ETWA 2026: Post-Trade Questions Every Commodities Firm Should Be Asking

Your records say one thing. Your counterparty's say another. Somewhere between execution and settlement, your operations team is left to untangle the issue.

Across our workshop and panel sessions at ETWA 2026 in Houston, the message was clear: AI adoption in commodities post-trade is uneven, but dedicated use cases are moving into production. The people who run post-trade operations are practical about it. They want control, clean data and a clear return.

Two-thirds of the room named the same bottleneck

Our panel asked the room which post-trade task was the most urgent manual bottleneck, or the lowest-hanging fruit for automation. The answer was clear:

• 67% said reconciling trades between their ETRM or CTRM, brokers and bank statements
• 17% said processing paper confirmations and email recaps
• 13% said downloading, verifying and explaining daily ISO and RTO invoices
• 9% said ingesting 15-minute meter data and executing ISO position schedules

When asked about AI, the room was split. A third (35%) are already implementing dedicated AI agents for document extraction, confirmations and reconciliation. Another 35% use general AI tools, such as large language models for training materials. Almost a quarter (23%) aren't using AI in post-trade at all.

So the problem isn't a lack of interest. It's that the work causing the most pain sits between systems, where no single system owns it.

Why working outside the ETRM creates risk

Ask most firms and they'll tell you their confirmations and reconciliations are automated. Then someone asks: what about the emails?

Broker and trader recaps, amendments and nominations still arrive by email, PDF and spreadsheet. Fees sit in the body of an email. Amended trades never make it back into the workflow. All of it lives outside the ETRM, and it's handled by hand. That creates risk in three ways.

Errors travel downstream. A price difference caught when a confirmation first arrives is cheap to fix. Caught at settlement, it can become a dispute, a delayed payment or a financial loss.

Automation can make it worse. As Xceptor's Younus Khan put it on the panel: if you automate messy data, you just get the same break faster.

Growth multiplies the problem. Our partners at Smart Communications made the point that manual processes become a blocker to growth. Every new asset class, counterparty or jurisdiction brings more documents, in more formats, with more client-specific variations. With power demand surging and LNG swinging between extremes, that volume and volatility is only rising.

Fixing this is as much a business problem as an IT one. Operations understands where breaks occur and which controls are required. Technology teams understand how data moves between systems. Designing the process together reduces manual work without removing the checks needed before settlement.

A checklist: six questions every commodities post-trade team should answer

These are the questions we put to the room in Houston. If any of them is hard to answer, that's where to start.

1. Can you actually trust your trade data?

Or are you quietly hoping it's right?

What good looks like: Data validated as it comes in, with static data cleaned up at source and one version of the truth across systems.

2. Do agreed terms and amendments make it to operations?

Or do they get lost in an inbox?

What good looks like: Every broker recap, amendment and nomination captured and brought into the workflow, whatever format it arrives in.

3. Can you prove you and your counterparty agree?

This is about proof, not assumption.

What good looks like: Every inbound confirmation extracted, checked and matched against your books, with a full audit trail.

4. Do your physical and financial records line up before delivery, invoicing or payment?

Before the mismatch becomes a dispute.

What good looks like: Reconciliations that run before money moves, not after.

5. Who owns the errors that keep coming back?

And can you put a number on the manual effort, risk and delayed cash they cost you?

What good looks like: Every exception routed to a named owner and visible on one dashboard, with match rates tracked over time, so root causes such as bad static data get fixed for good.

6. Where can AI really help?

Beyond the hype, AI makes the biggest difference with unstructured data.

What good looks like: AI handles defined tasks such as extracting confirmation data or proposing match rules, while people approve consequential decisions and every output can be checked and audited.

How Xceptor and Smart Communications help

One process for confirmations, end to end.

Together, Smart Communications and Xceptor cover the confirmation lifecycle from drafting and delivery to inbound extraction, matching and exception management. Outgoing confirmations are built from a pre-approved clause library and delivered electronically. Firms can manage complex confirmations in one controlled process, without maintaining hundreds of static templates or replacing their ETRM.

AI with two layers of control.

Xceptor's confirmations agent reads a counterparty confirmation in any format, with no templates or OCR. It classifies the document by asset class and trade type, extracts the trade economics and key legal clauses, and flags anything missing, with a confidence score and the reasoning behind each extraction. Matching against your books then runs on rules, not AI. Your team manages only the exceptions, inside your existing risk framework.

One control layer for everything else.

Xceptor brings emails, PDFs and spreadsheets into a single control layer alongside invoices, nominations, and broker, clearer and treasury reconciliations. You get one view of where your operational exposure sits and which items need attention. And because the data is cleaned on the way in, results improve fast. In one example Younus shared, a reconciliation stuck at a 30 to 40% match rate jumped to 90% once the static data was fixed.

The goal isn't fewer clicks or fewer people. It's less risk.

Let's talk

None of this means ripping out your ETRM. It's about automating the work that still sits outside it, so your post-trade operation can grow as fast as your trading business.

If you're asking the same questions, we'd like to hear about it. Book a demo today and see how Xceptor can power your post-trade team.

Want to understand where the bottlenecks are across your post-trade operations?

Our post-trade maturity assessment helps you uncover opportunities to reduce manual effort across confirmations, settlement, reconciliations, and cash movements.

In as little as six minutes, you'll receive personalised insights and practical recommendations: https://www.xceptor.com/post-trade-maturity-assessment