrupting production. This enables evaluation of various maintenance scenarios and inventory needs before any changes to procedures are made.
AI in supply chain management and optimization
MRO procurement is often complicated by the need to work with large numbers of suppliers. AI can help by:
- Finding patterns in lead time data that help optimize ordering and avoid expedited shipping
- Finding opportunities to consolidate purchasing for multiple sites
- Determining criticality at the component level
- Working across different data types to consolidate purchasing needs
- Monitoring supplier risk and identifying alternate suppliers
- Optimizing reorder points and consolidating spend
- Reducing emergency orders and planning for long-lead parts
Challenges and solutions in adopting AI for MRO
For much of the MRO industry, AI is a new tool that poses some challenges. Chief among these are:
- Quality of training data: Every AI tool is only as good as the data it’s trained on. To obtain useful results from AI MRO, the software must be trained on data sets that have been “cleaned” and checked for accuracy.
- Limited number of examples: For most industrial machinery, breakdowns are rare. This limits the number of cases the AI system can be trained on.
- Obtaining actionable recommendations: While AI is exceptionally good at finding patterns, this doesn’t automatically translate to actionable insights. A business adopting AI for MRO must consider how AI conclusions will be interpreted and applied.
- Determining component criticality: This requires expert knowledge of the machinery on which the part or component is used.
When implementing an AI-driven solution in MRO, managers must take the time to understand its limitations and requirements. Careful attention is required to training for maintenance teams, which should be followed by extensive testing. This will build the confidence in its capabilities.
Get expert advice on AI and industrial MRO
MRO hasn’t always been the most technologically advanced part of most industrial manufacturing businesses or even of the maintenance function. Today, a reliance on spreadsheets, paper-based systems and sometimes misguided use of the ERP system, is giving way to AI solutions.
AI inventory optimization promises to reduce the capital tied up in MRO and to address the problem of spare part obsolescence. In addition, AI will help optimize maintenance activities, augment human practice, and improve operational efficiency. Manufacturers slow to adopt AI in industrial MRO will almost certainly lose ground to competitors.
Maintenance is a complex activity. Implementing an AI MRO solution is a large project that requires detailed understanding of both AI technology and industrial maintenance practices. Businesses wishing to investigate this further should start by discussing options with their MRO provider or an expert in outsourced maintenance. Contact us to initiate that process.
