Few disagree that Gen-AI has the potential to shake up the aviation supply chain as an enabler of predictive maintenance and all-round optimisation.
In a similar fashion to the anticipation in warehouses, airports and CEP distribution centres, there’s growing excitement about how the industry can leverage Gen-AI to detect patterns not obvious to the human eye.
In the aviation supply chain, this can enable Used Serviceable Material (USM) suppliers to predict when airlines might need new parts for their planes before the parts begin to fail.
Furthermore, the prediction enables them to keep USM inventories fully stocked, so there is never a shortage.
But there’s a problem: Gen-AI needs data – lots of it. And, as things stand, most data relating to replacement parts in the aviation supply chain has not been digitalised.
Gen-AI works well when OEMs share all their data with the USM suppliers – and this is true in the case of modern aircraft. So in 20 years’ time, it’s safe to assume all data will be available for Gen-AI to work its magic.
But Gen-AI tends to struggle with parts produced 10-20 years ago. The parts might have serial numbers, but if their entire usage history hasn’t been digitalised, then USM suppliers can’t be sure about their track record and future durability, and Gen-AI can’t make predictive calls with certainty.
Such records have always required a great deal of paperwork – often files were uploaded to servers such as DropBox, making it inefficient to find the data years later – and often the entire history wasn’t adequately documented.
Even though parts without documentation are effectively worthless, only essential data was stored – all other information, even though it might be valuable further along the supply chain, was lost.
All of this leaves major gaps in the life-story of many parts.
Gen-AI is used to assess data on blockchains – a cryptographically linked chain of immutable data recorded digitally on a ledger (‘blocks’).
Since the turn of the decade, many aviation companies have been integrating blockchain into the way they keep records of parts and materials.
The blockchain includes all the relevant information pertaining to an item (an engine part, for example) or materials – such as their origins, usage history, maintenance logs etc – which provides a highly efficient way to keep track of them throughout their life cycle.
Accessible via a digital, globally accessible platform shared among ecosystem partners, all that is needed is the item’s hash or block number to check its blockchain.
But again, none of this is possible without data, which explains why the momentum behind blockchains has stalled of late.
Blockchains are a good fit for AI because they are secure – huge efforts are made to ensure the data cannot be falsified, and this is undoubtedly one of blockchain’s biggest selling points.
Blockchain offers huge benefits to the aviation supply chain industry, as it provides total transparency for critical parts used in planes, along with other tangible and intangible assets.
But many players in the aviation supply chain industry continue to only keep track of configuration data in their own systems. The result is too many loose ends.
By resisting blockchain – and not integrating their data with other companies’ data – collaboration in the supply chain will continue to be limited.
Blockchain has been identified as one of the technologies that will have a massive impact on not just the future of the aviation supply chain industry, but also aviation too. The hype has not been too dissimilar to the way many major aviation players viewed the arrival of the internet.
Already in 2018, 70 percent of respondents to an Accenture survey said they were hopeful blockchain could help the industry detect the increasing amounts of falsified data entering their systems.
In a 2018 note, Accenture highlighted the huge potential it saw in blockchain technology for the aviation supply chain industry:
“Blockchain is well-suited to improve the performance of one of the world’s most complex, globally interconnected and security-dependent supply chains. This elegant and paradigm-shifting technology has the potential to deliver profound benefits for the hundreds of suppliers typically involved in the manufacturing of a single aircraft.”
It won’t be long before blockchain becomes the standard, contends Deloitte:
“The future will likely be a place where every part has a trusted digital identity and historical record that is immediately accessible to those with the right credentials. Before long, many airline CFOs will wonder how they did things the old way. With all paths open, now is the time to start small – and to think big.”
Gen-AI’s slow development can also be seen outside the aviation supply chain industry.
Despite large investments, the core business of industrial enterprises is not yet AI-enhanced.
Again, industry surveys carried out by Deloitte and Accenture support the notion that data issues constitute the main reasons for the insufficient adoption of AI in industrial enterprises.
Data management, democratisation and governance issues are also doing their best to derail Gen-AI. Other areas of concern include metadata management, data architecture and data ownership.
Furthermore, there is a tendency to use Gen-AI in an insular fashion, thus engineering a polyglot and heterogeneous enterprise data landscape.
This makes systematic data management, comprehensive data democratisation and data governance challenging, preventing the widespread use of AI in industrial enterprises.
Much depends on the data ecosystem. Without defined platforms and roles, AI will continue to be insular at a time when it needs to become fully industrialised.
Nobody disputes that Gen-AI has the potential to radically improve the performance of the aviation supply chain industry with its real-time quality prediction. But there are concerns about its limitations due to the lack of available data. At present, Gen-AI’s timeframe to make a significant contribution is lengthy. Not only do large amounts of data remain undigitalised and non-integrated, but a fair proportion is unobtainable. The industry needs to feed Gen-AI, but the nutrients must be high quality. It’s in nobody’s interest for a ‘garbage in, garbage out‘ scenario to prevail.