Skip to content
Matthew Oshin
Back to entrepreneurship
Production build

Financial Communications Platform

An AI platform for investor relations, PR, and capital-markets teams: earnings prep, peer and investor intelligence, crisis command, and on-voice drafting in one workspace.

Overview

A single workspace for the work an investor-relations, PR, and capital-markets team does: earnings prep, peer and investor intelligence, crisis response, governance and activism monitoring, IPO readiness, and on-voice drafting, all grounded in live market data. Twenty-six modules sit on a Postgres brain fed by SEC EDGAR, FMP, FRED, Polygon, and other sources, with an AI layer on Anthropic Claude doing streaming generation, tool use, and retrieval over filings. I designed and shipped it as an in-house AI platform for financial-communications teams, prototyping most of it solo on Claude Code. This is a faithful recreation on sample data: every company, figure, holder, and filing in the demo is fictional.

What it involved

  • Earnings Hub: AI prep briefs, predicted analyst Q&A, a live earnings-call simulator, post-call analysis, and consensus tracking in one workspace.
  • Intelligence suite: investor targeting with 13F and ownership tracking, peer benchmarking with transcript mining, and conference prep.
  • Crisis Command: simulate a scenario, forecast market and media impact, and generate the full response playbook in minutes.
  • Corporate Comms: on-voice press-release drafting across ten templates plus a narrative-consistency checker across documents.
  • Grounded in live data: SEC EDGAR, FMP, FRED, Polygon, Finnhub, and more, behind an Anthropic Claude AI layer with tool use and retrieval over filings.
  • Built largely solo on Claude Code: Next.js 16, React 19, Supabase Postgres, and Prisma, deployed on Vercel.

Stack

  • Next.js 16
  • React 19
  • TypeScript
  • Tailwind v4
  • Anthropic Claude
  • Supabase
  • Prisma
  • PostgreSQL
  • SEC EDGAR
  • Vercel