Based on documented results at Södra Värö, Södra has decided to scale Viking Analytics’ AI platform, MultiViz, across the entire group. The framework agreement covers 3,650 sensors deployed across the mills in Värö, Mörrum, and Mönsterås, marking the next step in Södra’s commitment to data-driven, proactive maintenance.
Over the past several years, Södra Värö has worked closely with Viking Analytics to develop a more effective approach to vibration monitoring and condition-based maintenance. The results have been significant. By implementing Viking Analytics’ behavior-based AI platform, MultiViz, the number of alarms was reduced from about 500 to just 20 per week. At the same time, the efficiency of vibration data engineers and the maintenance organization increased, spare parts costs decreased, and plant uptime improved. Today, the total annual savings at the Värö mill exceed multiple millions.
From Large Data Volumes to Measurable Business Results
For Södra, this initiative is not simply about collecting more machine and sensor data. The real challenge is turning data into the right decisions at the right time.
" Vibration data has been available at our facilities for many years, but we have lacked tools to handle the large volumes of data generated by modern monitoring systems. For us, the value is not only in detecting faults earlier; it is in creating impact across the entire organization - increased plant uptime, lower spare parts costs, and more efficient analysis and maintenance operations. We see Viking’s technology enabling exactly that, allowing us to work more proactively and achieve significant annual savings.".
- Thomas Håkansson, Head of Technology, Södra Cell.
Unlike traditional threshold-based systems, MultiViz analyzes how each machine normally behaves across varying operating conditions. The system identifies deviations from that normal behavior and can therefore detect fault symptoms long before they become costly failures. At the same time, normal process variations are automatically filtered out, dramatically reducing false positives. At Södra Värö, weekly alarms dropped from about 500 to 20, freeing valuable time for analysts and maintenance personnel to focus on real issues rather than managing noise.
Proof That Behavior-Based AI Works at Industrial Scale
For Viking Analytics, the agreement represents strong validation that behavior-based AI can deliver measurable results in one of Sweden’s most advanced process industries.
"What makes Södra Värö particularly interesting is that these results come not from a pilot project or an isolated deployment, but from day-to-day operations in one of Sweden’s most advanced industrial facilities. By reducing weekly alarms from 500 to 20, data engineers and maintenance teams have been able to focus on the right problems instead of managing noise.
The results clearly demonstrate that when behavior-based AI is used to prioritize the right actions, the impact can be measured in increased uptime, lower costs, and greater organizational efficiency. Södra’s decision to scale the solution across the entire group is a strong validation of the technology’s robustness, stability, and business value.".
– Stefan Lagerkvist, COO and co-Founder, Viking Analytics
Next Step: Group-Wide Rollout
The new framework agreement enables Södra to progressively deploy the MultiViz platform across all its mills. The objective is to establish a common condition-monitoring approach in which data from thousands of sensors are automatically prioritized and transformed into actionable insights for operations and maintenance teams.
For Södra, this represents an opportunity to further strengthen asset availability, productivity, and cost efficiency. For Viking Analytics, the agreement marks an important milestone in the company’s continued expansion within the process industries. It serves as clear evidence of the value of behavior-based AI in industrial environments.
Photo/Viking Analytics