Product descriptions:
In today’s highly volatile food industry, the speed of data-driven decision-making directly determines productivity, operational efficiency and market responsiveness. Within this environment, Dongwon Group, a global seafood manufacturer, aimed to adopt “Work Designed for AI” as their core mission, fundamentally redesigning their entire operations around AI. Starting with tuna deep-sea fishing in 1969, Dongwon has expanded into food, packaging, logistics and secondary batteries, generating approximately 10 trillion won in annual revenue. However, their legacy data systems were siloed by business units, limiting their ability to support company-wide AI-based operations. To address these limitations, Dongwon adopted Databricks to establish a unified data management and AI-driven decision-making framework, which resulted in 96% accuracy in raw material price forecasting.
Relying on a technology stack that hindered integrated analysis
In recent years, predictive decision-making — such as demand forecasting, production planning and inventory optimization — has become a foundational capability for companies across the manufacturing, logistics and food industries. Dongwon Group, one of South Korea’s leading seafood and logistics conglomerates, was no exception. As the company sought to operationalize at an enterprise level, they quickly ran into the limits of their legacy technology environment, which imposed several structural constraints.
Although Dongwon possessed more than 50 years of accumulated data, their siloed data structure across business units complicated enterprise-wide utilization. Differences in data definitions, formats and aggregation cycles prevented the creation of unified management metrics and consistent analytical standards.
Additionally, fragmented data pipelines made it difficult to detect errors promptly. The absence of a stable platform for processing large-scale data also limited the use of real-time analytics and advanced predictive modeling. Ultimately, Dongwon’s tech stack shortcomings complicated the team’s need to apply AI at scale, making the need for consistent data, real-time processing and reliable model deployment increasingly more vital.
Because data engineers, data analysts and data scientists required different tools, languages and access controls, collaboration was inefficient and even frustrating at times. As Jihyeon Lee, AX Team / DT Division Team Leader at Dongwon Group, explained, “These structural issues delayed data-driven decision-making. Analysis lead times were long, and deploying models into production took significant effort, repeatedly impeding the realization of business value.” For these reasons, Dongwon decided to adopt the Databricks Data Intelligence Platform to enable integrated intelligence across the enterprise.
Unlocking enterprise-wide insights for unified decision-making
To enable unified analytics and enterprise-wide data utilization, Dongwon Group built an internal analytics platform on Databricks, creating a shared data foundation that supports consistent analysis and reuse across all subsidiaries. This shift expanded access to data for business users while reducing dependence on manual reporting workflows.
A key initiative was D:CUBE, a centralized dashboard that aggregated data from multiple operational systems across all of Dongwon’s subsidiaries. Using Databricks as the underlying analytics layer, data was continuously ingested, cleansed and standardized before being delivered to downstream business intelligence (BI) tools. With reconciled data automatically reflected in Power BI, key performance indicators (KPIs) for sales, finance and procurement were redefined using a single, enterprise-wide standard. This ensured that all subsidiaries could analyze performance and make decisions using the same data definitions and metrics. Jihyeon noted, “The accuracy and speed of management reporting have significantly improved, giving us more room to respond with greater agility to changing business situations.”
When it came to inventory management, Databricks enabled a consolidated view of inventory turnover across the entire group, spanning finished goods, raw materials and packaging materials. By analyzing these categories together rather than in silos, Dongwon identified early indicators of obsolete or slow-moving inventory and took proactive action. To make these cost reductions stretch further, Dongwon combined external market indices with internal procurement data to build a raw material price forecasting model, improving procurement agility and strengthening risk management in response to global price volatility.
With the full analytics and AI lifecycle — from data ingestion and cleansing to analysis, modeling and deployment — managed on Databricks, the company eliminated fragmented tooling and handoffs between systems and enabled teams to move from raw data to insight and action more efficiently. A Delta Lake–based architecture provided a unified and reliable data foundation, ensuring consistent schemas and definitions across subsidiaries while removing discrepancies that previously limited enterprise-wide analysis. Tapping into the power of Unity Catalog, data assets, access controls and usage histories were centrally governed, which gave Dongwon greater visibility, stronger security controls and improved compliance across the organization.
These changes paved the way for the beloved brand to establish the Dongwon Citizen Data Scientist (CDS) Academy, scaling their Databricks adoption beyond centralized data teams and embedding data-driven decision-making within all business functions. Through this program, employees gained hands-on experience working with governed data and analytics tools within the Databricks environment. By leveraging AI/BI Genie and Databricks Assistant, the group enabled natural language–based data exploration of the platform, allowing users to ask questions in plain language, generate insights more quickly and reduce reliance on technical specialists. As a result, data access and insight generation became faster, more consistent and more widely available. According to Lee Seok-hwan, Head of Platform at Dongwon Group, “With Databricks, we have transitioned to a way of working where anyone can instantly access the data they need to make the important decisions that move the business forward.”
Strengthening revenue and operational resilience using data and AI
Since adopting Databricks, Dongwon Group has achieved tangible results in revenue and profit protection. Even with extreme price volatility, their new raw material forecasting model achieved approximately 96% accuracy, becoming a critical component in reducing risk for large-scale procurement decisions. By integrating external indices with internal purchasing data, the group can more precisely time procurement and pricing decisions, thereby stabilizing production and minimizing cost fluctuations.
At the same time, inventory optimization models enabled the integrated management of inventory turnover across all subsidiaries, identifying obsolete stock and generating monthly savings of billions of won. Real-time production metrics analysis significantly improved manufacturing efficiency by reducing defect rates and increasing equipment utilization, while also accelerating anomaly detection and response. Monitoring cycles for manufacturing and logistics KPIs were shortened from daily to hourly, greatly enhancing operational responsiveness.
From an IT infrastructure perspective, consolidating previously fragmented systems into the Databricks Platform reduced redundant infrastructure costs, ensuring resources are now used only when needed. Additionally, batch processing times were reduced to mere minutes, establishing a solid foundation for real-time decision-making.
Finally, reporting practices shifted from PC-based document creation to direct, platform-based data utilization. Currently, about 10% of all office workers are directly identifying and executing data-driven initiatives via Databricks, with expectations to expand this to over 50%. With all employees collaborating on a shared set of data and standards, the organization is evolving into a truly AI-ready workplace. Jihyeon concluded, “Databricks has enabled Dongwon to transform our ‘Work Designed for AI’ initiative from a vision into execution — simply by using a single intelligent solution that enables enterprise-wide, data- and AI-driven business innovation.”