The company, the second world’s largest aluminum company, founded in 2007, was a pioneer of innovation in the mining & metal industry from the very beginning. The company aimed to lower its carbon footprint, so implementing innovative and energy-saving technologies was crucial.
However, the green-oriented approach was not limited just to producing metal but also a logistic part of operations. Read about how the company managed to automate logistic workflows and increase efficiency thanks to implementing AI.
Case Study Shortcut
Challenge
Manual and Fragmented Operations Across Departments
The company relied on spreadsheets, SAP ERP systems, phone calls, and emails for managing shipments and inventory. This manual, disjointed system resulted in inefficient communication, data duplication, and human error, making the process slow and prone to disruptions.
Poor Data Quality and Inconsistency
Data was stored in Excel files, flat files, and scattered across multiple sources, leading to poor-quality, incomplete, or outdated data. Lack of a unified view hindered real-time decision-making and prevented predictive planning.
Inability to Predict Disruptions in a Complex, Global Supply Chain
The multi-modal logistics process, involving sea and land transport, was vulnerable to external risks like geopolitical shifts and weather conditions. Without AI-based forecasting, the company couldn’t anticipate vessel delays or market fluctuations, leading to demurrage costs, missed deadlines, and increased CO₂ emissions.
Goal
In order to be able to make decisions based on real factors, not assumptions; processes needed to be automated, scalable, repetitive, and bulletproof, so manual labor had to be reduced to a minimum.
- Unified Platform for Managing Supply Chain
- Predictive AI Modules
- Decision-making is based on intuition, not on data
- Decreasing logistics costs and losses
- Reducing the amount of manual data processing
- Adding scalability and automation processes
Outcome
After implementing the AI platform developed by Addepto, Company team was able to stay abreast of the live status of their inbound major raw material vessels. This always-up-to-date knowledge allows the company increases the efficiency of inventory planning, reduce/avoid demurrage, and – lastly – increase the general operational performance.
Thanks to automation, managers were able to quickly make the most cost-effective decisions, and so logistic expenses were reduced (transportation and stock costs were reduced by an average of $1.5 per ton) and the number of customer orders completed on time increased. The overall order handling process got faster and the delivery process was released from unexpected disruption.
Before
- Manually data processing
- Slow and unscalable information exchange
- Processes sensitive to human mistakes
- Huge logistics costs, losses, and high gas emissions
- Inability to predict events and possible disruptions
After
- Automated data processing
- Unified data environment (one source of truth)
- Forecasting potential disruptions based on historical data
- Reducing logistics costs, losses
- Lowering carbon footprint
Integrate those solutions in your company
Case Study Details
Approach
Unified AI-Driven Supply Chain Platform
- Developed a centralized platform that integrates shipping schedules, inventory data, and order management into a single source of truth.
- Enabled the company to plan across multiple transport types (vessels, trains, trucks) while consolidating siloed data into one scalable system.
Predictive Modeling for ETA and Market Pricing
- Implemented machine learning models to forecast vessel ETAs based on historical patterns, real-time updates, and third-party APIs.
- Built predictive algorithms to estimate market vessel prices, considering macroeconomic indicators and seasonal trends, optimizing procurement costs.
Automated Logistics Planning and Optimization
- Designed an AI engine that evaluates multiple constraints: loading rates, sailing times, fuel consumption, deadlines, incoterms, and customer preferences.
- The system automatically groups orders into batches, determines the optimal dispatch date, and updates plans daily if new data arrives (e.g., from SAP).
Integration of Internal and External Data Sources
- Seamlessly connected internal ERP data with third-party services for ship tracking and weather insights.
- Allowed for real-time visibility and adjustments, helping prevent delays and reduce unnecessary costs.
Scalable and Future-Ready Architecture
- The platform is designed to scale across regions and terminals, accommodating more products and routes over time.
- It enables continuous updates and improvements without disrupting the core logistics flow.
Our Team Expert Opinion
The project was very challenging not only from a technology perspective but from a business as well. To build a proper optimization system we had to gather very detailed requirements and understand each part of the company supply chain. Thanks to the close collaboration with the customer and its team we were to achieve the best results in optimization system and supply chain optimization that brings real money benefits to the business.
Edwin Lisowski Co-founder – Addepto
The logistics of homogeneous marine cargo loading and transportation due to the high complexity of an optimization problem is problematic. Still, the potential for reducing costs is enormous. The manual data processing was time demanding, the results inaccurate, and costly. Finding the optimal solution is impossible not only because of high computational complexity but simply due to unrealistic customer expectations. We are incredibly pleased with the solution found by Addepto. Heuristics implemented for us makes it possible to get the satisfactory assignment of orders to vessels and vessels to piers in loading and unloading ports. Thanks to this solution, we have a lot of savings both on the costs of loading, storage, and transport, but also on the work of our employees.
Michał Bernardelli Data Scientist – Addepto
