Case study: Berentzen boosts bottling line efficiency with digital traceability and AI-driven production planning
Key Highlights
- Berentzen faced production losses due to lack of real-time visibility and manual data collection, impacting quality and profitability.
- The digital ecosystem integrated machines with SAP BTP, enabling centralized data access, KPI monitoring, and analytics for better operational insights.
- Implementation of the Paperless Production App streamlined workflows, improved traceability, and reduced errors caused by manual documentation.
- AI Order Sequencing optimized production planning, reducing changeover times and increasing line utilization by approximately 15%.
- Operator adoption was supported through targeted training and communication.
For beverage manufacturers managing complex packaging operations, delayed visibility into production losses can erode throughput and quality slowly. It’s difficult to notice day-to-day, until efficiency has dropped and no one knows why. Slow drift production losses don't appear all at once. They accumulate quietly until profitability starts to disappear. At the Minden, Germany bottling facility for Berentzen Group, a lack of transparency into the production line was making managing multiple packaging variants, smaller batch sizes, and increasingly complex production requirements even more challenging.
“The key trigger was missing line transparency in bottling, which led to productivity losses,” says Thomas Faust, head of operations at Berentzen. “If OEE drivers such as downtime, micro-stops, speed losses, and scrap are not captured cleanly and in a timely manner, problems are recognized too late and countermeasures take effect with delay.”
Berentzen Group operates across three core business segments: spirits, non-alcoholic beverages, and equipment for freshly pressed juice. Its operations include spirits bottling in Minden, Germany; non-alcoholic beverage production in Haselünne, Germany; and a central warehouse in Stadthagen connected to the Minden plant by rail. Through its Fresh Juice Systems equipment line (in Citrocasa in Linz, Austria), it also serves food service operators such as hotels and restaurants. Overall, the company bottles product for retailers, hospitality businesses, beverage and foodservice operators, and international distribution partners in both the branded and private-label markets.
Output was decreasing and unit costs were increasing, and the lack of visibility into production was affecting planning and traceability. To address these challenges, Berentzen undertook a broad digitalization initiative named Project “Mehrblick,” or “more insight.”
Slow drift production losses don't appear all at once. They accumulate quietly until profitability starts to disappear.
The problem: Limited production visibility and paper-based workflows
Before the project, Berentzen faced multiple operational challenges tied to disconnected data and manual processes. “Gaps in quality control were a major pain point prior to digitalization,” Faust says. “In spirits bottling, critical quality characteristics such as fill level, closure torque, tightness, label position, and batch coding were not always documented seamlessly and system supported.” The cost of poor quality will continue to increase when complaints turn to recalls, which can include penalties and reputational damage, on top of revenue losses.
The manufacturer also faced higher scrap and rework levels and blocked batches. “Without digital batch and process traceability, deviations often required precautionary blocking of larger quantities because affected units could not be clearly identified,” Faust says.
At the same time, production and quality data often arrived too late to prevent extended losses. As a result, issues such as rising scrap, drift in filling valves or capping heads, labeling errors, or temperature-related deviations could continue for extended periods before corrective action was taken.
Manual documentation processes created additional inefficiencies. “Quality checks performed on paper or in Excel tied up personnel, caused transcription errors, and complicated audits,” Faust says. The worst complication from this was when problems did occur, the evidence chain for investigation was weak.
The company also struggled to optimize production across multiple packaging variants and smaller batch sizes. “Without data-based transparency, changeover times, start-up losses, and SKU-specific performance were difficult to optimize,” he says. Ultimately, it needed better data to zero in cost transparency per SKU.
The solution: Building a connected production data ecosystem
For the digitalization effort, Berentzen worked with Process & Data Automation, a Krones subsidiary and certified member of the Control System Integrators Association (CSIA). The system integrator and packaging and bottling machine OEM Krones provided a unified architecture combining production data, analytics, and planning tools built around the SAP Business Technology Platform (SAP BTP).
According to Sabine Stengel, product manager for SAP solutions at Krones, the architecture was designed to integrate with existing enterprise resource planning (ERP) systems, while minimizing ERP customization.
“Most customers already run SAP ERP (ECC or S/4HANA),” Stengel says.
