Ongoing research monitoring runs without manual input. High-quality insights are surfaced and routed to the right place. Low-quality results are archived rather than cluttering the workflow. The system is live and running on n8n Cloud.
The Outcome
A scheduled agent that pulls research topics, runs Perplexity searches for each one, formats the results as structured notes, evaluates each note for quality using an AI scoring model, and routes the output to one of four destinations based on score: analyst queue, writer queue, inbox, or archive. Only high-quality insights reach active use. Everything else is filed automatically.
What I Built
Ongoing market and competitive research requires monitoring across multiple topics. Doing it manually means either spending hours every week or letting it fall behind entirely. Most of what gets collected is noise, and separating signal from noise is its own time cost.
The Problem
Analyst Agent