A two-speed journey: the impact AI and robotics will have on financial markets
In the last two decades new technologies have modernised and transformed the way business is conducted in financial markets and, in particular, the market infrastructure which underpins it. Recent examples include use of the public cloud and several promising distributed ledger technology (DLT) proof of concepts which are now being put into production.
The latest raft of emerging technologies which have already started to have an impact on financial markets are robotic process automation (RPA) and artificial intelligence (AI). This article will consider the current and future impact of these emerging technologies on financial markets and, more specifically, on market infrastructure.
While RPA and AI are often grouped together, the truth is that they serve different purposes and are at very different points in their maturity curves and implementation timelines. For example, while AI requires deeper technology expertise, it may provide more extensive strategic value, whereas RPA operates in human scale settings to automate tasks that people perform manually using hardware such as their keyboards for entering data or producing reports. Furthermore, AI can provide added value to existing automated systems, such as business intelligence applications.
In terms of timelines, RPA has the potential to transform the industry more quickly than other emerging technologies like AI because of its immediate ability to automate traditional processes for small scale and more tactical operations tasks, rather than supporting strategic initiatives. RPA is specifically designed to integrate into existing workflows and interact with existing user interfaces. In comparison, AI is more complex and is likely to have a longer implementation curve. It cannot, for example, be deployed on an “as is” infrastructure. AI is predicated on raw data and raw data is sometimes not available or preserved in organisations. As a result, many changes are required to undergo digital transformation to maximise the potential of AI.
Firms across the industry are already considering where best to apply RPA. Many are evaluating how tasks can be automated, actively assessing how RPA can be used tactically and strategically to deliver operational and client benefits today. Currently, many firms use RPA as a short-term efficiency play while more sophisticated automation approaches, which may take much longer to implement, are being developed.
Looking specifically at AI, it is actually considered an umbrella term which encompasses a wide range of technologies that mimic human intelligence such as machine learning (ML). AI can enable predictive analytics which has the potential to deliver innovative new services. In market infrastructure we expect AI to have significant impact in the area of data analytics, as well as in smart (AI-enhanced) robotics. In many of these systems, AI involves simulating human senses such as language and image recognition.
In fact, at DTCC we recently made enhancements to our Security Issue Database leveraging AI. The enhancements allow us to capture, analyse and distribute mutual fund security data in a better and faster way. Currently, market infrastructures are leveraging their vast amounts of data to identify where AI and ML could be best applied to provide meaningful business hypotheses. For example, cybersecurity is fast becoming a ripe opportunity for the application of AI and ML. These technologies are useful because they allow firms to monitor massive amounts of data related to network traffic and messages, looking for anomalies and suspicious behaviors in real time to fight against cyber threats.
Risk management is another emerging area that is being explored for AI and ML. For the wider financial services industry that includes activities such as AML, fraud, customer service and other front office applications.
We are also seeing the use of ML to recognise similar phrases from operator comments about application and system issues, and thereafter recommend solutions from a knowledge database. With the vast amounts of data within the industry, additional use cases are sure to come, providing new insights to enable even better human decision-making and response. Of note, there can be a material risk of false positives with AI, but that is where humans can be best deployed and leveraged – making sense of the outcomes, identifying true positives and determining appropriate next steps.
In the longer term, we are also likely to see convergence around emerging technologies. For example, DLT is simply a database, but it is a database that enables greater sharing of data and hence, creating a larger pool of data for AI to leverage. Over time, AI could be applied to DLT implementation and there are already examples of this in various stages of planning, proof of concepts and pilots.
That said, and as noted in DTCC’s framework for evaluating risks of fintech and its impact on financial stability (October 2017), fully automated decision-making associated with AI has the potential to introduce systemic risks, especially where the rules and situational dynamics are less well defined. In such cases there is greater need to balance AI and human decision-making, pointing more towards solutions characterised by “augmented intelligence” with a greater reliance on human perspective.
Emerging technologies have the potential to revolutionise existing processes in financial markets according to different time scales. While we are beginning to see the benefits of RPA applied to small scale and tactical operations tasks, the full potential of AI is still not yet fully understood. While undergoing this transformative process, it is essential that the implementation of these technologies is done with risk management and due diligence at the front of mind. Providing these principles are adhered to, we expect to see both RPA and AI successfully leveraged across certain post trade processes, in the near term and mid-longer term, respectively.
By Rob Palatnick, chief technology architect, DTCC
“Currently, many firms use RPA as a short-term efficiency play while more sophisticated automation approaches, which may take much longer to implement, are being developed.” – thanks for bringing this up! Indeed, the maturity stage of RPA equals that of scaling towards enterprise-level.
Automation is enticing – once you start with it, you’d better keep moving forward in the direction of an enterprise-wide use for maximum gain. In fact, given that scalability is among the top benefits of RPA, it comes natural that many CEOs consider adopting it at scale.
Scaling requires a certain degree of foresight, operationalised in terms of medium- and long-term strategic planning. All along the way, potential difficulties must be acknowledged and addressed, if possible before they become actual obstacles. Think employees’ resistance, keeping expectations realistic, deciding upon the right sequence of processes to be automated, figuring out the targeted scope of automation, etc.