“Basing extensions on SAP BTP allows us to keep ERP clean, fit-to-standard, while delivering innovation and integration on a scalable platform.”
The system also supports modular deployment. “BTP supports event-driven and API-first patterns,” she says. “We aim for configurable building blocks rather than heavy custom code.”
At the production level, the Krones Ecosystem connects machines and line systems into a centralized data environment. “The Krones Ecosystem creates an integrated data network that connects a customer's machines and entire production lines with Krones,” Stengel says.
The cross-over architecture makes production data accessible across the organization. “By combining information technology (IT) and operational technology (OT), the Krones Ecosystem forms the basis for making machine and line data readable and usable,” she says.
A Kronos portal gives customers access to their data dashboards. “The data is pre-processed at machine level by a hardware and software package installed in the machine,” Stengel says. “From there, the data is securely transferred to the cloud.”
Krones also uses edge devices to make better use of the cloud. “We use edge devices on machine level, where we collect and pre-aggregate some of the machine data for more efficient cloud resource utilization,” she says.
Digital traceability and analytics turn production data into operational decisions
A central part of the project involved expanding production visibility through Krones Analytics, which aggregates and contextualizes the machine and line data. According to Stengel, the analytics platform calculates KPIs according to EN-415-11 and Weihenstephan standards. These KPIs include availability, performance, quality, technical availability, overall equipment efficiency, MTBF, MTTR, reject reasons, and actual versus nominal line speeds.
Beyond KPI tracking, the system categorizes downtime events and identifies bottlenecks. “Analytics aggregates, classifies, and categorizes the downtime events, turning fragmented raw machine data into actionable information,” Stengel says. The platform also identifies the machine responsible for line stoppages, including its fault message to help operators prioritize bottlenecks and coordinate maintenance response actions.
The ability to contextualize machine data across lines and facilities has improved operational decision-making. “Machine and line data can be contextualized, allowing comparisons across lines, varieties, or even entire plants,” Stengel says. This data transparency helps both production management and shift leaders make better informed decisions.
The analytics platform quickly uncovered small but recurring production losses, such as delays caused by material infeed or finished goods removal, or machine maintenance or repair needs. “Frequent but short stoppages, one by one, doesn’t impact the line performance significantly,” Stengel says. But in total, the cumulative effect over time does impact line performance.
For Berentzen, the project gave production management measurable performance analysis. “Decisions are no longer based on gut feeling but are data-driven,” Faust says. The company can now identify how individual SKUs, materials, and disturbances affect production losses. “Cause-and-effect relationships are transparent: which SKU causes which losses, which disturbance costs how many minutes, and which material leads to higher scrap,” Faust says.
The first step in the project was the introduction of the Paperless Production App in 2023. “The Paperless Production App has streamlined and digitalized all production relevant workflows – from order management to batch tracing and quality control processes,” Stengel says. Quality data can now be captured directly from connected sensors or entered by operators. All quality information is linked to the associated production order to support traceability and compliance. The system also guides users through completing and fully documenting quality checks and control standards.
Traceability is very important in beverage production, and material validation at Berentzen is supported through scanner-based verification. “Digital traceability is ensured by tracking all materials consumed in a production order together with their batch information,” Stengel says. “A scanner-based material validation process confirms that the correct components are used at the right time.”
Faust says digital audit trails and automated quality records for batch-level traceability have reduced risk overall and shortened response times when production deviations occur.
AI-based sequencing improves production planning
Berentzen is also serving as the pilot customer for Krones’ AI Order Sequencing application, which is designed to optimize production order planning. “We are now in the process of implementing AI Order Sequencing at Berentzen, where the company is acting as our pilot customer,” Stengel says. The project will optimize production order sequencing with the aid of AI. “The goal is to minimize changeover and cleaning times, improve line utilization, and ultimately increase overall efficiency,” she adds.
According to Krones, the AI-based planning system does not rely on trained machine learning models. “The knowledge-based AI does not need to be trained as it is working with predefined criteria that can be combined and used as restriction or optimization parameters to determine the best possible production sequence,” Stengel says. So far, the pilot test has shown that improved planning quality increased performance by about 15%.
Faust says the additional production transparency has made scheduling more realistic. “When real line speeds and real changeover times per format are available, detailed planning becomes more realistic,” he says. “This reduces rush orders, overtime, and replanning.”
