Behind the Balance Sheet (Substack)

How One Investment Firm Is Building AI Analysts

Phoenix Asset Management showcased how it is embedding AI agents into its equity research workflow, creating custom tools such as the Himilco screening system and an AI‑driven version of its DREAM evaluation model. The firm leverages large‑language models, workflow platforms like n8n and LangGraph, and data‑search tools such as Supabase to automate document ingestion, comparative analysis, and scenario testing. These AI assistants accelerate knowledge ramp‑up, expand data coverage, and have already cut research cycles from months to weeks, while still relying on human judgment for intuition and forward‑looking assessments. The initiative positions Phoenix as a front‑runner among boutique managers in AI‑enhanced investing.