Trade surveillance has transformed dramatically over the past two decades. A recent BroadPeak webinar brought together Kendra Fergusson (Manager, Asset Management at TC Energy), Michael Berry (Managing Member at MRB Compliance Consulting and former Director of Compliance at PetroChina), and Vivek Pathak (Co-Founder and COO at BroadPeak) to discuss how surveillance technology has evolved and what firms need to get right.
Why human judgement still matters
Before 2008, surveillance in energy and commodities was largely manual. Teams relied on reports and exception checks, and many risks went unnoticed. The financial crisis changed everything. Regulations like MiFID II and Dodd-Frank forced firms to adopt a more structured approach. They had to capture more data, monitor for manipulation, and prove they could detect abuse.
The next major shift came with technology adoption. Initially, firms were skeptical about using the cloud for sensitive trading data. Over time, however, the cloud became the default infrastructure. It reduced costs, simplified system management, and made it easier to handle large volumes of data. This shift enabled more cross-asset monitoring and opened the door to big data approaches.
With larger datasets in place, firms began experimenting with artificial intelligence and machine learning. The promise was compelling: faster analysis, better detection of unusual behavior, and more efficient monitoring. However, there are limits. AI can help process information and highlight patterns, but it should not replace human judgment.
The AI said so
Artificial intelligence can process billions of records, detect correlations, and reduce false positives. What it cannot do is explain itself in a way regulators accept. Compliance teams cannot tell regulators “the system flagged it.” Regulators expect a clear, auditable rationale.
AI should be treated as an assistant. It highlights anomalies and saves time on routine tasks, but it cannot replace human oversight.
Kendra noted that AI works well for speeding up processes and handling data, but compliance officers must remain responsible for interpretation and decision-making. A regulator will not accept “the AI said so” as an answer.
Michael pointed out that AI may produce results that are hard to trace back to the rules regulators require. This is why human oversight remains essential. Compliance professionals know the context, can interpret the findings, and can defend decisions if questioned. Vivek added that firms should integrate AI insights with human expertise to make them actionable. AI enhances surveillance, but people remain accountable.
Get the data right
Advances in cloud and data management have made modern surveillance possible. Firms can now bring together data from exchanges, OTC trades, and bilateral deals. Every speaker agreed that surveillance is only as good as the data beneath it. If data is fragmented, delayed, or inconsistent, alerts will be unreliable. If data is integrated, structured, and real-time, surveillance can be more effective and scalable.
Vivek stated that this is where BroadPeak makes a difference. The platform can integrate data quickly and flexibly. It connects to more than 30 exchanges and delivers structured data in real time. This makes it possible to include physical, OTC, and listed trades in a single view. Without this foundation, even the most advanced algorithms fail.
Technology alone isn’t enough
Technology matters, but people remain at the core of effective programs. Systems can generate alerts, but it is analysts who provide the context that determines whether activity is meaningful or benign. Compliance professionals play an equally critical role in translating complex technical or market findings into clear explanations regulators can understand. Michael Berry emphasized that this ability to distill and communicate complexity is one of the most valuable skills in compliance today.
Vivek noted that strong surveillance outcomes come from collaboration across the business. Compliance cannot operate in a vacuum; it must work closely with trading desks, IT, and risk teams to understand both the technical mechanics of trading and the commercial drivers behind it. This collaboration ensures that alerts are not only investigated but also interpreted in a way that reflects real-world trading behavior.
Ultimately, effective programs are built on more than just tools and data feeds. They require a mix of legal knowledge, market insight, risk awareness, accounting expertise, and data analysis. When these skills are combined, blind spots shrink, surveillance becomes more proportionate to the business, and compliance shifts from being reactive to actively shaping the control environment.
Avoid rushing into new technology
Some firms look at new surveillance technology as a quick fix. They hope it will remove pressure from compliance teams. The panel warned against rushing in. If you cannot explain or unwind the system manually, you should not rely on it. Surveillance tools must remain transparent and controllable.
The evolution of trade surveillance has been driven by better data, more powerful technology, and clearer regulatory expectations.
The foundation of effective surveillance has not changed: it still requires skilled professionals who understand markets, regulations, and trading behavior. Technology amplifies their capabilities, it does not replace them.
