Why Aircraft Maintenance Needs Better Information
UPS is exploring how artificial intelligence could help identify alternative aircraft parts and improve supply chain processes as the aviation industry faces persistent maintenance challenges. Lauren Purcell, an aircraft engineering manager at UPS, indicated that finding alternative parts had been a priority over the previous year.
The initiative reflects a broader industry concern. Aircraft operators are managing aging fleets, limited parts availability, rising maintenance expenses, and workforce shortages. However, UPS has not disclosed which AI systems it is evaluating, whether they are deployed, or what measurable results they have delivered. The more important question is how AI in aviation maintenance can support complex operational decisions without compromising reliability, traceability, or human oversight.
According to Oliver Wyman’s 2026 MRO survey, the aviation maintenance, repair, and overhaul market exceeded $136 billion in 2025, representing 8% annual growth. The industry faces several simultaneous pressures, including component shortages, geopolitical uncertainty, inflation, and tariff impacts. Approximately 90% of survey respondents reported tariff-related effects. Finding alternative aircraft components could help maintenance teams evaluate sourcing options when preferred parts are unavailable. However, identifying an alternative is not the same as approving it for installation.
Are you looking for developers?
Replacement parts must satisfy applicable aviation approval, compatibility, and installation requirements. Parts Manufacturer Approval, or PMA, is a regulatory approval for producing certain replacement or modification aircraft parts. AI cannot independently determine whether a suggested component is authorized for a particular installation. For UPS, the opportunity lies in improving access to relevant information while qualified personnel retain responsibility for maintenance decisions.
Parts shortages are not the only challenge. Oliver Wyman reported that approximately two-thirds of survey respondents considered recruiting aircraft technicians and mechanics moderately to very challenging. Labour rate inflation remained around 5.5% to 6% across most maintenance categories. UPS is also exploring how AI and automation could support training as experienced technicians retire.
Southwest Airlines is examining AI applications for maintenance troubleshooting. A company representative noted that newer mechanics may need up to two years to become fully efficient at troubleshooting aircraft issues. These developments suggest useful applications for AI assistants. Technicians could potentially search maintenance manuals, retrieve service bulletins, examine previous repair records, and identify recurring technical problems without manually navigating multiple information sources.
Such tools would not replace experienced technicians. Their value would come from making approved technical knowledge easier to locate, interpret, and apply within established maintenance procedures. Maintenance organizations often manage aircraft records, inventory data, repair histories, technical documentation, and operational workflows across different software environments.
Are you looking for developers?
An AI assistant cannot produce dependable recommendations if the underlying information is incomplete, outdated, inconsistent, or inaccessible. This is where data engineering and software integration become essential. Organizations need reliable data pipelines, backend services, APIs, access controls, and structured information management. These components allow different systems to exchange information while preserving accuracy and traceability.
The Oliver Wyman survey reinforces this challenge. Approximately 58% of respondents remained in the experimental stage of AI development, unchanged from the previous year. Around two-thirds reported AI initiatives delivering value that met or exceeded expectations. However, poor data quality and limited data availability continued to restrict progress. For UPS and similar organizations, an important lesson is that data integrity begins with established processes for capturing, maintaining, validating, and using information. Adding an advanced model cannot compensate for unreliable records.
Japan Airlines provides another example of the infrastructure work surrounding aviation maintenance. In June 2025, the airline initiated a project to implement IFS Cloud for Aviation Maintenance, with plans to replace a maintenance management system that had operated for more than 15 years. The planned platform covers aircraft, engine, and component maintenance management, spare parts inventory, and supply chain operations.
IFS describes predictive maintenance, real-time analytics, and workflow automation among its platform capabilities. However, those descriptions do not establish that Japan Airlines has deployed every capability or achieved specific operational results.
The broader significance is architectural. Before organizations can connect AI tools with operational workflows, they often need to modernize the systems that store and exchange critical information. Deploying AI in safety-sensitive environments requires more than an accurate model. Applications must integrate with existing systems, follow authorization rules, maintain audit trails, support monitoring, and undergo appropriate testing. Recommendations affecting aircraft maintenance require review through established technical and regulatory processes.
Are you looking for developers?
Observability is particularly important. Engineering teams need to understand which information influenced a recommendation, whether an integration failed, and whether the system relied on outdated records. These requirements also extend beyond aviation. Healthcare, logistics, manufacturing, and other complex industries face similar challenges when connecting AI with operational systems. For companies seeking to move from experimentation into production, specialized software engineering capacity can help address these integration requirements.
Square Codex, a Costa Rica-based nearshore software development and staff augmentation partner serving North American companies, can support organizations through custom software development, backend engineering, API integration, data engineering, cloud solutions, workflow automation, and AI integration.
The objective is not to replace domain specialists or assume responsibility for regulated decisions. It is to help internal teams build reliable software around their existing business processes while retaining ownership of architecture and operational decisions. UPS’s exploration illustrates a broader lesson: successful operational AI depends as much on trustworthy data and connected software as it does on model capabilities.
For organizations working in complex environments, lasting value will come from combining dependable information, integrated systems, qualified professionals, and engineering practices that make AI assistance reliable, measurable, and accountable.