A Global Manufacturer
Industry
Manufacturing
Year
2025
Stack
Axium, Azure Cloud
⸻ Business Impact
Accelerating Manufacturing Performance with Axium
Faster time-to-production, lower planning overhead, and scalable automation for made-to-order manufacturing
We recently worked with a global industrial manufacturer that builds customized products on a make-to-order basis. Their production model demands precision and agility; every order must be configured, packaged, and scheduled with minimal room for error.
MobiLab enabled the company automate two of its most expert-intensive processes: production line configuration and weekly production smoothing. By combining Axium’s Semantic AI with logic-based automation and targeted machine learning, we delivered intelligent automation that outperformed the previous setup both in accuracy and scalability.
The result is a smarter, leaner production operation:
Accelerate AI ROI, with a solution ready to scale to new products, materials, and rules.
75% reduction in manual configuration workload, enabling experts to focus on exceptions.
95%+ optimization in weekly production planning, reducing spikes and underuse without shifting orders across weeks.
A future-ready path to decommission costly legacy solutions.
Freed expert capacity to focus on high-value tasks
MobiLab’s solution enabled the company reduce operational friction, improve throughput, and lay the foundation for broader smart factory automation.
⸻ STARTING POINT
Manual Planning and Hidden Logic Slowing Growth
The company operates in a just-in-time, made-to-order production model, building customizable components for industrial use with little to no warehousing. Each order must be configured precisely before production, considering packaging, material layout, and sequencing constraints.
However, two critical challenges slowed operations:
01. Order configuration: ERP recommendations for packaging often need correction. Experts manually determined configurations and material placements, which created delays, inconsistent outcomes, and FTE bottlenecks.
02. Weekly production smoothing: To optimize production lines, product types needed to be balanced across the workweek. While daily planning was automated, weekly smoothing to prevent spikes and underutilization was still performed manually with a rule-based system, outside of the corporate data platform.
⸻ SETTING UP THE FRAMEWORK
Faster Results with Structured Knowledge and Semantic AI
We started by working with domain experts to identify the business logic driving order configuration and production balancing. Instead of replacing it all, we mapped it into a semantic layer using Axium.
- For order configuration, we identified which attributes followed deterministic patterns – ideal for logic-based implementation – and where variability warranted machine learning support.
- For weekly planning, we designed a lightweight scheduling algorithm that shifted only flexible orders to balance product mix across the workweek, preserving integrity for time-critical orders.
Axium tied everything together. It connected expert rules, historical data, and predictions into one shared layer. The result: automation that was accurate, transparent, and quick to deploy. Data preparation and algorithm deployment were done within the customer’s Azure environment using familiar tooling, ensuring fast adoption, cost transparency, and low integration overhead.
⸻ SOLUTION
Automated Order Configuration and Weekly Production Planning
We delivered two main solutions:
1. Order Configuration Automation
We automated the logic for assigning packaging and positioning based on order characteristics and implemented
- Encapsulated rules in Axium’s modeling environment
- Trained a Random Forest model to predict the configuration for edge cases
- Applied confidence-based routing to decide when to use logic vs. ML
- Added full traceability and fallback logic
The outcome: 75% of orders were configured automatically, dramatically reducing reliance on manual input. For major product classes, accuracy exceeded 98.5%.
2. Weekly Production Smoothing
We built a custom planning algorithm that:
- Filters and locks fixed-date orders
- Calculates optimal distribution targets
- Minimizes product-type variance across the week
- Preserves schedule integrity (no order shifting between weeks)
The outcome: 95%+ smoothing accuracy, significantly improving operational consistency without increasing planning complexity.
Both components are fully integrated into the customer’s SAP environment and require no change in interface or workflow for end users.
⸻ Conclusion
Scaling Smart Operations with Axium Semantic AI
This project delivered a clear shift from manual planning to smart, explainable automation without losing control.
By combining logic-driven rules with predictive modeling both grounded in Semantic AI, we helped the customer move from manual production planning to a system that automates and adapts. With Axium as the foundation and a practical approach to AI, we accelerated time-to-value without introducing unnecessary complexity.
Tangible business benefits:
➞ Faster to Market
Automated order configuration and weekly planning shortened lead times and improved responsiveness.
➞ No Overreliance on SAP or Legacy Systems
Logic and intelligence were decoupled from rigid systems, enabling flexibility and maintainability.
➞ Reduced Warehouse & Processing Costs
Better planning accuracy reduced overproduction, rework, and unnecessary storage.
➞ Scalable, Guardrail-Based Automation
Rules and models handle routine planning, with transparency and flexibility for business users to adjust.
Start small and scale.
Axium is designed future-proof, for all businesses to organize their knowledge and leverage the power of data and actionable business insights.
Let's go forward. Together.
Every business is unique. Let us show you how Axium can accelerate your data journey. Get in touch with us today!