Product descriptions:
SK Shieldus is South Korea’s leading converged security company, offering integrated solutions across physical security, information security and industrial safety. As data and AI have become essential for anticipating threats, personalizing customer service and improving operational efficiency, SK Shieldus is rising to the challenge, delivering advanced security services to more than 1 million customers nationwide. Despite operating the country’s largest security network, the company’s previous data infrastructure lacked the structure and scalability needed to support advanced data-driven services. This limited their ability to fully leverage data across the business. To overcome this, SK Shieldus adopted Databricks and, within just six months, built a unified data analytics platform. As a result, they quickly achieved major milestones — including increasing the accuracy of their machine learning models by more than 20%, surpassing 90% overall accuracy.
Dealing with legacy limitations like a subpar data infrastructure
SK Shieldus set out to evolve their security service offerings to be more responsive, efficient and customer-centric. The goal was to better serve clients by identifying risks, detecting incidents earlier and tailoring support in real time. To achieve this, the company wanted to integrate a range of data-driven machine learning technologies, such as cancellation prevention and predictive control models, into their workflows. At a more granular level, the cancellation prevention scoring model would identify customers likely to churn, enabling proactive outreach. Meanwhile, predictive control models could support early detection of abnormalities or potential incidents in security control areas to allow for faster, more effective responses.
Yet, executing on these initiatives proved to be difficult as SK Shieldus’ existing data infrastructure lacked an environment that supported such advanced technology. With no systematic data or AI system in place, the company faced other limitations. For instance, data was spread across multiple operational systems in different formats and structures. Plus, each business unit was managed by their own operational IT department, making integrated analysis difficult. Ensuring the compatibility and quality of data collected from source systems was also a substantial issue that created even more need for an integrated solution.
Additionally, the lack of a common platform between the engineering team, who collected and provided data, and the data science and analytics teams, who utilized it, led to collaboration that often felt more chaotic than cohesive. With a gap in their standardized data processing procedures, it was also difficult to prepare the necessary data at the right time, especially considering disparate systems frequently led to added delays in data utilization. "The low accuracy of data analysis in our existing environment resulted in the poor quality of our data-driven decision-making," a Data Group Leader at SK Shieldus said.
Overall, these obstacles prevented the organization from fully realizing the value of their data assets cross-functionally, which harmed business competitiveness. As a result, SK Shieldus set up proofs of concept (POCs) for various solutions on the market, including Snowflake, and decided to adopt the Databricks Data Intelligence Platform because of its ability to support unstructured data and unify analytics and AI into a single foundation.
Maximizing data utilization with an integrated data analytics platform
In adopting the Databricks Data Intelligence Platform, SK Shieldus improved their legacy data processes by building an integrated data analytics platform. The new platform conducted data collection and management from a variety of source systems and streamlined the entire process, from collection and filtering to analysis and delivery — maximizing the value of the company’s data assets while establishing a culture of data-driven decision-making.
This investment in Databricks also enabled the successful execution of a variety of easily repeatable tasks that consumed an inordinate amount of time. First, the cancellation prevention model proactively identified customers likely to churn by scoring complaints and after-sales histories. Then, it provided tailored responses using generative AI–powered customization guidance to improve both consultation quality and retention rates. Second, the predictive control model analyzed abnormality signals and automatically identified routine responses, call-in responses and whether or not actual dispatch was needed, which ultimately lowered the number of unnecessary dispatches to boost operational efficiency. Lastly, geospatial analytics, full-funnel analytics and other various data were supported to enhance SK Shieldus’ overall business competitiveness.
On the data quality management side, Delta Lake and Databricks Lakeflow Jobs were leveraged to build automated filtering pipelines and real-time monitoring dashboards. The combination of both tools solidified data accuracy and reliability while taking the data quality of their machine learning models to the next level. Because of these improvements, the relationship between the data engineering team and data science and analytics teams became stronger and more productive. Since the data collected and filtered by engineering was immediately available, tangible analytical outputs, such as cancellation scoring, retention guidance and predictive control models, were easier to produce. "With Databricks, we could systematically connect across teams and create an organic collaborative environment between specialized areas without the usual amount of friction we experienced prior," the Data Group Leader explained.
Continuing to create a scalable, secure data environment, SK Shieldus adopted Unity Catalog alongside Delta Lake and Lakeflow Jobs for systematic management of data security and access controls. This ensured data was shared seamlessly across teams while maintaining governance and precision. Since Delta Lake paved the way for flexible data filtering and the configuration of domain-specific data marts, MLflow easily worked in conjunction to help the company establish a consistent environment for model experimentation and tracking. With the tactical combination of these Databricks tools, SK Shieldus laid the foundation for faster iteration, more reliable insights and greater operational efficiency — unlocking measurable improvements across customer experience and security operations.
Building end-to-end analytics in just six months
SK Shieldus successfully built a data analytics platform based on Databricks in just six months, further amplifying their data utilization capabilities. In particular, a cancellation prevention model that predicted customer churn was developed and applied, which reduced call times and improved retention rates. With the new model, the team effectively retained customers who were at risk of churn. Not to mention, the use of predictive control models reduced unnecessary dispatches by 2%.
Machine learning model development and MLOps automation also accelerated, improving model accuracy from less than 70% to over 90%, a 20% increase. "The Databricks Platform enabled faster decision-making through integrated data analytics and AI workflows," the Data Group Leader concluded. Because SK Shieldus reduced the complexity of data utilization, they dramatically advanced collaboration across the data engineering, data analytics and machine learning teams.
By adopting cloud-based solutions, SK Shieldus optimized budget operations by reducing hardware and maintenance costs, minimizing IT infrastructure burdens and empowering the scalability of their computing resources. With Databricks, the company has shaped a highly flexible environment where large amounts of data can be processed and analyzed quickly, making work that was once difficult more approachable to significantly expedite business productivity.
All in all, these technological investments have bolstered the analytical capabilities of SK Shieldus’ data management and contributed to their newfound business agility and competitiveness. Looking ahead, the team expects that Databricks will continue to play a large role in their data-driven innovation, AI process automation, operational efficiencies and cost savings. Not only does SK Shieldus plan to drive automation and intelligence through generative AI–powered tools across the enterprise, but the organization is also proactively responding to security threats using Databricks’ advanced capabilities. In parallel, they aim to forecast market trends and evolving customer needs to inform future strategy. With this integrated approach that mimics their own business model, SK Shieldus expects to achieve future-oriented growth and gain a sustainable competitive edge.