Adding AI to Revolutionize Supply Chain Management

 

The fine art of supply chain management is all about delivering the right products to customers at the right time and at the lowest total cost. Supply chain managers seek the sweet spot where the costs of manufacturing, storage, delivery, and other factors come together. This convergence creates the best possible business case for their company and customers. Finding this sweet spot involves collecting and analyzing data from the supply and demand workflow.

How AI the Role of Supply Chain Managers 

On the customer side, supply chain managers supervise incoming orders, carry out mid- and long-term forecasting to anticipate future demand, and adjust production accordingly.

On the production side, they monitor their company’s production facilities to get an overview of the capacity of its machinery. They balance customer demand with manufacturing capability to satisfy customer expectations and optimize the use of raw materials and equipment.

On the logistics side, they manage storage and warehousing and calculate the most efficient way to transport goods from the factory to the customer.

In the age of Industry 4.0, everything is highly digitalized. Data from across the complete supply chain continuously flows into complex Enterprise Resource Planning (ERP) systems. AI-powered ERP systems provide supply chain managers with actionable insights for optimizing logistics and manufacturing workflows.

Akkodis AI Solutions for Supple Chain Challenges

Monitoring all these parameters and making the right decision at the right time is a tough challenge. This is especially true in the Life Sciences industry, where supply chain managers face extremely high-quality requirements and time pressure. However, AI-driven logistics strategies help overcome the unique challenges in the Life Sciences supply chain sector.

A global healthcare manufacturer turned to Akkodis for help optimizing its supply chain management. The client manufactures hospital products, ranging from syringes and custom plastic bags containing pharmaceutical solutions to machines for intensive care units. They wanted to better balance customer orders with production capabilities across several sites, storage, and distribution. All these constraints were to be satisfied while minimizing costs.

AI-Powered Tools for Supply Chain Optimization

Mehdi Mounsif, AI Tech Lead at Akkodis in France, heads a team of engineers and researchers working on cutting-edge projects in natural language processing, computer vision, and conversational AI. Analyzing the customer’s supply chain workflow, Mounsif saw that decision-making involved a great deal of manual labor. Additionally, the company’s analyses ran the risk of being biased. Although decisions were built on hard facts supplied by ERP systems, an element of intuition was inevitably involved.

“Unconscious preconceptions can drive people to make flawed analyses, Mounsif said. “Because they’re humans, their approach is not 100% rational and formalized. We saw a clear potential for improvement there. We could give them a decision support tool and a rational methodology to evaluate the data they base their decisions on.”

The Traveling Salesman Problem

The team’s solution is scientifically known as Combinatorial Optimization, a subfield of mathematical optimization that deals with issues often exemplified by the travelling salesman problem (TSP): "Given a list of cities and the distances between each pair of cities, what is the shortest possible route that visits each city exactly once and returns to the origin city?"

Combinatorial Optimization is a powerful technique well suited to solving complex supply chain problems. However, such a system requires highly specialized people, primarily data scientists, who are scarce resources in the workplace.

We are building an architecture that enables our customers to manipulate high-level expertise that is very scarce in the workplace.

Mehdi Mounsif, AI Tech Lead, Akkodis France

Leveraging LLM’s in Supply Chain Management with AI

To address that obstacle, Mounsif’s team created a hybrid system enabling supply chain managers to communicate directly with their Combinatorial Optimization algorithms. The team achieved this by creating large language model (LLM)-based “agents” that can contextualize the requests fed into the system, written in plain language by the supply chain managers.

"We are building an architecture that enables our customers to manipulate high-level expertise that is very scarce in the workplace. This can be done without requiring them to undergo any formal training,” Mounsif said. “And via GenAI, they can bring in their business logic. In this case, it is the prioritization of orders for medical supplies from hospitals, the capabilities of different production sites, transportation and logistics, etc.”

He added: “The GenAI filters the data and angles the tool and its algorithmic logic to give the client the answers they want within their business perspective. ”In this way the Akkodis AI experts brought rationality and formality to a workflow that used to be mostly manual and at risk of flaws through bias and intuition.

Akkodis’ Industry-Agnostic Solution

Customers from diverse sectors are eager to partner with the Akkodis AI team to achieve similar gains in managing their supply chains. Apart from Life Sciences, Mounsif is currently working with a large vehicle tire manufacturer to improve its sales forecasts across different types and sizes. The goal is to reduce discrepancies between forecast and actual sales to optimize production and earnings.

There is potential nearly everywhere, Mounsif said. “We give our customers the power to manipulate data without having to hire data scientists. Our system provides a bridge between humans and highly accurate and deterministic techniques, allowing for tremendous gains in resource management, forecasting, logistics, and much more. All these issues fall under the umbrella of Combinatorial Optimization, and I can safely say that they are present in any industrial sector.”