AINsight: Too Much AI Talk, but Maybe It Can Help MROs

During roundtable discussions earlier this week at AIN’s Corporate Aircraft Leadership Summit - Maintenance (CALS MX) in Scottsdale, Arizona, I asked maintenance directors about the software that they use and what missing features could improve efficiency. The wide-ranging discussions with those responsible for maintaining corporate fleets covered everything from predictive maintenance to troubleshooting and tracking and artificial intelligence (AI). (Ed. note: CALS participants are encouraged to speek candidly, and AIN agrees not to identify them.)

Although some participants felt that they are being bombarded by AI-related solutions, there was some agreement that the technology might be helpful. Some companies are using Camp’s Corridor software to help manage their maintenance departments, and the roundtable participants were interested to hear about Aero Star Aviation’s development of its own Corridor-based AI tool to help mechanics learn about maintenance practices and troubleshooting solutions that are part of their peers’ tribal knowledge.

“As a new mechanic, you can’t always find the lead [to ask a question],” said one participant.

A representative from maintenance documentation and tracking provider Veryon said, “I imagine it's like that in every hangar." Veryon users can search across all of the information in the company's products, he pointed out, and it will cite the source so the user can confirm the quality of the results. "The generative AI is only as smart as what you feed it," he concluded. 

One participant said he was appalled to learn during an AI webinar that a company was just dumping maintenance manuals and information into an AI tool, without considering that the manuals are proprietary and belong to the aircraft OEM. Although this may be happening more than OEMs prefer, it is up to OEMs to decide how their information is used. Aero Star’s AI tool, for example, uses only its own information based on work done by its mechanics and doesn’t attempt to input any OEM proprietary products.

In any case, despite some resistance by maintainers to adopt AI tools, they did admit that there are plenty of opportunities to apply AI in the maintenance arena. One company is seeing benefits from its flight operational quality assurance program, not just for pilots but also as a resource to tease out maintenance-related issues.

One of the biggest complaints of many of the maintenance roundtable participants is that there still is no universal software package that does everything they need. The result is that their maintenance operations depend on multiple software platforms and, despite many promises, integration between these platforms is often lacking.

Software developers often promise that their system solves all these problems, but that isn’t necessarily the case. It is important for companies considering a particular software platform to test it against their processes and expectations and not to make any assumptions about what the software can do before buying and implementing the package.

AI may be able to help with these challenges, but still has a long way to go, many participants agreed, and doesn’t alleviate the need for mechanics to learn their trade. Mechanics can’t just rely on downloading data from an aircraft to figure out what’s wrong, one participant said. “Go and figure it out,” he said. The laptop or iPad isn’t going to fix that leaky strut, he added. Physical skills still matter.

Another maintenance director pointed out that understanding how the system works is still important, as is putting hands physically on the problem. While working on helicopters in the military, he was faced with an intermittent fuel gauge, which turned out to be caused by a wire bundle that was getting inductance generated by a nearby metal flight control tube. “We only found it by pulling the bundle away,” he said.

“You have to be hands-on,” said another participant.