Problem
Competitive content research in B2B was a manually-intensive process: open ten tabs, skim ten articles, synthesize gaps, write a brief. A two-day job every time. At scale with multiple clients, this was unsustainable.
Built
A hub-and-spoke agent system that takes a keyword as input and returns a machine-readable competitive brief. DataForSEO pulls the top-10 SERP. A scraper agent extracts and chunks each URL. An analysis agent identifies content gaps, semantic clusters, and structural patterns. A brief-writer agent outputs a structured document for the content agent.
Result
Average research-to-brief time went from 2 days to 4 minutes. Brief quality measured via human review showed a 94% acceptance rate with minor edits. Now running in production for 8 active clients across DACH B2B verticals.