How Microsoft Azure Machine Learning is Reshaping Manufacturing: Hands-On Experience of Sandvik, 3M and GOLDBECK

Summary

• Sandvik implemented Azure Machine Learning for remote monitoring and predictive maintenance, therefore enabling real-time insights and reducing CO2 emissions.
• 3M Developed 1,500 custom machine learning models on Azure to automate sales forecasting, improving forecast accuracy and reducing manual efforts.
• GOLDBECK used Azure and HoloBuilder to digitize construction site management, automating documentation and defect detection, leading to faster project delivery.

Manufacturing is undergoing a significant transformation as companies adopt advanced technologies like artificial intelligence and machine learning. Microsoft Azure Machine Learning is playing a key role in this shift, providing manufacturers with the tools to optimize operations, improve product quality, and make better decisions. 

This article presents real-life case studies from Sandvik, 3M, and BASF, showcasing how these industry leaders are applying Azure Machine Learning to address unique challenges, enhance efficiency, and drive innovation across their operations.

Refining Production Processes at Sandvik with Microsoft Azure Machine Learning

Sandvik AB, a global leader in engineering, recognized that their traditional production monitoring methods weren’t keeping up with the demands of their complex operations. They needed a more effective way to improve efficiency and reduce operational costs.

Company Background

Sandvik AB is a global engineering group specializing in mining and rock excavation, metal-cutting, and materials technology. Headquartered in Stockholm, Sweden, Sandvik is a leader in manufacturing advanced equipment, tools, and services that cater to various industrial sectors. 

In 2023, Sandvik reported a 13% increase in revenues compared to the previous year, reflecting the company's strong market position and continued growth. 

With a focus on innovation and sustainability, Sandvik continuously seeks to improve its production processes to maintain its competitive edge and meet the evolving needs of its global customers.

Challenge

Sandvik faced the challenge of optimizing its production processes to keep pace with growing demand and increasing complexity in manufacturing operations. 

The company needed to improve its ability to monitor and analyze production data in real-time to identify inefficiencies and bottlenecks. The traditional methods of process monitoring and optimization were not sufficient to handle the scale and complexity of Sandvik’s operations, –leading to delays and increased operational costs. 

Sandvik recognized an opportunity to take advantage of machine learning to gain deeper insights into their production processes, enabling them to make data-driven decisions.

Solution

To address their challenge, Sandvik shifted to a Microsoft Azure environment, which was key in standardizing and streamlining their data. The key elements of the solution included:

• Remote Monitoring Service: Sandvik created the Azure-powered Remote Monitoring Service, a new concept that converts data into actionable information. This solution extracts data from machines and uses AI to measure and analyze it.
• Data Integration and Processing: Sandvik began gathering more data from its machines using IoT Hub and storing it in Azure Data Lake Storage. The data was then moved to Azure Synapse Analytics, where it was cleaned and prepared using Azure Databricks.
• AI-Powered Insights: With data inside Azure Synapse, AI models and predictions were created using Azure Machine Learning. These models provided predictive insights, enabling Sandvik to monitor equipment in real-time and make informed decisions based on the data.
• Scalability: Azure’s cloud infrastructure allowed Sandvik to deploy and scale this solution across multiple production sites globally, ensuring consistent improvements in efficiency and performance.

Impact

The implementation of Microsoft Azure Machine Learning had a significant impact on Sandvik’s operations:

• Predictive Maintenance: Sandvik’s Remote Monitoring Service enabled the company to predict when a machine is likely to fail, allowing for timely intervention before failures occur. This has elevated productivity, safety, and energy efficiency.
• Operational Efficiency: The ability to monitor equipment in real-time has led to substantial improvements in operational efficiency, reducing machine breakdowns and lowering operational costs.
• Sustainability Gains: Sandvik’s solution has delivered significant environmental benefits. They improved fuel efficiency and reduced CO2 emissions are among the sustainability achievements, benefiting Sandvik and also the broader mining sector.
• Industry-Wide Impact: The ripple effects of Sandvik’s innovations extend beyond its operations. As the mining industry increasingly demands sustainable practices, Sandvik’s technology contributes to the development of more sustainable mines, supporting the sector’s shift towards electrification and environmental responsibility.

