Building Agentic Due Diligence for High-Ticket Marketers: Architecture, Risks and Open Questions

AgentVN · reddit · 2026-09-06

The author proposes an automated due-diligence app: users input a company, person, domain, social profile, or landing page, and a network of specialized AI agents investigates high-ticket course sellers, agencies, and aggressive e-commerce brands, producing an evidence-backed structured report.

Key design points:

Every finding includes screenshots and sources, with similarity flagged as possible matches with confidence levels. The author seeks feedback on architecture, false positives, legal risk, and whether this is truly an agentic use case or better built as a deterministic pipeline with selective LLM analysis.

Related event: Developer Proposes Multi-Agent AI Tool for Vendor Due Diligence(2 posts)→

Original post →

More from coding & agent

coding & agent channel →