The Real Problem in Commodity Buying: Decisions Trapped in Spreadsheets
Every commodity team I know does three things brilliantly: negotiate, move fast, and hustle. But the decision layer is broken. Prices change by the minute; approvals take days. Data sits in emails, PDFs, and sheets. The result: late hedges, missed quotes, avoidable P&L swings.
The Problem (What Actually Breaks)
- Fragmented data: Exchange prints, FX, supplier notices, positions, and MTM live in different places.
- Lag in decisions: The market moves in minutes; our workflows move in hours.
- No common view: Procurement, treasury, and finance don't see the same risk at the same time.
- Weak governance: Roles/approvals/audit trails are bolted on, not built in.
- Over-reliance on instincts: Great for direction; bad for repeatability and scale.
What "Good" Should Look Like
- One live workspace for pricing, exposure, and approvals.
- Explainable signals (Buy / Wait / Hedge) that anyone on the team can trust.
- MTM + audit stitched into every action.
- Latency & cost under control as volumes grow.
Why Teams Don't Fix It (Yet)
- "Our sheet works." (Until it doesn't in a fast market.)
- "This is IT's job." (But IT won't encode desk intuition.)
- "Too many feeds/licences." (Compliance is real, but solvable.)
- "We'll do it next quarter." (Volatility won't wait.)
What We're Building at Novaex
- A decision engine for commodity buyers: live pricing, exposure, and approvals in one place.
- Signals, not noise: model-driven Buy / Wait / Hedge with context you can explain to your CFO.
- Enterprise controls: roles, policies, audit history by default — not afterthoughts.
- Built to run low-latency, cost-sensitive inference so it works in the real world.