The complete post is available where it was originally published on this site
In Chicago, Tax Increment Financing (TIF) districts publish annual reports in PDF form and detailed expenditure data in CSVs. Both formats contain valuable information — but parsing them, cross-referencing them, and answering even simple questions is often too labor-intensive for the average reporter.
Over the last couple months the author built a prototype system that tries to bridge that gap: a command-line tool that uses an AI agent to understand user queries, search across both structured and unstructured data, and return a clean, well-sourced answer.
The result is a functioning demo of agentic search applied to public finance: a system that combines SQL, PDF retrieval, and language model reasoning in a transparent pipeline.
Agentic Search for Investigative Journalism was originally published in Generative AI in the Newsroom on Medium, where people are continuing the conversation by highlighting and responding to this story.