The project was mission to create an intelligent logistics co-pilot using the RAG (ROTIVAL-AUGMENTIED Generation) approach that can act integrated with LES (Logistics Execution System) to transform traditional logistics operations into intelligent processes. The solution was designed to interpret inputs in natural language and cross these inputs with operational data such as labels, storage positions, inventory layouts, movement times and inventory status. With this, the co -pilot now offers direct support to those in the operation, eliminating the need to manually seek information on different systems, facilitating quick decisions and delivering more accuracy in the logistics flow.

---

The biggest challenge was to develop a solution capable of dealing with the complexity and fragmentation of logistical data, ensuring that the copilot could understand not only the commands and questions asked in natural language, but also the context behind these questions. It was necessary to connect the co -pilot to different systems (such as WMS, ERP and movement historical) and ensure that they could translate complex consultations into actionable insights. In addition, it was essential to build a system that dealt with typical factory floor challenges: variations in operations, unprecedented deviations, scattered information, and constant need for real -time decisions.

---

The idea was to create an intelligent multimodal architecture, where the logistics copilot acts as a link between operators and systems. It receives questions and commands in natural language (for example, “where are the replaced materials not distributed?” Or “What is the main cause of picking delay?”), Seeks the answers on updated bases and delivers organized, called context -enriched information. The application combines analytical intelligence, operational history and connectivity with legacy systems, turning gross data into answers that help not only to understand what is happening, but to quickly decide what to do.

---

• Material allocation optimization on the bins

• Reduction of Picking Time with Intelligent Grouping

• Dynamic suggestions for non-distributed substitutions

• Prevention of traffic jams and work zones

• Identification of deviations and failures in logistics processes

• Intelligent Adjustment of Rework and Fallback Strategies

• Fluid integration between systems and historical

• Process Standardization and Manual Failure Reduction

• Creation of an intelligent logistics copilot for real-time support

---

Benefits include optimization of material allocation in the bins, ensuring that the right features are always in the right places, reducing errors and waste. The system significantly reduces picking time by suggesting intelligent groupings and optimized sequences, accelerating operations and increasing productivity. It also allows you to quickly identify replaced or unsubscribed materials, avoiding bottlenecks and stops in the process.

In addition, the copilot prevents traffic jams and work zones, improving the overall flow within the warehouse. It dynamically adjusts rework and fallback strategies, ensuring that failures are resolved quickly and without compromising the operation. Fluid integration between legacy, historical and real -time data systems increases traceability and strengthens the governance of operations. With standardized processes and reduction of manual failures, the system generates clear gains in efficiency, accuracy and agility, creating an intelligent logistics copilot capable of supporting actual time operational decisions.