Commodity and energy trading

A commodity trade starts with a deal at the desk. It then passes through confirmation, logistics, contract administration, risk, finance, legal and tax before the money settles. The ETRM (energy trading and risk management) system is the system of record for deals, positions and exposure, and most other tools exist to feed it or read from it.

Trading houses grow fast and enter new markets, so each team tends to build its own way of doing its part. The cost shows up as inconsistent confirmations, hand-offs that depend on a person, and KPIs that each team reads differently.

• Training for the people who touch a trade. AI avatars can play a counterparty, a colleague or a coach for traders, operators, contract administration, risk, finance, legal and tax. See AI avatar training, delivered for a trading company in the UAE.
• Reporting across the lifecycle, so a delay at one stage is visible before it reaches the next. See the trade KPI report.
• Standard process and standard documents: one trade lifecycle model with owners, controls and KPIs, and one trade recap form. See business process management across the trade lifecycle and the trade recap application.
• Tooling around the ETRM system instead of a replacement for it. The bunkering blotter captures a deal once and passes it to Aspect, ION’s ETRM platform.
• Document work over confirmations, contracts and invoices, where a failure is a flagged clause for a person to review. See the tiered document pipeline.

Process standardisation comes first. Reporting and training depend on a lifecycle everyone describes the same way, and AI added to an inconsistent process automates the inconsistency.

A trading business answers to auditors, lenders and counterparties, so every step needs an owner and a record. The ETRM system stays the single source of truth for deals. Anything we build around it writes back to it, keeps the original input, and logs who changed what. Where a model drafts or extracts, a person approves before a document leaves the company or a trade is booked.

Rules depend on where the business trades and which instruments it uses. EU wholesale energy trading falls under REMIT, which sets market conduct and reporting duties.¹ In Switzerland, derivative trading falls under FinfraG, and the classification of a contract decides which duties apply.² Counterparty and employee data is personal data under the FADP.³

The AI Act does not name commodity trading as a high-risk use case. A system used for HR decisions or worker monitoring can still be high-risk, so classify each system by its function.

This page is a screening note, not legal advice. Ask your counsel before you rely on a classification.

• Building a tool that bypasses the ETRM system, so two records of the same deal drift apart.
• Standardising the documents while each team keeps its own definition of the KPI.
• Sending counterparty or pricing data to a general-purpose model without checking contract confidentiality terms.
• Letting a model confirm or book a trade without a human approval step.

Start with the AI opportunity assessment. In one to two weeks it names a metric and an owner, and maps the trade lifecycle stage where AI has the clearest payoff. See also AI regulation in Switzerland.

• WorkBusiness process management across the trade lifecycle
• WorkA standardised trade recap application for a commodity trading house
• WorkA bunkering blotter integrated with Aspect
• WorkA trade KPI report across the trade lifecycle
• WorkAI avatar training across the trade lifecycle