Macro-cycle AI research

Classic cycle frameworks (Merrill, Pring) inside an AI agent: auto-collect macros, quantify cycle position, replay asset history, and propose allocation against desk holdings—end-to-end traceable.

1. 01
   Scattered data

   Manual pulls from FRED, BLS, stats bureaus, Wind and more
2. 02
   Subjective cycle calls

   Hard to reproduce senior-analyst judgment
3. 03
   Slow historical replay

   Manual work to score equities, bonds, commodities by regime
4. 04
   Research–position gap

   Conclusions rarely become concrete allocation moves

Dual-track research balances objectivity and flexibility: a rule engine for reproducible baselines; an LLM referee when boundaries blur.

Automation frees analysts for judgment. Cycle calls become quantitative and reproducible. Allocation advice ties to desk holdings with confidence and risk notes.

Rolling out with securities, asset management, and fund clients. Demo and POC available on chosen history windows and asset classes.

Intranet production deploy with local inference; domestic-stack ready with full audit logs. Delivery & support →