No Experts, No Access, No Problem

The Quiet Power of Scrappy AI Research

Yan Gao | Founder of alphaguru.ai

When I left my former fund in May, I stepped away from everything that had been my safety net for over a decade:

The sell side pipeline, endless expert calls, quiet backchannels to management teams.
That infrastructure shaped how I thought, wrote, and decided.

Without it, I felt stripped down.

But scarcity can be an honest teacher.
With every layer of support gone, I had to face a question I’d quietly avoided for years:
Which parts of my process truly generate alpha, and which are just habits?

What emerged wasn’t a weaker version of institutional research. It was a different practice altogether, scrappier, leaner, and oddly, more fundamental.

AI became my leverage

Not the polished, enterprise AI shown in boardroom demos.
Truthfully, few commercial tools really wowed me, because most were just clever search engines or RAG-based workflow layers, quickly becoming commoditized.

So I built my own approach. I treated AI as a thinking partner, an analyst, and a co-PM. It let me tear down old routines and rebuild my research process from first principles.

Instead of sprawling checklists and processes for its own sake, AI forced me into tighter loops.
I’d start with a single, critical question, the one that, if answered, would truly flip conviction or sizing.
From there, only the sub-questions that could change the outcome survived. Everything else got cut or reimagined.

1. Replacing Expert Calls with AI