An auto parts manufacturer with a large and complex supply chain operation struggled to keep track of material inventory, considering the end-to-end supply chain process (open orders, material in transit, delays, etc.). This challenge directly impacted:

• Compartmentalized knowledge by planner and material segment, limiting the scalable level of management due to a very customized and manual analysis workflow.
• Infrastructure limitations in dealing with scattered data from multiple sources.
• Costly manual extractions for frequent execution of the analysis workflow, relying on technical experts in the data foundation to provide accurate inputs.

VF integrated a data layer for inventory (including order history, safety stock levels, and a tracker of orders in transit) into Palantir Foundry and built a control tower application. The solution brought together many layers of information that, integrated with AIP, enabled insight generation and alerts, leading to informed and anticipated decision making.

The work included:

• Integrated lookup: access to materials information in a single environment, replacing RPA extractions in the workflows it covers.
• Inventory simulation: evaluating stock scenarios to support planning.
• Materials in transit: visualizing materials in transport to assess their availability.
• Operational indicators: criticality and availability metrics to flag items that need attention.

The solution was disruptive, making it possible to track all updated information automatically in a single application, in a reliable and consistent form. The main gains were:

• Decrease in inventory levels, based on accurate historical data on material transit and consumption.
• Decrease in losses due to material shortage.
• Anticipation of potential inventory gaps that can be addressed on time at lower cost (avoiding air freight).
• Customized interface to generate simulations and reports and to provide information for decision making.
• Standardized rules and processes across material planning.

Production planning across South America sites depends on complex data manipulation and specialized knowledge to keep everyday operations active. The complexity of dealing with a large amount of information, with segmented knowledge and manual analysis, limits efficiency and accuracy, resulting in higher inventory and production that is mismatched with client demand.

This setup concentrates knowledge in a few people and makes it hard to coordinate planning with operations, providing limited visibility into execution and a slow response in identifying the need to replan.

VF designed a solution that integrates demand data from different sources and combines it with internal supply (such as qualified labor, material availability, and production line capacity, among others), resulting in optimized and reliable views that coordinate everyday operations, including:

• Long-term planning: automatic generation of plans with a roughly 6+ month horizon.
• Short-term scheduling: converting the approved plan into a detailed schedule for the following two weeks, with optimized allocation of orders.
• Decision support: AI-driven alerts and recommendations, plus integration with the existing tool for monitoring material-shortage risk.

The project is moving to a first validation at a pilot plant. Expected benefits include:

• Decrease in inventory level, due to tailor-made modeling for supply/demand balance.
• Decrease in line stoppages caused by material shortages.
• Customer demand met in a more timely manner.

Results will be measured once the platform is rolled out to the remaining plants.

A pharmaceutical company relied on manual data analysis to forecast sales and calculate net revenue across its portfolio. The simulations combined volumes and drug types with many policy attainment mechanisms.

This approach had three main drawbacks:

• Decentralized analysis: separate spreadsheets per product and portfolio made it hard to consolidate projections.
• Risk of calculation errors: combining multiple variables in spreadsheets made building and reviewing scenarios error-prone.
• Limited visualization: the available charting tools made it difficult to analyze projections and compare scenarios.

VF moved the calculation models to Palantir Foundry and built an interface for creating and analyzing financial scenarios. The solution brought calculations and dashboards together on a single platform, with a focus on ease of use for the teams responsible for the projections.

The work included:

• Scenario simulation: configuring volumes and commercial conditions to forecast sales and calculate net revenue.
• Planning interface: adapting the logic used in the spreadsheets to allow scenarios to be created and reviewed in the new environment.
• Projection visualization: charts and dashboards to support analysis of simulation results.
• Flow validation: verification mechanisms to reduce errors while building scenarios.

The solution provided a centralized environment for financial planning, with benefits to the team's day-to-day analysis:

• Decrease in calculation errors, supported by validation mechanisms.
• Simpler and faster scenario creation, enabling assertive decisions.
• Access to the different dashboards on a single platform.

A fashion retail group used fragmented reports and data to track the commercial performance of its brands. They had access to sales indicators, but identifying revenue opportunities still depended on additional manual analysis.

The main challenges were:

• Limited opportunity analysis: the available reports didn't systematically identify underperforming stores or estimate sales opportunities.
• Difficulty scaling: the existing structure limited how well the analysis could adapt to the needs of different brands and business units.

VF's proposal was to build an MVP on Palantir Foundry, initially covering two of the group's brands. The scope combines data integration, comparative store analysis, and an AI assistant.

The solution included:

• Performance comparison across similar stores: clusters of similar stores, grouped by traits such as size, location, and type, set the ground for the analysis.
• Opportunity estimation: calculating a store's revenue potential based on its cluster (sister stores) made it possible to identify specific root causes and address them.
• AI assistant: brings the cross-store analysis to attention, with alerts and suggestions.

The MVP aims to help leadership and store managers identify and prioritize commercial opportunities. The expected benefits include:

• Increase in revenue for stores that have a gap compared to their cluster.
• Prioritized actions based on estimated revenue potential.
• Efficient sales and marketing plans based on granular data.
• A data and model foundation that can extend the solution to other brands and new uses.