This project explores an agentic AI system designed to interact with and analyze a marketing dataset using natural language. The chatbot is capable of answering analytical questions, generating insights, and supporting follow-up reasoning over structured data.
The goal is to build an intelligent assistant that can query a marketing dataframe and provide data-driven answers. It supports exploratory analysis, trend detection, and basic visualization generation through natural language interaction.
The system combines a large language model with structured data operations. It uses 3 tools which I have designed, for simple filtering and aggregation, more advanced filtering and aggregation, and plotting a feature. The LLM interprets user intent, selects appropriate analytical operations, and interacts with the dataframe to compute results. This allows it to behave like an analyst capable of reasoning over marketing performance data in real time.