From an unsupported trading hypothesis to an evidence-linked analyst workflow

Customer problem

Institutional analysts receive more information than they can review. Historical comparison is manual, relevance is uncertain, and research reasoning becomes fragmented.

Target

User: institutional research analyst
Buyer: Director or Head of Research
Gatekeeper: CTO or Head of Data

Customer-discovery pivot

Six interviews emphasized prioritization, visible evidence, workflow fit, governance, and analyst control. They did not validate adoption or willingness to pay.

Value proposition

Help investment teams decide what to read first, understand what changed, and support conclusions with visible evidence.

Working MVP

Compare consecutive 10-Ks, classify meaningful changes, check earlier 8-Ks and headlines, expose exact excerpts, and preserve an auditable record.

150 companies · 995 changes · 629 high-confidence · 1,990 verified offsets · 0 timing violations

What research found

The product demonstration is operational. Incremental alpha was not supported; the disclosure-change risk test was inconclusive.

Final classification: INSUFFICIENT_EVIDENCE.

Next experiment

A 2–4 week pilot with 3–6 analysts measuring time, recall, agreement, correction burden, trust, repeat use, and stored-context value.