Founded in 1873 and headquartered in Kohler, Wisconsin, Kohler Co. is a global leader in manufacturing kitchen and bath products, with major, sustainable operations in hospitality.

With so many products and customers, it can be difficult for Kohler to manage purchase orders and inventory. Due to this, Kohler’s goals are to streamline the way they access information and interact with transactional systems, most importantly their Enterprise Resource Planner.

Our Multi-Agent Buying Assistant is a Microsoft Teams App that streamlines the inventory stocking process by supplying clients with real-time data analysis to aid in the decision making that comes with filling inventory.

Our ecosystem supports purchasing managers’ decision-making by coordinating multiple specialized AI agents that analyze a multitude of factors that contribute to Economic Order Quantity.

Through a centralized Copilot experience, users query the entire supply chain ecosystem using natural language instead of manually referencing the spreadsheets and databases.

The application has numerous pathways for employees to explore. Upon asking the agent a question, multiple specialized agents with specific, tailored skills each generate a personalized response that is evaluated by our judge, who presents a final culmination and optimal solution.

The custom-made dashboard provides a general system overview and is also accessible in the Teams apps, where employees generate designable skills for the AI agents through a familiar and easy-to-use interface.

Our system utilizes a front end built with ReactJS and TypeScript, and a back end built with Python, PostgreSQL, Uvicorn, Fabric, and FastAPI. Our AI agents are built using Microsoft Copilot Studio with Direct Line API communication.