Plant capacity without hardware investment
Ingredion needed data-driven predictions for optimal response times. Precise forecasts for all production conditions increased capacity by up to 10% in six months.
- Customer
- Ingredion
- Industry
- Food & process industry
- Project duration
- 6 months
Response times in production were controlled by employees based on experience, laboratory analyzes and Excel data. The manual control tended to react defensively - the product should definitely not react too quickly. This slowed down plant capacity and drove up energy consumption.
We developed an AI-powered assistance system that predicts the optimal response times for each batch. To do this, we integrated process data, laboratory values and historical batch data into a common model and implemented it directly on the system as a decision-making aid.
- Precise predictions of optimal response times across all production conditions
- Increase system capacity by up to 10% without new hardware
- Reduction of specific energy consumption through shorter, optimal response times
- Relieving the burden on system operators through data-based recommendations
Deliverables
+10%
more system capacity
6 months
Time-to-value
reduced
Energy consumption
More success stories
AI guided repair assistant
An MVP for guided repair plus automatic quoting in just two months. The assistant suggests the procedure to technicians directly in the service software with 90% accuracy.
Data lakehouse as a central intelligence platform
SAP, Salesforce, production and sensor data transferred to Databricks — a central data platform for AI, BI and GenAI use cases of many data teams.
Next-Best-Action Marketing im Telko-Konzern
Data science-supported control of call centers and sales activities: 20% more call acceptance and, depending on the product, 2–15% more sales.
KI-getriebene Glasfaser-Gebietsausbauplanung
A planning tool that identifies optimal expansion areas and predicts expansion costs and sales forecasts in a data-driven and explainable manner.