ABOUT
I trained in fundamental equity research at Comgest and Moneta, alongside a background in M&A. A large part of the job is deciding, document after document, what actually deserves attention – and that judgement is never generic. It depends on the thesis you are underwriting and on what your firm considers signal rather than noise.
Testing existing AI tools on my own research kept exposing the same limitation. They read quickly and summarize well, but without calibration they do not know what matters in the way you invest. General-purpose models default to generic notions of materiality; specialized platforms make information retrieval dramatically easier, but that does not by itself encode the judgement that determines what deserves attention.
That judgement is precisely what distinguishes one investment firm from another, and much of it remains tacit. Evidence Desk grew out of that observation: the hardest part is not choosing another model or building another workflow. It is making explicit what an investment team already knows, but has never had to formalize.
ONE DESIGN PRINCIPLE
An analytical tool should support judgement, not replace it.
This is as much a technical choice as a conviction. Overly procedural workflows that force every case through the same rigid framework tend to degrade analytical quality. Evidence Desk therefore focuses on completeness, context and traceability, then stops short of making the investment judgement itself. The tools gather the evidence, put it in context and source it. The conclusion remains the analyst’s.