Intelligent Communication Surveillance and the Power of Prediction Boosts Market Integrity

If one aspect of trade surveillance has remained consistent throughout the years, it is that change is constant. Whether driven by regulatory enforcement or environmental factors, surveillance practices are compelled to advance at the same (or faster) pace as trading strategies themselves. And the myriad surveillance challenges presented by transforming communication norms as firms adjusted to pandemic-era remote working environments has continued to evolve – but fortunately so has risk detection and surveillance alert technology.

The explosion in communication channel usage added fresh dimensions of complexity and technological challenges for surveillance program managers – especially concerning firms’ suddenly heavy reliance on video conferencing and rapid adoption of new e-communications platforms. Even early adopters of expanded communication data source coverage used AI processing in alert generation to tackle these challenges and implemented up-to-task detection and monitoring capabilities as quickly as they were brought to market. Since then, the trajectory of communications surveillance tools has continued to evolve and now provides surveillance managers with increasingly powerful capabilities.

The Power of Prediction

The use of AI to generate predictive surveillance alerts is one such strategy that is now a reality. Effective incorporation of predictive processing is a potential game-changer for trade and market oversight efforts. The introduction of new screening factors and contextual dimensions, for example, could dramatically improve accuracy and at last help surveillance managers meaningfully reduce false positives. In addition, these capabilities free human review resources to focus on genuinely critical case reviews which is essential to better protection for markets, firms, and employees.

Predictive processing, deployed in conjunction with proven static detection algorithms, significantly enhances risk detection capabilities. It is important to note that predictive alert technology is meant to be an enhancement, not a replacement, to existing successful detection and parameter settings. By building on a well-honed monitoring program, the integration of predictive scenarios could add significant accuracy to behavioral analysis tools such as employee profiles and risk scores. Successful implementation could expand correlation possibilities in previously unattainable ways. Leveraging enhanced contextual data would improve the usefulness of other unstructured data analysis tools, such as NLP and facial recognition.

Great Power, Great Responsibility

The operational risks of using predictive processing in trade and market surveillance efforts are relatively low in technology. Alerts can be prioritized efficiently by utilizing "auto-close" options or "ring-fencing" alerts into separate queues. However, like any new technology deployment (especially involving AI processing), correct model training and beta testing management is crucial to successful use, and the sensitivity of results to "noise risk" is magnified. Data source disaggregation can also amplify the sensitivity of alert results, so program managers will need to adjust parameters with approved channel options in mind to avoid an onslaught of false positives/false negatives.

Lastly, with ever-more powerful monitoring and alert capabilities in use, boundaries of activities subject to surveillance must be clearly defined and communicated to address both compliance expectations and employee privacy, particularly in the post-Covid era of fungibility between home and office working environments. With such powerful alert generation possibilities and the expanse of data sources used in analysis, compliance managers must ensure employees clearly understand coverage boundaries and are provided with the proper tools to maintain compliance.

The next generation of surveillance alert processing opens a new world of possibilities to promote market integrity and enhance firms’ self-regulation and employee protection. Detecting potential or recent violations before the regulator authority does is always preferable. Moreover, predictive behavioral analysis also carries promising implications for monitoring employee well-being and other operational functions beyond surveillance. Thus, the vision that accuracy and efficiency will increase significantly in 2022 is undoubtedly promising.