Lessons Learned

Sandvik’s experience with Azure Machine Learning offers several key takeaways:

• Real-Time Data is Vital: Monitoring production processes in real-time allows for immediate identification and resolution of inefficiencies, leading to more efficient operations.
• Predictive Analytics Drive Improvement: Utilizing machine learning to predict and prevent bottlenecks before they occur can significantly strengthen production efficiency.
• Scalability Matters: Implementing a scalable solution is essential for global companies to ensure consistent improvements across all operations.
• Data-Driven Decisions Lead to Success: Access to comprehensive data and predictive insights enables more informed and effective decision-making, ultimately driving better business outcomes.

The Sandvik's case study is available on the Microsoft website, and you can access it by clicking this link: https://customers.microsoft.com/en-us/story/1518521945649112744-sandvik-discrete-manufacturing-azure-en-sweden

3M Improves Forecasting with Microsoft Azure Machine Learning

3M improved its sales forecasting process by implementing Microsoft Azure Machine Learning and other Azure AI services, developing 1,500 custom models tailored to its regions and divisions. This solution provided finance, sales, and marketing teams with timely insights, significantly enhancing forecast accuracy and automating manual tasks.

Company Background

3M, a global conglomerate with $33 billion in sales, employs 85,000 people worldwide. The company has continued to set new milestones, as highlighted in its 2023 Annual Report. Known for its commitment to innovation, 3M constantly explores advanced technologies to enhance its operations and maintain its competitive edge across various industries.

Opportunity

3M is known for bringing science to life through innovative products for various markets, including home, business, and industrial customers. As a large, global enterprise, 3M recognized the need to standardize and automate its sales forecasting process, which was previously managed through multiple, highly manual methods. 

By adopting Microsoft Azure Machine Learning and other Azure AI services, 3M developed and implemented 1,500 custom models tailored to its regions and divisions. These models now provide timely insights to finance, sales, and marketing teams, streamlining processes and freeing employees from time-consuming manual tasks.

Solution

3M take advantage of the power of Microsoft Azure to build a robust, cloud-based forecasting solution, focusing on the following components:

• Platform Choice: Azure was selected for its advanced machine learning capabilities and compatibility with 3M’s existing data infrastructure, including a Snowflake data warehouse already hosted on Azure.
• Many Models Solution Accelerator: 3M used this tool to train and score approximately 1,500 machine learning models, each tailored to specific region-division combinations, enabling precise and customized forecasts.
• Scalability and Speed: Azure’s ability to provision compute clusters and nodes on demand allowed 3M to scale up rapidly and reduce development cycle time, delivering the first iteration of the solution in just three months.
• Automation and Efficiency: Azure’s automated machine learning capabilities helped refine the forecasting process, reducing manual effort and increasing accuracy across the company’s global operations.

“With Azure AI capabilities, business conversations become more meaningful, and strategic decisions become more targeted.” - Jeff Neilson: Data Science Manager, 3M

Impact

The implementation of Azure Machine Learning at 3M has yielded significant positive outcomes:

• Enhanced Forecast Accuracy: The Azure Machine Learning pipeline, developed in collaboration with business partners, provides more frequent and accurate sales predictions, improving decision-making at the executive level.
• Improved Efficiency: By automating data analysis and forecasting, 3M's finance experts can now focus on strategic tasks rather than manual data entry, leading to more meaningful business discussions.
• Business Buy-In: The successful integration of Power BI and Azure Machine Learning has garnered positive feedback, with increased adoption expected to further refine sales planning and operational efficiency.

“As more of our groups rely on the Azure Machine Learning solution, our finance experts can focus more on higher-level tasks and spend less time on manual data collection and input.” - Jeff Neilson: Data Science Manager, 3M

Lessons Learned

3M's experience with Microsoft Azure Machine Learning offers several key insights:

• Standardization is essential: Moving from multiple manual forecasting methods to a standardized, automated solution significantly improves accuracy and efficiency across a global enterprise.
• Business Integration is Key: Collaborating closely with business partners ensures that technical solutions effectively address real-world challenges, leading to better adoption and impact.
• Scalability Enhances Impact: Maximizing scalable cloud solutions allows for rapid deployment and consistent improvements across all regions and divisions, ensuring uniform benefits.
• Automation Frees Resources: Automating data-heavy tasks enables employees to focus on higher-level, strategic activities, leading to more meaningful business discussions and decisions.

