Intelligent VMI powered by AI: the new predictive execution engine in Hinode

A hypercompetitive industry that requires speed, precision and total synchronization

‍

The beauty and cosmetics sector in Brazil is characterized by its strong dynamism: constant launches, short life cycles, intense seasonality and competitive pressure that requires maximum agility to avoid bankruptcies, excesses and loss of business opportunities. In this context, Hinode faced a critical challenge: Synchronize product availability across hundreds of franchises scattered across the country, each with different consumption patterns and high sensitivity to news.

Before Pyplan, the company relied heavily on Manual and poorly integrated processes, based on multiple spreadsheets and local analyses without a single view of inventory, sales and real needs. This made it difficult to anticipate bankruptcies, identify excesses and execute a consistent replenishment. As described in the challenge sheet, the chain required end-to-end synchronization, early visibility, and greater operational resilience to react faster to market signals.

The lack of integration created operational risks such as late refills, overinventory at some points and low availability at others. In a business with more than 350 products and monthly releases, even small deviations could result in significant sales losses and a disruption in the experience of the franchisee and final consumer.

‍

‍

From intuition to data: the urgency of modernizing supply and unifying decisions

‍

To sustain its growth and increase network performance, Hinode needed to evolve into a process collaborative, automated and based on data, especially for supply planning and inventory control. The objectives were clear:

• Connect demand, inventory and replenishment in a single integrated flow.
• Raise the Replenishment accuracy using statistical models and logic for DFU (SKU x store).
• Align the operational replenishment process with practices S&OE modern, enabling faster and more informed decisions.
• Incorporated Predictive Intelligence to anticipate the effects of releases, atypical patterns and commercial events.
• Reduce manual work, increase decision consistency and improve the franchisee experience.

The project required capabilities today considered essential for modern execution: demand sensing, demand control, inventory optimization and collaborative workflows, all of them components of the Pyplan S&OE module that make it possible to translate tactical planning into coordinated daily execution.

In addition, Hinode sought to strengthen the resilience of its chain — an explicit need in its initial diagnosis — through mechanisms that would allow preventive action against deviations and not just react when they had already materialized.

‍

‍

Intelligent VMI powered by AI: the new predictive execution engine in Hinode

‍

Hinode implemented a complete Vendor Managed Inventory (VMI) model on the Pyplan platform, designed to operate with the level of detail required by its franchise network. The solution integrated sales, inventories, deliveries and stock policies into a unified, flexible and scalable model, able to adapt to each store and category, and evolve with the launches and business rules.

The architecture combined key S&OE modules: Demand Sensing to capture real consumption signals, Demand Control to manage diversions, Inventory Optimization to balance costs and service, and collaborative VMI to automate replenishment. This integration allowed us to align demand, supply and daily execution, reducing friction between areas and increasing the speed of response in the network.

‍

The incorporation of generative AI and intelligent agents took the process to another level. The algorithms eliminated outliers, improved release management, and adjusted behavioral curves through multiple iterations, achieving more accurate forecasting and more stable revisions. The agents automated daily analyses, anticipated anomalies and enabled predictive scenarios, transforming replenishment into a prescriptive and non-reactive cycle.

The results were convincing: +30% of the positivity of the mix, +23% of sales and — 11% of inventory in the stores with the highest adherence. Hinode went from managing its network with manual processes to operating with data, AI and automation, consolidating a modern, collaborative VMI ready to scale.

‍

Discover how Pyplan can transform your end-to-end planning.
Learn about our solutions or talk to an expert: https://pyplan.com/es/entre-en-contacto/

‍

‍