Driving workforce adoption and measurable performance gains
Berentzen also focused heavily on operator adoption during the rollout process. “On the shop floor, acceptance increased noticeably once employees experienced that the digital solutions provide real relief in daily work,” Faust says.
10 lessons learned: What manufacturers can apply from Berentzen's digitalization project
1. Start with visibility, not AI.
Berentzen's first challenge was limited insight into line production losses. Capturing reliable data on downtime, micro-stops, scrap, and speed losses created the foundation for later analytics and AI initiatives.
2. Digitalize quality records early.
Replacing paper-based quality checks improved traceability, reduced transcription errors, simplified audits, and accelerated investigations when production deviations occurred.
3. Connect production and business data.
Integrating shop-floor information with planning and ERP systems helped production, quality, and supply chain teams work from the same data set.
4. Focus on hidden losses.
Analytics revealed numerous short-duration stoppages that individually appeared insignificant but collectively reduced line performance. Small losses often represent large improvement opportunities.
5. Build traceability into material handling.
Scanner-based material validation and batch-level tracking reduced risk, improved compliance, and helped isolate issues without blocking larger quantities of product.
6. Use data to improve scheduling accuracy.
Accurate line speeds, changeover durations, and SKU-specific performance data enabled more realistic production plans and reduced replanning, rush orders, and overtime.
7. Keep ERP customization to a minimum.
Using a platform approach allowed new capabilities to be added without extensive ERP modifications, reducing complexity and supporting future scalability.
8. Prioritize operator adoption.
Short, task-based training sessions and support at the workstation helped employees adapt to new workflows faster than classroom-style training alone.
9. Show employees what's in it for them.
Reduced paperwork, easier access to work instructions, and fewer manual transactions helped build acceptance by demonstrating practical day-to-day benefits.
10. Treat digitalization as a continuous improvement program.
Berentzen plans to expand the solution while refining the system based on user feedback, reinforcing that digital transformation is an ongoing operational journey rather than a one-time deployment.
Planners at Berentzen adapted quickly because the system improved scheduling and coordination visibility from the start, Faust says. However, operators initially required more support as they adjusted to new devices and workflows. “What worked best were short, repeatable trainings closely aligned with real work steps,” Faust says. “It was particularly effective to provide support directly at the workplace.” This was even more important in the initial shifts after the upgrade roll-out to build worker confidence.
Berentzen tried to communicate to workers and frame the upgrade in the ways it would improve and support their work, such as less paperwork and fewer errors. “A key principle was not to overwhelm people, but to build confidence through small steps repetition, clear instructions, and an environment where questions are explicitly welcomed,” he says.
The project has already produced measurable operational improvements at Berentzen’s bottling lines at the pilot facility, improving OEE time losses, changeover duration per SKU, and the share of unplanned downtime.
Berentzen has seen “less scrap and rework through systematic limit monitoring, for example, fill level, closure torque, label position, and lot code readability,” Faust says. Other improvements were driven by real-time feedback from the line, such as material consumption and recorded disturbance causes.
The system also improved visibility into production losses and material yields, which led to “higher yields through accurate recording of spirits losses, product changeover losses, flushing volumes, leakages, and overfilling,” Faust says.
Berentzen reports line efficiency improvements of 13% to 15% with AI-based planning. Inventory levels have also been reduced, including both safety stock and stock shortages for raw and packaging materials.
The company also expects labor savings through the automation of manual transactions and documentation workflows. Faust says the digitalization of manual processes will reduce the workforce needed to do live postings of goods receipt, consumption, and auxiliary material inventory transactions.
Following the Minden deployment, Berentzen plans to extend the solution to additional facilities. “By leveraging the insights our teams gain in their daily work, we’ll further refine both the system and the user experience,” Faust says.
About the Author
Anna Townshend
managing editor
Anna Townshend has been a journalist and editor for almost 20 years. She joined Control Design and Plant Services as managing editor in June 2020. Previously, for more than 10 years, she was the editor of Marina Dock Age and International Dredging Review. In addition to writing and editing thousands of articles in her career, she has been an active speaker on industry panels and presentations, as well as host for the Tool Belt and Control Intelligence podcasts. Email her at [email protected].