Highlighting AI applied to the supply chain
In an increasingly complex world, AI can produce analyses and projections fine enough to support organizations in their quest for agility. Back to the basics of the technology and its benefits for the supply chain.
Sales forecasting, inventory management, planning, scheduling... How to anticipate the vital functions of the supply chain in an unstable environment and in the face of demand that is becoming volatile? The answers include artificial intelligence and its most well-known branches, machine learning and deep learning. But how do they work? What use cases can be identified? And which companies exist on the market to serve these needs? To answer these vast questions, four experts from the sector took turns on September 30 during a webinar organized jointly by the Lab Digital de France Supply Chain and Wavestone.
Demystifying artificial intelligence
and understand the data

To demystify the subject, we must first understand it. So, what is AI? According to Ghislain de Pierrefeu, partner in charge of the Machine Learning Data Lab at Wavestone, it is the ability of a machine to perform complex intellectual tasks previously specific to humans. Machine learning, an application of AI, gives computer systems the ability to make decisions based on learned data. Deep learning, on the other hand, has the ability to imitate the functioning of the human brain in the processing of data and the creation of models. This trio is intertwined with the fields of data science and Big Data Analytics. For Ghislain de Pierrefeu, once these concepts have been integrated, the question must be asked: "Do I have data and how canI use it? »
Because the data can be of different types and AI algorithms too. The latter can be divided into two categories. The first, supervised learning, which consists of developing a predictive model based on input data and results. The second, non-supervised learning, is based on input data, divided into subgroups considered homogeneous: "The idea here is to use the data to build clusters and analyse the orientations found, using common human sense and understanding of the profession," explains Ghislain de Pierrefeu.
From exploration to exploitation
of AI for the Supply Chain

From a practical point of view, these AI algorithms have a positive impact on supply chain management: on data, demand forecasts, sales strategy, but also on stock replenishment, supervision and forecasting. Ivan Baturone, Michelin's Supply Chain Innovation Manager, gave a concrete presentation on the development of SAAM (Stock Analysis & Alerting Machine), a tool that "deals with the tsunami of supply chain data for the distribution of our products from our factories to our commercial warehouses. We have developed machine learning algorithms to help us detect three weeks before an item is out of stock at a distribution point. Ultimately, the Holy Grail would be to achieve a self-analysing and self-regulating supply chain. Thus, over 2019, thanks to SAAM, Michelin has notably increased its "product" availability by 7 points.
A radar, a panorama and exchanges
around AI applied to the Supply Chain

To reach this level of maturity, to understand AI and its benefits, many solutions are available on the market. The radar established by Wavestone and France Supply Chain was created to give an overview of these. At the same time, the France Supply Chain Digital Lab, the originator of this webinar, will very soon publish its panorama of digitalization 2020/2021.
There is still time to complete the survey on the association's website before the results are published on November 17 at the Supply Chain Event.
Two webinars will also be held, the first in November on RPA (Robotic Process Automation) with the participation of Michelin and the second in December, led by Daher on Supply Chain Innovation.



















