In the digital age, the leaders of the aeronautics sector are opting to transform themselves at any price in order to face the competition increasingly fierce. In many companies, the transformations remain superficial, not changing the daily lives of the vast majority of users - including their productivity. In large companies, specialized teams "Big Data", teams specializing in "Machine Learning", or the "Internet of Things", etc. Although competent, these teams are cut off from the rest of the company, both at the organizational level and especially at the cultural level. They are confined to producing "Proof of Concepts" or PoC but which are not operational at the operational level. It is considerably more difficult to produce value for the company, in production and at scale. That is, moving from the stage of the beautiful demonstration to an operational stage having a real impact on the company's revenues.
The aeronautical industry does not escape this observation. This industry produces aircraft in very small quantities (10000th aircraft produced by Airbus in 2016). Producing a plane is a very unachievable task that is often a matter of craftsmanship: the complexity lies in the immense quantity of systems to integrate, test, and follow in their life cycle. The stakes are real and the consequences of an anomaly can be catastrophic humanly and financially, as Boeing is currently paying the price.
But the small number of aircraft, very complex, products does not lend itself well to conventional Big Data and Artificial Intelligence algorithms, which were invented in the first place to exploit the highly structured data of billions of users. Critical information is often hidden in disparate, heterogeneous and unstructured data, such as office documents, or reports written by technicians.
The real IT challenge facing manufacturers is to manage this huge complexity, relying on all their users, not just a few handfuls of digital natives. This requires the implementation of modern applications based on advanced technologies, such as natural language processing (NLP) and machine learning (ML).
Thus, we can trace the history of decisions and design choices made ten years ago, or bring out a significant risk expressed by a driver hidden among a thousand other reports.
This requires leaders to rethink their digital transformation strategy.
Source : www.actuia.com of 1er juillet 2019
#transformationdigitale #usineconnectee #bigdata #digitalisation