The 3M's case study is available on the Microsoft website, and you can access it by clicking this link: https://customers.microsoft.com/en-us/story/1504342377134122633-3M-manufacturing-azure-machine-learning

GOLDBECK Transforms Construction Site Management with HoloBuilder and Microsoft Azure: Enhancing Efficiency Through AI and Robotics

GOLDBECK uses HoloBuilder's 360° image capture and Microsoft Azure's AI capabilities to digitize construction site management. This solution integrates construction progress documentation with digital building plans and automates defect detection using Azure Machine Learning, reducing the need for manual data entry and analysis.

Company Background

GOLDBECK is a family-owned construction company founded in 1969 and headquartered in Germany. With over 8,500 employees across Europe, GOLDBECK specializes in constructing commercial properties and is known for its innovation in digitizing building processes to enhance efficiency and project management.

Challenge

Managing construction sites traditionally requires physical presence, making it difficult to document and analyze progress without being on-site. Site managers spent significant time manually capturing and organizing information, which often led to inconsistencies and errors. GOLDBECK needed a digital solution to improve site documentation, track project progress accurately, and reduce the burden on site managers.

Solution

• HoloBuilder Platform: GOLDBECK implemented HoloBuilder’s solution, which uses 360° cameras and AI to capture and document construction progress. The images are integrated with digital building plans and stored securely in the Microsoft Azure cloud.
• Azure Integration: The solution uses Azure Machine Learning to analyze the captured images, recognize building components, and detect discrepancies between the actual and planned progress. Site managers can easily identify issues and notify contractors, ensuring timely resolution and minimizing delays.
• Robotics Support: For autonomous photo capture, GOLDBECK utilizes Boston Dynamics’ Spot robot, which can navigate the construction site independently and upload images to HoloBuilder. This functionality further reduces the need for site managers to be physically present at all times.

“Robotics and artificial intelligence will have a major impact on the future of the construction sector. This is already clear from the positive effect HoloBuilder and Azure are having on our site managers’ workload—which ultimately benefits our customers.” - Maximilian Schütz: Director of the BIM Department and Managing DirectorGOLDBECK US Inc.

Result

GOLDBECK’s use of HoloBuilder and Azure has improved documentation accuracy, reduced manual workload for site managers, and accelerated project delivery. The solution was awarded the Intelligent Manufacturing Award by Microsoft and Roland Berger in recognition of its impact on the construction industry, and GOLDBECK has rolled out the platform to around 50 of its building sites​.

Lessons Learned

• Human-Centric Approach: Focusing on digital tools that complement and augment site managers’ work has ensured higher adoption and engagement.
• Automation and AI: Leveraging AI and machine learning has reduced time spent on repetitive tasks and improved defect detection, enabling faster issue resolution.
• Scalability: The success of HoloBuilder and Azure at multiple sites demonstrates the scalability of the solution, making it suitable for deployment across GOLDBECK’s other projects.

The Goldbeck’s case study is available on the Microsoft website, and you can access it by clicking this link: https://customers.microsoft.com/en-us/story/1422938843831293737-goldbeck-HoloBuilder-azure-machine-learning-en

Conclusion

The experiences of Sandvik, 3M, and GOLDBECK highlight the transformative impact of Microsoft Azure Machine Learning in manufacturing. These companies have successfully optimized operations, improved decision-making, and automated processes through AI and cloud-based technologies, leading to enhanced efficiency and sustainability. These examples demonstrate the power of integrating advanced technology with industry expertise, positioning companies to innovate and meet future challenges more effectively.

If you're interested in exploring how Microsoft Azure can help your business, reach out to us and learn more about how we can support your journey toward greater efficiency and innovation.

If you’re interested in exploring how Microsoft Azure can help your business, we encourage you to book a call with our team of experts. As a Microsoft Gold Partner, we can discuss how you can leverage machine learning and cloud technologies for success.