How RTX is using AI and data to transform aircraft maintenance

How RTX is using AI and data to transform aircraft maintenance

Aviation maintenance experts at RTX facilities around the world are using data and AI-enabled analysis to accelerate inspections, spot potential needs earlier and deploy resources faster - all to help make repairs more predictable, shorten shop visits and keep flights on schedule.

Here's how they're doing it.

"We've advanced from early-generation digital tools to models trained on a far more robust dataset. That's what unlocks the speed and accuracy our customers have been asking for."

• Enhanced inspection
• Full-flight data analysis
• Managing MRO resources

Enhanced inspection: Smarter, faster checks from engine core to nacelle

Inside the engine maintenance facilities of Pratt & Whitney, an RTX business, one of the most time-consuming jobs historically has been borescoping, where a technician inserts a tiny camera through a narrow port to find microscopic signs of wear or distress in turbine blades, compressors and other critical components.

A full inspection can take many hours, with thousands of images and video frames to review. And until recently, writing the final report - downloading videos, cropping photos, formatting Word files, typing descriptions - could add several more hours of manual effort.

AI is accelerating that process.

In mid-2026, Pratt & Whitney acquired Aiir Innovations™ software company whose AI-powered borescope tool has been used on the V2500 engine and recently completed pilots on the Pratt & Whitney GTF™ and F135 engines, with plans to expand its use across the company.

As the inspection moves through the engine, the software counts blades, tracks surface patterns and flags areas worth a second look.

"It's like a spellcheck, but for borescope inspections," said Miriam Huijser, associate director for AI Inspection Technology at Pratt & Whitney. "The operator still makes the decisions - the AI just helps make sure nothing slips through."

The software then compares what the inspector found with what the AI flagged, in a way that avoids distracting or biasing the inspector. Then, with a single click, generates a complete, standardized report. Clerical work that once required significant manual effort now takes minutes.

Beyond speed, the AI improves accuracy by serving as a second set of eyes to catch things that humans might miss.

Behind those improvements is software powered by years of data. Aiir Innovations spent nearly a decade training its product on one of the largest and most diverse real-world datasets in the industry - built from actual inspections across multiple engine families and operating conditions.

That depth of data matters because engines around the world wear differently: When they fly over hot, dry places, they accumulate dust. When they fly over oceans, they take in salt. Climate, region and operating conditions change how wear patterns develop - and the AI has learned from all of it.

That broad learning allows the model to recognize patterns across fleets, generalize between engine types and perform well even on platforms it wasn't originally trained on.

Technicians are seeing similar gains at the Goodrich Aerostructures Service Center Asia, also called GASCA - a facility jointly owned by Collins Aerospace, an RTX business, and Singapore Airlines Engineering Company that services nacelles, the large composite structures that encase aircraft engines. The team uses a robotic machine-vision system nicknamed "Spot" to automatically scan what are known as "quick engine change" kits, which consist of more than 150 items, to ensure all are there and ready for installation. That step reduces manual workload by roughly 70%.

Also in Singapore, digital mapping tools give teams a clearer picture of wear across composite nacelle structures. Laser-guided ply-installation systems help them position layers of material more accurately, reducing rework and improving overall repair quality.

Ascentia: Turning full-flight data into foresight

Where Pratt & Whitney's borescoping software provides insight about what's inside the engine, Collins Aerospace's Ascentia® technology gives a look at what's ahead of the entire aircraft.

The Ascentia platform ingests full-flight data - thousands of parameters per second - then blends it with maintenance records and identifies repair needs long before cockpit alerts appear or before technicians can see anything physically.

"Some signs of wear whisper before they shout," said Seth Babcock, general manager for FlightAware® and Tech Ops at Collins Aerospace. "Ascentia is designed to hear those whispers."

Early on, some operators were skeptical and "wanted to test us," Babcock recalled. Rather than act immediately on Ascentia's recommendations, they monitored the specified parts closely. When those parts later showed the exact degradation Ascentia had identified, operators' skepticism turned to trust.

That shift has spread quickly across the industry, and airlines now "expect this level of predictive insight to be part of any major maintenance agreement," said Nicole White, vice president and general manager for Connected Aviation at Collins Aerospace.

Now, replacements can be planned during scheduled downtime instead of during a delay. Recurring patterns across a fleet become visible trends, not surprises.

And rather than troubleshooting an unexpected need, those teams can now incorporate upcoming maintenance into routine work.

"Tomorrow's delays are in today's datasets," Babcock said.

Across several fleets, operators using Ascentia have reduced unscheduled, unexpected removals of components, cut down minimum equipment list write-ups and avoided hours of disruption.

It's similar to the predictive health-monitoring capabilities Pratt & Whitney applies to its own engines. Those capabilities draw from detailed operations and maintenance data to detect potential needs early and plan repairs with greater precision.

Together, these systems give operators a clear view of what's ahead for their fleets in the months and years ahead - and how best to handle it.

FlightSense: Reducing surprises and strengthening operator confidence

Predictive insights are powerful - but they only pay off if the repair is ready to go.

Putting the pieces in place to do that is the job of the FlightSense™ platform - a premium service that helps operators plan for repairs years ahead of time.

"Operators are focused on protecting the passenger experience - avoiding the delays and disruptions that cascade through their schedules," said Katie Atiyeh, who leads airline support for Power & Controls Aftermarket at Collins Aerospace.

In the Asia-Pacific region - one of the fastest-growing aviation markets - the FlightSense service and Ascentia often work hand in hand. With aircraft in heavy use, even small things can snowball, and parts need to be available exactly where and when operators need them.

"Operators don't want surprises - they want confidence," said Kreesan Govender, director of FlightSense at Collins Aerospace. "They want transparency. They want time to plan."

Predictive alerts allow FlightSense teams to pre-position inventory, adjust MRO output and support airlines during peak travel periods. And because Collins carries the required pool of spares, operators don't have to invest upfront in large stockpiles of parts.

For example, in one long-haul fleet, predictive alerts flagged signs of cooling-unit wear. That warning enabled Collins to prioritize aircraft, coordinate repairs and reposition inventory - keeping the fleet on schedule and in service.

And in a world where a single delay or cancellation can spark viral frustration, avoiding it can make all the difference.

"One hour of delay is incredibly frustrating to a passenger - especially when you're stuck on a hot tarmac in Orlando," Atiyeh said.

RTX Intelligent MRO by the numbers

The future: One intelligent MRO ecosystem across RTX

Pratt & Whitney's AI-enabled inspections and Collins Aerospace's predictive analytics show what happens when data helps technicians make faster, clearer decisions.

The next step is connecting those capabilities across RTX - using data created in the design, testing, manufacturing and service life of products, and building a unified predictive backbone for the entire aircraft.

"We're moving from prediction to prescription," White said. "Not just knowing a part will wear out, but modeling its entire life cycle - when to replace it, how to optimize it and how to design the next version better."