Goldman Sachs and Deutsche Bank Deploy AI Agents in Trade Surveillance

Banks are experimenting with a new form of artificial intelligence, known as agentic AI, that goes beyond simple keyword scanning or pre-defined rules. Rather than depending solely on static alerts, some trading desks are starting to implement systems capable of reasoning through patterns in real-time and highlighting behavior that may require human attention.

Bloomberg detailed how Goldman Sachs and Deutsche Bank are exploring or implementing so-called "agentic" AI tools for trade surveillance. The objective is to enhance oversight of orders and trades by deploying software agents that can analyze activity as it occurs and spot patterns indicative of potential misconduct.

Adaptive agents

Major banks utilize automated surveillance systems to monitor trading activity, which often depend on predetermined rules: if a trade surpasses a specific size, deviates from a benchmark, or matches a known risk pattern, it triggers an alert. Compliance teams then manually review these cases.

The difficulty lies in the scale and complexity. Modern markets produce enormous amounts of data across asset classes, time zones, and trading venues. Static rules can lead to a high number of false positives, while more nuanced forms of manipulation may not fit established patterns.

According to Bloomberg, the newer agentic systems are designed to move beyond this approach. Instead of merely matching trades against a checklist, these AI agents examine trading behavior using multiple signals, compare it with historical activity, and detect unusual combinations of actions.

These tools are not intended to replace compliance officers. Rather, they serve as an extra layer of monitoring, bringing to light cases that merit closer human examination.

Deutsche Bank’s work with Google Cloud

Bloomberg reported that Deutsche Bank is collaborating with Google Cloud to develop AI agents capable of monitoring trading activity. The system is built to analyze large volumes of order and execution data and flag anomalies almost in real time.

The bank has been expanding its AI initiatives over recent years, and this surveillance project illustrates how financial institutions are applying generative and large language model technology beyond simple chat interfaces. Here, the AI is not answering customer queries but analyzing both structured and unstructured data streams related to trading behavior. These AI agents can help pinpoint "complex anomalies" in orders and trades, suggesting the system may evaluate relationships between trades, timing, market conditions, and trader history rather than isolated events.

Human compliance staff continue to be responsible for reviewing flagged cases and deciding whether further action is necessary.

Goldman Sachs’ agentic AI strategy

Goldman Sachs is also exploring the use of agentic AI for surveillance, as reported by Bloomberg. The bank has made significant investments in AI within its trading and risk systems in recent years, and this initiative appears to extend that work into compliance.

The focus, according to the report, is on deploying AI agents that can operate with some autonomy in scanning for signs of misconduct. The system may identify patterns that do not fit a specific rule but still appear unusual.

For regulators, the advantage is clear: earlier detection can help reduce market harm and reputational risk. For banks, there is also an operational benefit. Compliance departments are under pressure to manage large volumes of alerts while upholding strict oversight standards. Tools that can reduce noise without compromising scrutiny are likely to gain traction.

Why “agentic AI” matters

The term "agentic AI" refers to systems that can take goal-oriented actions rather than simply responding to prompts. In practice, this means the software can decide which data to examine next, compare multiple signals, and escalate findings without constant human intervention. In a trading environment, this could involve monitoring order flows, price movements, communications metadata, and historical behavior to determine whether activity aligns with normal patterns.

This does not mean the system makes disciplinary decisions on its own. Financial institutions operate under strict regulatory frameworks, and accountability remains with human supervisors. The agent's role is to identify and organize information more effectively than static systems can.

Part of a wider compliance shift

What seems new is the application of more advanced generative AI architectures to internal control functions.

Regulators in the US and Europe have encouraged firms to improve their monitoring of market abuse and manipulation. While rules do not specifically require agentic AI, they do mandate that firms maintain effective systems and controls. If AI tools can help meet this standard, adoption is expected to increase.

At the same time, using AI in compliance raises its own questions. Banks must ensure that models are explainable, free from bias, and able to withstand regulatory scrutiny. Model governance, data security, and audit trails remain critical concerns.

What changes for the industry

If agentic surveillance tools prove effective, they could transform how compliance teams operate. Instead of sifting through numerous simple alerts, staff may devote more time to evaluating complex cases identified by AI agents.

This shift would not eliminate the need for human judgment. However, it could change where human effort is concentrated. In markets where speed and data volume continue to grow, the ability to analyze patterns in real time is increasingly difficult to achieve with rule-based systems alone.

See also: Mastercard’s AI payment demo points to agent-led commerce

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.