Weekly Stock Drop Report

Every week, markets move. Understanding why a stock dropped matters just as much as knowing that it dropped. This agent automates that research pipeline with a strict focus on precision and cost-efficiency.

Deterministic Filtering

Rather than sending broad queries to an AI, the agent starts with the numbers. It pulls 5-day closing prices for every tracked ticker from standard financial data providers and ranks them by percentage change. Only the worst-performing stocks with confirmed negative returns advance to the next stage, which eliminates guesswork and unnecessary processing costs.

Headline Triage

For each declining stock, the agent fetches raw RSS headlines from public news feeds. It only pulls the titles rather than full article content. These lightweight strings are sent to an AI model with a strict prompt to identify the top headlines most likely to explain a short-term price drop. Because headlines use very little text, this step costs a fraction of what full-article analysis would. Generic price trackers, standard institutional filings, and positive news are filtered out before a single full article is ever downloaded.

Targeted Deep Analysis

Only the selected headlines have their full article text extracted and scraped. That specific content is then passed to the AI to produce a concise summary for each article, along with a synthesized conclusion explaining the broader picture behind the stock's decline.

The Final Report

The result is a fully rendered HTML report that requires zero manual research. Each falling stock gets its own card showing the drop, the relevant headlines, individual article summaries, and an overall analyst-style conclusion. The entire process is generated automatically at a minimal token cost every time it runs.