Suzuki Garphyttan is a global industrial company with roots in the Swedish steel industry. The company, owned by Japanese Nippon Steel, manufactures advanced wire products used in the automotive and engineering sectors. Production takes place in Sweden, the United States, Mexico, England and China, with headquarters and the largest factory in Garphyttan outside Örebro.
Historically, Suzuki Garphyttan has been strongly associated with valve spring wire for internal combustion engines – a product that has long been the company's most profitable. But as electric cars take greater market share, demand for these components is declining. At the same time, the need to find new markets and applications is increasing, placing new demands on innovation, efficiency and customer focus.
"In the past, we could rely on having the best product in the world. Now we must also be the best supplier. This requires us to work more data-driven, with better control and insight into our processes," says Andreas Rendius, Digitalization Manager at Suzuki Garphyttan.
The company management formulated two clear goals: process optimization and increased customer focus. To achieve these goals, a modern data architecture was required – one that could bring together both business data and machine data and transform them into valuable insights in real time.
Suzuki Garphyttan chose to explore Microsoft Fabric as the platform for its new data environment. The choice was natural as the company was already using Microsoft Azure extensively, which provided good integration opportunities. To get started quickly, a POC (Proof of Concept) was carried out together with Random Forest, on the recommendation of Microsoft.
The purpose of the POC was to test how Fabric could be used for both business data and OT (Operational Technology) data – process data directly from the machines. The team wanted to see if it could stream data in real time, make calculations, identify anomalies, and even send control signals back to the production equipment.
“We wanted a platform where we can run machines more efficiently – perhaps even unmanned in the long term – while also giving operators better tools," explains Rendius.
From a technical perspective, the solution was built on the company’s existing Ignition platform for collecting data from machines, PLC systems, and wireless sensors – around 200 machines globally. From there, data is sent to Fabric, where it is combined with business data from the company’s cloud-based ERP system. With support from Random Forest, real-time dashboards were also set up in Fabric, enabling analysis and visualization for different user groups.
"Suzuki Garphyttan has really dared to invest in innovation. By connecting OT data from the factory floor with business data in Fabric, we have been able to create a solid foundation for faster and smarter decisions in real time," says Alexander Wittlock, BI consultant at Random Forest.
During the project, challenges arose that quickly yielded important lessons. For example, data governance and naming standards proved to be crucial. Because the same sensor data could be called "instant value" in one machine and "actual value" in another, it became difficult to consolidate the information.
“We thought we could address governance issues later, in parallel with development – but that didn’t work. You need a clear data standard from the start, otherwise you end up having to redo everything,” Rendius notes.
Another insight was that data quality and availability varied widely. Many important data were held by operators and not recorded digitally. To solve this, Suzuki Garphyttan developed a manual data collection application – where operators record, for example, lubricant usage in real time. In the long term, automation is also planned via scales and sensors that record material consumption directly.
The results of the POC project clearly show the potential of Fabric and the new way of working. Suzuki Garphyttan now has a unified platform for both business and process data, where insights can be used to optimize production, improve quality and streamline maintenance work.
“For example, we were able to build an application where engineers and operators can monitor machine temperatures in real time and receive notifications when adjustments are needed. That’s a concrete step toward more proactive and automated operations,” says Rendius.
The project has also helped to strengthen the internal data culture. Through workshops and training, many employees have been engaged and contributed ideas on how data can create value in everyday life. At the same time, the company has established domain ownership and clearer governance models – and moved automation engineers into the same team as data engineers, simplifying collaboration between IT and production.
Although the journey has only just begun, Suzuki Garphyttan has already seen how modern data management can drive innovation. The Fabric platform provides the opportunity to scale and build towards more advanced analytics, AI and machine learning as data quality improves.
"This project has shown us the importance of combining technology, processes and people. Random Forest has been a strong support in both technology selection and implementation. We are now ready for the next step in our digital journey – with data as our most important raw material," Rendius concludes.
Listen to Andreas Rendius talk about the collaboration at the Data Day 2025: