A recent study (Pre-COVID19) found that the companies lose between 9% and 20% of their value in just 6 months due to supply chain problems. The problems, accentuated by the current pandemic, range from parts shortages, excess finished product inventory, underutilized installed capacity, unnecessary warehousing costs, to unnecessary transportation of supplies and goods.
Due to the obligation of suppliers to keep their information systems and operating advantages in reserve, problems recur more frequently, with no possibility of breaking the inertia.
Artificial Intelligence (AI)
Crises demonstrate that risk assessment and preparation of contingency plans in advance of a drastic change in the Supply Chain can prevent significant losses. To achieve this, it requires proactive strategies and the ability to accurately forecast the probability of occurrence, as well as the impact of risks.
This ability can be achieved using one of the many techniques provided by Artificial Intelligence, until now little used in local industries:
-Mathematical Optimization Techniques.
-Neural Network Approximations, which represent a set of possible states and transitions. Artificial Neural Networks (ANN).
-Methodologies that adopt multi-agent interaction modeling.
Approaches using automated knowledge-based reasoning.
-Machine Learning and Big Data techniques.
-Classification Machine Algorithms, of which SVM (Support Vector Machine) is one of the most popular.
How does it work?
Artificial Intelligence must be enabled for autonomously decide on the course of action that will lead to the successful realization of the objectives of the Procurement areas and to do so in an environment that is partially unknown of the Supply Chain. IA takes fundamentals from mathematical representations, evolutionary computation and statistics.
Data Analytics
Process standardization is the baseline for mitigating the risks inherent in the supply of goods and services. With it we achieve "expected" behaviors of the agents and with data analytics we can measure and understand with ample certainty the impact that each process has and finally, establish if adjustments are required.
With its implementation we can also ensure the accuracy of transactions, that there are no misinterpretations with suppliers, it suggests through interpretation how to handle exceptions and above all it allows us to make strategic decisions that guarantee the continuous improvement of the objectives.
Use Cases
Automated Supplier Selection based on risk identification.
Automated Bid Selection using parameters defined in algorithms of Sorting Machines.
3. Modeled propagation of the risk in the supply network.
4. Identification and Prediction of deceptive practices in the supply chain.
5. Explanation of disasters data oriented.
6. Support end-to-end to facilitate collaborative disruption in management.
Advanced Applied Examples
Identification of critical variables ("Rapid trust", and "Transparent information sharing".) during a disaster event (such as Nepal earthquake in 2015) to achieve a reliable supply chain.
2. The Chinese government analyzes its supply data to train sorting machines. This data is used to determine whether a company has the capacity to produce, demand, supply through external disruptive events.
3. Bayesian predictions are used for determinate the risks of delays in supply chains air.
4. Food sensors in supermarkets that dynamically adjust prices depending on its maturity and temperature profile.
Taking the first step towards the use of data
Although Latin America has so far ventured to take the first initiatives, it requires a courageous management team that believes in the adoption of new technologies to unleash the power of the data that their companies produce every day.
This does not mean that Artificial Intelligence and Data Analytics are the only ones that should be taken into consideration when making strategic decisions, but they should be taken as crucial tools to discover new insights. In this way, the real decision makers can combine them with their own experience and come to optimal resolutions.
Sources:
-Supply chain risk management and artificial intelligence: state of the art and future research directions.
