Trading and Analytics
AI in Commodity Trading: Time to Accelerate or Reassert Human Control?
Artificial intelligence (AI) is being increasingly embedded across the commodity trading ecosystem - not just as a support tool, but as an agent actively reshaping:
- How data is handled
- How decisions are made
- How business strategies evolve
From my own conversations across the industry, and from insights shared during recent webinars hosted by Commodities People, it’s clear that while AI’s potential is vast, the path forward demands a more deliberate, informed, and human-centric approach.
In this article I explore key AI trends that are emerging across the global commodities landscape, and what the increasing emergence of new solutions means for talent.
AI as an Enabler – But Only as Smart as We Train It
Across trading, logistics, analytics and operations, AI is already helping firms reduce costs, streamline decisions and gain more immediate access to market intelligence. But the effectiveness of these systems depends heavily on two things:
- The quality of the data they ingest
- The understanding of those who operate them
There remains a significant disparity between firms that have embedded AI into their core processes and those still experimenting at the edges – deploying tools like ChatGPT or copilots without a clear governance structure or optimisation strategy.
Without skilled human input – prompt engineers, analysts who understand the context, and leaders who have the knowledge to shape a strategic AI roadmap – the promise of AI often defaults to automation rather than transformative change.
Control, Trust and the Human Feedback Loop
What has emerged as a key topic is the growing tension between automation and control. AI can undoubtedly accelerate insights and operational efficiency, but without rigorous oversight, its outputs risk becoming misleading or even dangerous.
For example, in trading scenarios, acting on AI-generated data without understanding its source or methodology could expose firms to both commercial and regulatory risk. There’s a growing consensus that humans must remain firmly “in the loop” – validating outputs, interpreting context, and maintaining accountability. The long-term concern isn’t just about job losses but a wider erosion of critical thinking and decision ownership.
Governance, therefore, is not optional. AI must be deployed within a framework of internal alignment, vendor accountability, explainability and continuous feedback. This is particularly vital in regulated environments, where the consequences of poor decisions – even if AI-driven – still rest on human shoulders.
Skills, Strategy and the Executive Agenda
Perhaps the most profound shift AI is triggering is at the leadership level. Traditionally, strategic decision-making in commodity trading has relied on instinct, relationships, and experience. Now, executives are being asked to interpret model outputs, vet third-party AI tools and assess the integrity of underlying data sources – including complex external feeds like weather or macroeconomic indicators.
This requires a different skillset at the top. It’s not enough for senior leaders to understand the commercial or geopolitical dynamics of their markets – they now need fluency in AI governance, model validation and data trustworthiness. The challenge isn’t just in recruiting technical specialists, but in equipping the C-suite with the capability to lead in a hybrid environment, where decisions are increasingly shared between human judgment and machine inference.
Equally, talent strategies must now account for upskilling existing teams. Data literacy, AI literacy and even ethical reasoning are no longer the domain of niche innovation roles. They are becoming core competencies in modern trading houses – from junior analysts to board-level decision makers.
The Road Ahead: Strategic Integration, Not Tactical Hype
The rise of algorithmic trading in energy markets is a case in point. AI is opening access to a broader set of market participants, accelerating response times and enabling new trading strategies. Yet as one panellist warned, legacy IT systems and siloed data remain a major barrier. Even more pressing are the “blind spots” between users and AI providers – gaps in understanding that can undermine the effectiveness of even the most powerful tools.
The next chapter is not about choosing between AI and human expertise. It’s about integrating both in a structured, strategic and explainable way. AI is a tool – and like any tool, its impact depends on the intentions and skills of those who wield it.
Firms must move beyond the novelty of AI and into a phase of sustainable, controlled adoption. That means robust internal processes, clear data stewardship and thoughtful vendor partnerships. And above all, it means leadership that is ready to embrace AI without abdicating responsibility for its outcomes.
Final Thoughts
The use of AI in commodity trading is not a question of if, but how.
Whether enhancing decision-making, optimising risk strategies or enabling algorithmic trading models, AI is here to stay. But so is the need for human oversight, contextual judgement and strategic foresight.
We are not at the end of the human-led trading era – we are at the beginning of a more complex, co-dependent one. The firms that thrive will be those that understand this, and who invest not just in the tools, but in the people and processes required to wield them wisely.
For a conversation about your firm’s talent strategy and how AI is influencing the skills needed at senior and executive level, please don’t hesitate to get in touch.
Katie Dunbar
Associate Partner | Commodities | EMEA
T: +44 12 7364 8080 E: katie.dunbar@weareprocogroup.com
