Automating order classification using AI
A leading manufacturer of flexible packaging films in Central Europe (the "Client") implemented an AI solution for the automatic classification of orders into production compatibility groups. The project, carried out in collaboration with Aspecta, uses AWS services and has improved production planning efficiency, accuracy, and scalability.
13. jún 2025 ┃ 6 minút čítania
Challenge
Before automation, manufacturers faced increasing product variability and increasingly demanding production planning. Each order contained a number of parameters that had to be correctly assigned to avoid downtime and reduced efficiency. In an environment with a high volume of orders, even minor inaccuracies resulted in significant losses of time and capacity.
Increasing portfolio complexity: hundreds of products defined by combinations of layers, materials, and print types required manual classification.
Dependence on experts: the process of assigning orders to production compatibility groups depended on senior planners.
Low scalability: expanding the product range threatened to slow down production and increase costs.
Risk of error: Manual assignment led to inconsistencies and delays.
Solution
Aspecta designed an AI classifier on AWS (SageMaker, Lambda, Glue, S3, API Gateway). This solution was chosen over a traditional rule-based system or on-premise application because it enables rapid model training, scalability, and easy integration into existing client processes. The AWS cloud architecture minimizes maintenance costs and enables continuous model improvement through a feedback loop. The model uses historical order data, layer parameters, print types, and materials to automatically determine the correct production compatibility group.
Main components of the solution:
Dátová vrstva: centralizované úložisko (S3), ETL pipeline (Glue / Lambda).
Modeling: SageMaker with XGBoost / LightGBM / CatBoost, metrics: accuracy, F1, confusion matrix.
Deployment: SageMaker Endpoint or Lambda + API Gateway for real-time predictions.
Integration: API connected to ERP/PLM.
Monitoring: CloudWatch, SageMaker Model Monitor, feedback loop.
Security: IAM, encryption, audit logs.
Results and benefits
The implementation of the solution has brought measurable results that confirm its practical benefits. Key performance indicators show not only improved accuracy and speed, but also a fundamental shift in the efficiency of the entire planning process.
Classification accuracy: 96.2% accuracy (target ≥ 95%).
Processing time: from 5–10 minutes to ~7 seconds.
Operational impact: faster planning, fewer errors, more consistent assignment.
Scalability: no need for manual rule updates.
“Automated decision-making has accelerated our planning and reduced dependence on expert know-how.”
— Head of Planning Process
Conclusion
The project confirmed that a properly designed AI model can fundamentally change the way manufacturing responds to the growing complexity of its portfolio and the demand for quick decisions. By combining expertise, AWS cloud infrastructure, and integration with ERP systems, we created a solution that is not only accurate and stable, but also sustainable in terms of operation.
Are you considering how AI can increase the efficiency of your decision-making processes and planning? Contact us for a free introductory consultation.
Subscribe for newsletter
Share article
