Market Research & Copywriting Agent is a project for copywriters and marketers who need a better understanding of their target audience before creating content. Traditional audience research can take several days, while relying on a single LLM prompt can produce assumptions instead of evidence. My goal was to make this research process faster while keeping the results grounded in real online discussions.
The application first researches the target audience across X and the wider web. It turns the collected evidence into a structured market research report and a psychological analysis. From this, it creates an audience agent that represents the researched audience and can be questioned through chat. The research, analysis, and audience perspective are then used to generate marketing assets. I built the system as a multi-step AI workflow using LangGraph, multiple LLMs, external search tools, RAG, FastAPI, React, SQLite, and LangSmith tracing.
The biggest learning for me was seeing how much stronger the results become when different AI approaches are combined. Models alone are useful, but connecting them to external data, tools, structured workflows, and retrieval produces much more grounded outputs.
My next step is to fill the RAG knowledge base with real performance data, improve the quality and consistency of the generated outputs, and continue preparing the project for commercialization.
