AI Insights

Using AI to Make Supply Chains More Sustainable While Also Saving Costs

April 26, 2024

article featured image

To meet the ever growing needs of businesses and consumers around the world, International Supply Chains have maintained a spirit of constant innovation and progress. In the modern day, Supply Chain Management is a science that is dedicated to achieving maximum efficiency while handling a wide and complex network of hubs and destinations around the world.

However, the area of innovation that demands immediate and serious consideration in recent times is the controlling of carbon emissions. International Supply Chains require a mind-boggling number of ships, trains, planes, trucks, and other vehicles running near constantly. This and other activities lead to Supply Chain Management Companies having quite sizable carbon footprints.

In this article, we explore how a global supply chain management company partnered with Omdena to create an AI-powered solution to reduce carbon footprint in supply chains, leading to a 10% reduction in emissions and $5M in annual savings.

The Background

Freight transportation alone is responsible for around 8% of the world’s total annual greenhouse gas emissions. 

If we include emissions caused at warehouses and ports, this figure rises to 11%. This makes Logistics and Supply Chain Management one of the industries with the highest carbon footprints.

However, these activities occur on a global scale. Billions of tons of cargo gets shipped every year and tracking exactly where carbon emissions are highest is a grand undertaking. Also with the level of time sensitivity that these companies have to function within means that launching large scale initiatives to address this issue could result in a high financial impact.

The Goal

A global supply chain management company was facing the challenge of reducing its carbon footprint. The company had a complex supply chain with operations in over 100 countries. This made it difficult to track and measure the company’s carbon emissions, and to identify opportunities to reduce them.

Challenges to Overcome

Making the perfect Dataset

The dataset required to achieve our goal needed to account for a massive amount of shipping data and also the myriad of ways this shipping was being done. This meant that we had to strike a balance in the dataset between scale and data diversity

Building Trust

In the high pressure environment that is international supply chain management, adoption of new technology requires managers to trust it. One key way to form this trust is to offer transparency in how the system works. This meant that the system had to be built to be interpretable.

Ease of Use

Considering the wide geographic and hence cultural diversity of the people who would have to use this tool, ease of use and accessibility had to be taken as priority.

Our Approach

The company partnered with Omdena to develop an AI-powered solution to reduce its carbon footprint. Omdena’s team of data scientists and machine learning engineers developed a solution that uses machine learning to analyze the company’s supply chain data and identify opportunities to reduce carbon emissions. The solution also develops tailored recommendations for the company’s supply chain managers on how to implement these opportunities.

Machine Learning Models Used:

Random Forests

This method builds multiple decision trees during training to predict outcomes. It’s effective for analyzing complex relationships in supply chain data, aiding in understanding factors affecting carbon emissions.

Gradient Boosting Machines (GBMs)

GBMs correct errors made by previous trees to improve accuracy. They predict carbon emissions based on supply chain activities like transportation methods and fuel types, offering precise insights into emission levels.

Clustering Algorithms (e.g., K-means)

These algorithms identify groups with similar emission characteristics in supply chain data. By pinpointing high-emission areas, they facilitate targeted interventions to reduce carbon footprints.

Neural Networks

Deep learning techniques like CNNs and RNNs forecast emission trends by analyzing historical and external data. They excel in capturing temporal and spatial patterns, aiding in comprehensive emission analysis.

Development of Tailored Recommendations

The AI solution doesn’t just stop at identifying opportunities; it goes further by developing tailored recommendations for supply chain managers on how to implement these opportunities efficiently. Here’s how it’s achieved:

Feature Importance Analysis

The solution assesses the most influential factors affecting carbon emissions by analyzing feature importance from models like Random Forests and GBMs. This prioritizes recommendations based on their potential impact on emissions reduction.

Scenario Simulation

Predictive models enable the system to simulate different interventions’ effects, such as changing transportation modes or optimizing routes for efficiency. These simulations guide specific recommendations for significant emissions reductions.

Optimization Algorithms

Linear programming and other optimization methods identify the most efficient changes in supply chain operations to reduce emissions. This includes optimizing logistics routes, inventory levels, and production schedules for sustainability.

User-Friendly Interface

Recommendations are presented through an intuitive interface, ensuring accessibility and actionable insights for supply chain managers. Visualizations illustrate potential impacts, accompanied by step-by-step guides for effective implementation.

By combining these sophisticated machine learning models with a user-centric approach for developing recommendations, the solution empowers supply chain managers to make data-driven decisions that significantly reduce carbon emissions while maintaining operational efficiency.


