Transforming Vessel Design and Planning with Snowflake

Situation

The New vessel  Build team managed critical ship design, engineering, safety, and planning data across multiple Excel workbooks maintained by different teams. As vessel planning activities increased, the organization needed a scalable, governed data and decision-science approach to support faster design evaluations and future ship planning.

Problem

• Vessel design and specification data were distributed across multiple spreadsheets and teams.  

• Manual reconciliation and version management created inefficiencies and increased error risk.  

• Benchmarking vessels and evaluating design trade-offs required significant analyst effort.  

• Lack of governance, auditability, and standardized metrics hindered decision-making by reducing trust in data and increasing validation efforts 

• Business users depended heavily on manual reporting and data preparation.  

Solution

Implemented a Snowflake-native data and intelligence ecosystem that: 

• Established Snowflake as the single source of truth for vessel design and planning data.  

• Standardized and integrated data across architecture, engineering, safety, and design domains.  

• Delivered a Streamlit in Snowflake data entry portal with role-based approval workflows.  

• Enabled self-service benchmarking, simulation, and regression analysis through interactive dashboards. 

• Incorporated Snowflake Cortex AI capabilities to allow natural language querying of structured and unstructured information using Cortex Analyst and Cortex Search.

Key Components:

• Snowflake Data Foundation, Framed by AoPSS:  Applying Mu Sigma’s Art of Problem Solving System to map the full vessel design problem space, ship engineering, and regulatory data were centralized into governed Snowflake tables and views – creating a single source of truth grounded in systems thinking, not just data consolidation. 
• Data Standardization & Modeling:  Leveraged Snowflake SQL and Snowpark, guided by internal decision-science methodologies, to harmonize metrics, resolve data inconsistencies, and build curated datasets that reflect real engineering trade-offs.
• Governed Data Management:  Built a Streamlit in Snowflake portal with RBAC-based approval workflows, designed using Mu Sigma’s EoC principles to serve as a shared canvas for decision scientists and engineers – strengthening transparency, traceability, and confidence in every data entry and validation step.
• Self-Service Analytics & Simulation: Delivered interactive benchmarking, design comparison, scenario modeling, and trade-off analysis through Streamlit applications, structured around systems-thinking approach to surface insights beyond isolated metrics. 
• AI-Powered Natural Language Insights:  Integrated Snowflake Cortex Agent, Cortex Search, and Cortex Analyst – combined with behavioral science and design thinking expertise – to enable intuitive, conversational access to structured and unstructured ship design data. 
• Security & Governance:  Enforced role-based access controls, approval workflows, and governed data consumption within Snowflake’s native security framework, reinforced by EoC governance model for organization-wide trust and reuse.
  
  Together, Snowflake’s data and AI capabilities, applied through Mu Sigma’s systems-thinking approach, turned fragmented vessel design data into a governed, AI-ready, and confidently-decisioned analytics ecosystem.

Impact

• 90% reduction in vessel design benchmarking cycle time, accelerating analysis from 8–16 hours to 30–60 minutes.  

• Up to 85% reduction in manual data reconciliation effort, through centralized Snowflake governance and automated workflows.  

• 60–80% fewer re-validation and rework cycles, improving data quality while enabling faster, self-service planning and design decisions.