Energy Trading Week Middle East launched in Dubai this September. Gen10 was there as a Gold Sponsor and our CEO, Richard Williamson, shared his insights on commodity technologies, trading, and AI. He moderated Wednesday’s session on building a unified data culture and joined Thursday’s panel on digitising the energy supply chain.
Despite the event’s name, the content was relevant for all commodity traders, so whether you’re in ags, metals, or energy, this recap is for you.
The room agreed on the problem. Silos, still.
“Unified data” has been on trading conference agendas for many years now, and the talks in Dubai spent a lot of energy on the same problem: risk, ops and trading looking at different versions of the truth. The good news is that we as an industry are becoming more ambitious about the types of information silos we’re addressing. It isn’t only internal any more.
Your teams are using more systems to access more data than ever before, but none of those systems were built to match your unique workflows in your core CTRM. Breaking down silos now means interoperability with external systems, not just data your teams have extracted from them.
AI in commodity firms: agentic, not just analytical
Where the conversation got interesting was in what AI is being asked to do. Most AI and CTRM vendors discussed analytical activities: summarise this, flag that, tell me what changed overnight. But the sessions Richard took part in included the bigger ambition of AI that can act, or “agentic AI”. An agentic AI can create a new deal or execute a workflow end to end in your CTRM, taking actions rather than just giving information.
Whether you can use agentic AI or not often comes back to the same silos issue we’ve been discussing for years: is your data in order? If you are looking to add an AI point solution onto an existing CTRM, the two systems both need to be able to interconnect (usually requiring APIs and AI-native MCP servers). But the data within the CTRM also needs to be both structured and retrievable.
The problem with many of these AI bolt-ons is that they are pulling CTRM data across into the AI system in batches, then having the very valuable reporting conversation in that external system. But then, the AI can’t take action back in the CTRM, so it is advising but not actioning. And this approach also means that every conversation you have is about yesterday’s data, or at least data that is already becoming out of date.
Several of the larger CTRM vendors at the event explained that they are taking this bolted-on approach over developing integrated AI. It can be a good option for getting insights out of a CTRM that has historically been less flexible, but it’s worth considering what you’d be trading off against an agentic conversation with your own live book, with AI that can act on your conversations.
What building AI into the CTRM gets you
The alternative to the bolt-on model is a CTRM with AI built into it directly, reading and acting on the same live book your traders and logistics teams are working from. That means that the AI assistant can do something about the data you’re interrogating, or take actions based on a conversation with your team. You maintain human oversight as per your own controls, but tasks are carried out automatically and accurately, as opposed to your team needing to go back into the CTRM to make manual updates. Everything occurs inside the system of record, with nothing to reconcile afterwards and no lag between the conversation and the data behind it.
This is what that looks like in practice
AI is also changing the equation of whether to bolt new systems onto a CTRM or to replace it with a CTRM or Commodity Management System with all the functionality you need. Sticking with an ageing CTRM used to feel safer, because replacing it meant years of disruption. But nowadays, with more efficient CTRM implementations, the downsides are increasingly obvious: the CTRM remains the ceiling on what your other systems, especially your AI, can do, every year it stays in place. Replace it now, and your AI has the system it needs to act, not just advise, from day one.
That’s a smaller task than it used to be, and AI is a big part of why. In innovative CTRM companies including Gen10, AI is already doing real work inside implementation projects, speeding up data mapping, configuration and testing. An AI-ready CTRM is not a multi-year project, it should be live and operational within months, and can be a faster build than a system without AI baked in.
Talk to Richard about where you are
Every business we spoke to in Dubai was at a different point on this: some are already running AI reporting, others are trying to gauge which system(s) would suit them best. Wherever you’re up to, understanding the potential of AI for commodity traders is much more helpful as a conversation rather than a blog post. Let Richard know where you are with AI, and he’ll give you his take, grounded in what Gen10’s built and what our customers are doing with it. No sales call attached.