Every year, billions of dollars’ worth of equipment is returned on warranty claims - only to be found in perfect condition. Can AI prevent this wastage?
When three hundred passengers are already loaded onto a plane, airline customer satisfaction demands that line maintenance simply swaps out a troublesome module instead of undergoing a multiple-hour repair out on the tarmac. This, however, leads to staggering wastage, and not just in aerospace — nearly every industrial segment experiences this problem.
The future for competitive enterprises is to produce data-driven solutions that address this wastage by linking all stakeholders across the supply chain — gathering the data and domain expertise from each constituent and normalizing the aggregated intelligence into one democratized AI platform. If they cannot link the stakeholders, they will be left behind by those that do.
Currently, in the world of competitive product support and brand promise, it is necessary for parts suppliers & OEMs to honor a warranty claim before they are able to run the component through a full diagnostics. The replacement parts are supplied and paid for before the warranty departments can completely assess the claim’s validity.
The trouble codes coming from equipment help increase diagnostic accuracy, but they do not always perfectly capture the actual fault in the equipment. Complex equipment with many interdependent subsystems will sometimes fire codes unrelated or only tangentially related to the actual cause of the problem. As such, functional equipment is frequently returned either mistakenly or fraudulently. These are “No Fault Found” or “No Trouble Found” warranty returns.
This is no small problem. US-based Aerospace companies spend roughly $1.5 billion a year on warranty claims, and the worldwide automotive industry spends roughly $50 billion on warranty claims. Other manufacturers of mission-critical equipment with complex subsystems face similar challenges.
Statistics on the prevalence of No Fault Found claims are less public, but a 2007 report noted that rates are frequently as high as 50% - and our conversations with industry insiders have confirmed that they have improved only slightly in the decade since.
The complexities of modern equipment have made it difficult for front-line service providers to provide their technicians with diagnostic tools and heuristics accurate enough and rapid enough for them to quickly assess the true cause of the servicing scenario. Comprehensive diagnostics take considerable time.
Dispatchers, technicians, warranty managers, product engineers, product support specialists - connecting all stakeholders in the warranty ecosystem with data-driven, intelligent tools would go a long way towards reducing NFF inefficiency.
These products can deliver the expertise of the engineering departments (where the faults, or lack of them, are eventually found) immediately to the servicing situation. This will allow more comprehensive diagnostics tests to be performed at a fraction of the time, increasing the scope of problems that can be assessed and the accuracy of the assessments.
This data-driven ecosystem can only be successful if there is continuous learning, powered by artificial intelligence, providing visibility across all of the stakeholders, taking the issues in the field and converting it into usable insights for everyone involved.
For example, this would help engineering departments increase the correlation between a trouble code and the actual problem with the equipment — which then helps proper routing of issues at the front line and field diagnostics.
The brick wall is that this ecosystem is not vertically integrated.
The biggest challenge with this solution is a lack of visibility across the supply chain. Much of the data collected from equipment does not make it back upstream — currently, OEMs, independent servicers, and dealerships have little reason to share valuable IoT and servicing information.
New supply chain models instituted on a broad, strategic level can bake incentives into the contracts in order to ensure that data from the final product is passed up the chain to each of its constituents. Maybe the incentive is an overall lower cost of parts and services. Or, distinct parts of the supply chain can do the heavy lifting with the data, and supply intelligent solutions to the rest of the constituents in order for access.