kandyb studio
Portfolio · Case study 03

Hossa AI

Long-term AI quality scoring for every stock on the Warsaw Stock Exchange & NewConnect — six-pillar composite.

Web · live 762 tickers 6 pillars PIT + Basel III v5.10
Hossa Score 87 Premier · 762 tickers

About

Hossa AI scores the long-term business quality of every stock listed on the Warsaw Stock Exchange (GPW) and NewConnect — 762 tickers in total. Each gets a 0–100 Hossa Score across six pillars: Hard Fundamentals (30%), Soft Fundamentals (24%), Market Signals (10%), Filing Sentiment (10%), Risk (18%) and Multi-Year Stability (8%).

Philosophy: business quality on a 5+ year horizon, not 5-week hype. 80% of the composite weight operates on a 3–10 year timeframe. Point-in-time (PIT) aware scoring eliminates look-ahead leak; quarterly filings are extracted via an AI batch pipeline.

Banks are scored via a dedicated Basel III pillar (NIM, CAR, Tier1, LCR, NSFR, NPL). Developers benchmark against their order book. Hossa AI is educational; it does not constitute investment advice (MAR EU art. 20).

Key features

  • Six-pillar composite — long-term DNA (80%), short-term signals (20%)
  • Frontend-safe labels (Premier / Solid quality / Stable / Caution / High risk)
  • Point-in-time scoring — no look-ahead leak, audit-ready lineage
  • Dedicated Basel III scorer for 13 banks
  • Order book baseline for 31 construction & real-estate developers
  • Educational only — does not constitute investment advice (MAR EU art. 20)

Screenshots

Hossa AI overview screen
Hossa AI scoring detail

Want to see Hossa AI in action?

Reach out — we are happy to show a demo and walk you through the scoring methodology.

Get in touch →