{"slug":"bayer-s-model-store-copilot-plugin-saves-researchers-months-by-surfacing-existing-predictive-models","url":"https://findausecase.com/use-cases/bayer-s-model-store-copilot-plugin-saves-researchers-months-by-surfacing-existing-predictive-models","title":"Bayer's Model Store Copilot Plugin Saves Researchers Months by Surfacing Existing Predictive Models","description":"Bayer's Crop Science division, working with Microsoft Teams engineering, built the Model Store, a Microsoft Copilot plugin using Teams extensibility and PowerApps to search Bayer's scientific knowledge repository with natural language, helping data scientists and laboratory researchers quickly locate existing studies and predictive models instead of digging through decades of unstructured slide decks and reports, preventing duplicate model development and saving a researcher two to three months of work.","company":"Bayer","industry":"Life Sciences","country":"Germany","aiCapabilities":["Generative AI","Retrieval-Augmented Generation","Conversational AI"],"businessFunctions":["Research & Development"],"technology":["Microsoft Copilot for Microsoft 365","Microsoft Teams","Power Apps"],"deployment":"Public Cloud","problemStatement":"Every day, new data is added to the decades of existing research at Bayer's Crop Science division, creating difficulties for teams using unstructured data sources recorded in slide decks or reports. Teams faced obstacles identifying information in studies, finding existing models, and knowing who was conducting what research because of the overwhelming volume of information.","solutionApproach":"Working with Microsoft Teams engineering, Bayer developed the Model Store, a Copilot plugin that leverages extensibility provided by the Teams platform to search for information using natural language, with PowerApps embedded in Teams as the interface to the scientific data stored in Bayer's knowledge repository. Researchers and data scientists use a few key phrases to quickly locate studies and predictive models, and the Model Store also identifies who is responsible for the research, making it easier to connect with the expert source.","businessValue":"By using Copilot to investigate the Model Store, a researcher in the United States identified a predictive model already developed by a researcher in Germany, preventing a duplicate model from being developed and saving two to three months of work. Previously, it could take days, if not weeks, to find the information; Microsoft Copilot immediately extracts the relevant models for a certain task and connects researchers to colleagues in Microsoft Teams.","evidence":{"band":"high"},"sourceUrl":"https://www.microsoft.com/en/customers/story/1703594261178267318-bayer-microsoft-copilot-germany","dates":{"publishedAt":"2026-10-08T05:45:55.199Z","publishedAtSource":"pipeline","updatedAt":"2026-10-08T05:45:55.199Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/bayer-s-model-store-copilot-plugin-saves-researchers-months-by-surfacing-existing-predictive-models. Bulk republication requires permission."}