The growing complexity of procurement operations at a leading pharmaceutical company has led to an overload of dashboards, increasing reporting inefficiencies and limiting data-driven decision-making. To address these challenges, our project explores how Generative AI (GenAI) and automated data visualization can optimize procurement analytics. We developed a proof-of-concept (PoC) chatbot that allows procurement managers to use natural language queries to generate dynamic, accurate visual insights without relying on traditional dashboards or advanced technical skills. Leveraging open-source tools, notably LIDA (a grammar-agnostic library that generates visuals and infographics), the system connects to structured procurement data, processes user queries through Large Language Models (LLMs), and returns contextualized visual outputs. The framework was tested across four high impact use cases, including vendor spend summarization and forecasting, achieving a 96% success rate in producing accurate and contextually appropriate outputs. Key outcomes include reduced dependency on the BI team, significant time savings, and enhanced decision-making efficiency. The project also outlines a scalable deployment strategy, emphasizing data governance, user training, and prompt engineering to mitigate challenges like jargon misinterpretation and dataset scalability. This approach offers a sustainable, cost-effective alternative to commercial solutions, empowering procurement teams to focus on strategic value creation while maintaining flexibility for future AI advancements.