Since implementing the solution, the company has reduced its carbon emissions by 10% and saved $5 million in annual costs.

Not only did we achieve our primary goal of reducing the company’s overall carbon footprint, we went a step further and designed the system in a way that it also saves on costs. This demonstrates how steps towards sustainability in the private sector do not necessarily have to come at the expense of higher costs. With technology like AI, both these goals can be met simultaneously.

Innovating Further in Real World Scenarios

Optimizing Delivery Routes in Latin America Using AI Planning 

In a different project, Omdena partnered with Carryt, a technology company focused on improving logistics through artificial intelligence and route planning, to tackle urban congestion and enhance last-mile logistics efficiency. 

Through an 8-week challenge, 50 AI engineers collaborated to develop a route optimization tool leveraging Google’s Operations Research (OR) Tools and Open Street Map. This tool enabled Carryt to visualize optimal delivery routes in Bogota, with potential applicability in other congested cities worldwide. 

The project aimed to optimize the deployment of vehicles and drivers, thus improving customer service levels, reducing carbon emissions by minimizing unnecessary travel, increasing drivers’ well-being, and alleviating the overall impact of the logistics industry on urban congestion. 

This initiative reflects an effort to address the significant challenges posed by rising e-commerce demand and urban population growth, which contribute to increased carbon footprint and congestion, especially in cities like Bogota, Lima, Mexico City, and Rio de Janeiro.

Why use AI for SCM and Logistics?

The AI-powered solution to reduce carbon footprint in supply chains has provided a wide array of benefits to the company, which extend beyond immediate carbon emission reduction and cost savings. These benefits include:

Improved Fuel Efficiency

The solution offers insights into logistics and transportation operations, enabling the company to implement strategies that reduce fuel consumption, such as route optimization and transitioning to fuel-efficient vehicles.

Reduced Waste

By analyzing production and distribution processes, the AI solution helps in minimizing waste production at various stages of the supply chain, including excess inventory, packaging materials, and product spoilage.

Increased Use of Renewable Energy

The system identifies opportunities for integrating renewable energy sources into the supply chain operations, such as solar-powered warehouses or green logistics solutions, further reducing the carbon footprint.

Enhanced Supply Chain Visibility

The AI solution provides detailed insights into every aspect of the supply chain, offering unprecedented visibility that helps in identifying bottlenecks and inefficiencies.

Data-Driven Decision Making

With comprehensive analytics at their disposal, supply chain managers can make informed decisions that are backed by data, leading to more strategic and effective management practices.

Compliance and Reporting

The solution aids in tracking regulatory compliance related to environmental standards and simplifies the reporting process for carbon emissions and sustainability efforts, essential for transparency and accountability.

Increased Customer Satisfaction

By adopting greener practices and demonstrating a commitment to sustainability, the company can enhance its brand image and appeal to environmentally conscious consumers.

Scalability and Flexibility

The AI-powered solution is designed to scale with the company’s operations and adapt to changes in the supply chain, ensuring long-term sustainability and efficiency improvements.

These benefits collectively contribute not only to the environmental objectives of the company but also enhance operational efficiency, cost-effectiveness, and market competitiveness, underscoring the transformative potential of AI in supply chain management

Other Applications of This Model


Our system can analyze manufacturing processes to identify opportunities for reducing carbon emissions by optimizing energy usage, material consumption, and production efficiency.

Energy Production

The model optimizes energy generation processes to reduce carbon emissions from power plants and renewable energy sources, supporting the transition to cleaner energy production methods.

Agriculture and Farming

Our system can analyze agricultural practices to minimize carbon emissions from farming activities, such as crop cultivation, livestock management, and fertilizer usage, while promoting sustainable farming methods.

Technology and IT Services

Our model can help technology companies optimize data centers and IT infrastructure to minimize energy consumption and carbon emissions, supporting the transition to greener computing practices.

Construction and Real Estate

The system can evaluate construction materials, building designs, and energy usage in real estate projects to minimize carbon emissions throughout the construction process and during the lifecycle of buildings.

Want to work with us too?

media card
How We Leveraged Advanced Data Science and AI to Make Farms Greener
media card
Environmental Sustainability and AI: A Synergistic Approach for a Greener Future for Companies
media card
AI-Powered Solutions for the European Sustainability Reporting Standards (ESRS)
media card
AI-Driven Sustainability Solutions in a Changing World