Cleaning the House Before Adding AI: Our Take Ahead of Digitalisation in Commodities Online

On Tuesday 10 November, Gen10 CEO Richard Williamson will be joining a panel at Digitalisation in Commodities Online called Cleaning the house before adding AI: Pragmatic tech reinvestment in commodities trading. The title is a question we keep coming back to on this blog: what needs to be in order before AI can deliver serious value to your business?

Who Is on the Panel?

Digitalisation in Commodities Online runs on 10 and 11 November and Gen10 is an event partner. The sessions cover CTRM innovation, AI in trading systems and balancing speed and governance when it comes to AI.

Richard's panel runs from 15:00 to 15:45 on Tuesday and is moderated by James Gracey, Global Technology Manager, Procurement and Commodity Trading at JDE Peet's. Richard is joined by Felipe Augusto Ribeiro, Data Engineer at Louis Dreyfus Company, and Uche, a Commodities ETRM, Pricing and Risk Specialist at a leading commodities trading company.

Among other things, we’ll be listening out for these three questions:

Platforms, Data or People?

The first question is where technology reinvestment creates the most value in a trading business. Gen10’s view on this has always been the same; we believe that technology is an enabler of people. A good platform takes on the work that does not need human expertise or judgement, so traders, risk managers and operations teams can spend their time on the interesting part: the decisions that add real value and the problems nobody has solved before.

Where Do Legacy CTRM Stacks Fit?

The second question is how to simplify legacy CTRM stacks and establish clean, shared data layers before adding in new systems. There is no single route, because every business starts from different systems, integrations and history.

Richard has been asked something similar before, and he described a path many firms take. Connect your data with minimal change to your operational systems, build a data lake, then a warehouse on top. It gives you answers, but with no link back to the systems where the work happens, an AI agent has nowhere to act. You can read his full answer in the AI test every trading firm should be running , or listen to the panel in November to hear how NaNi integrates AI directly into the CTRM and system of record to nullify this problem.

Questions of data quality make this data organisation more pressing than ever. CTRM Center's latest report on AI in commodities found that a significant proportion of hallucinations and unreliable AI outputs trace back to upstream data deficiencies, not the model itself. We are keen to hear how Felipe, who works with this data every day, and Uche, who sees it across ETRM, pricing and risk, deal with this issue, and what they’ll be changing going forwards.

From Batch Data to Real Time

The third question is how to rebuild core architecture from batch processes to real-time optimisation.

Where AI sits beside the CTRM and pulls data across in batches, every conversation with it is about yesterday's data. Live decisions need a live connection. And agentic AI, where a system takes actual actions rather than answering questions, also needs a route back into the systems where the actions happen, something a batch set-up can make very challenging. We’re looking forward to hearing how traders are responding to this challenge, as well as sharing how NaNi can solve it.

Meet Us in-Person in the Meantime

We’d love to start a conversation about your own data systems and their impact on your AI plans. Richard and Brett Anderhub, our Head of Business Development, will be at two in-person events before the panel.

ICA Istanbul, 14 and 15 October. Richard and Brett are at the International Cotton Association annual event at the Hilton Istanbul Bomonti on both days. Cotton is a good example of the questions the panel raises. Certification data can often sit outside the trading system, in the Better Cotton Platform or a spreadsheet. But when it is managed alongside the relevant contracts and bales in one system, it can easily be valued and traced, and it is ready for AI to use. We would like to hear how delegates are approaching this. Read our ICA Istanbul preview .

LME Week, from 19 October. Richard and Brett will be in London, joined by Advisory Board member Bruce Tozer. Join us at the LME Metals Seminar on Monday 19 October at the QEII Centre or set up a meeting with our team to discuss any of the themes of the week and see just how transformative a joined-up trading and AI ecosystem can be. Our LME Week guide has the full agenda.

Join the Conversation

And wherever you are on your own AI journey, we would love to hear about it. You might be just starting out, running a pilot, or rethinking your data foundations. Get in touch to arrange a meeting with Richard or Brett, either in-person or online.