The recent explosion of data volume and computing power has led to a dramatic re-evaluation of Artificial Intelligence (AI) in several industries including the financial sector. AI’s challenges and limitations are creating a “moving target” problem for leaders: it is hard to maintain a leading edge in something that’s always advancing so fast.
This short article explores the commercial implications of AI, its possible applications for securities servicers and some potential factors of success and challenges for firms applying it.
AI, which first became a formal field of academic research in 1956, is not a single technology but a combination of algorithms leveraging mathematics and computer science to create tools which learn through the use of training data and backpropagation algorithms.
Artificial Intelligence applications include, but are not limited, to Machine Learning (ML), a range of algorithms that can scan large amounts of unstructured data and use statistical analysis to predict an output, while updating outputs as new data becomes available; Natural Language Processing (NLP), Automated Reasoning and Cognitive Robotics (a technology enabling robots to learn from and respond to real-world situations, as opposed to pre-programming them with specific responses to certain expected events).1
AI can also support the human decision-making process by giving additional insights: this sort of collaboration is called Augmented Intelligence.
Within the financial industry, AI is expected to impact a number of key areas in the next few years such as:
Investment analysis: of internal and external data. Asset managers can benefit of useful investment signals while securities servicers can get insights to data on prices, holdings and trading flows, as well as the provisions of data as a service;
Portfolio management: AI can monitor investments, enhance asset allocation and portfolio construction;
Middle office: performance evaluation and risk management by providing a faster exposure to fraud or money laundering;
Investment operations: AI tools have a wide range of potential applications within the back office as, interaction between platforms, automated labour-intensive processes, by identifying, capturing, scrubbing and acting on events such as Redemptions or Dividends. (transfer agency and corporate actions).
Functions such as query handling and automated response to trade queries, predictive analysis to spot transactions that are likely to fail, would improve customer support functions.
ML can teach less advanced robots to enhance straight-through processing (STP) by learning to identify and correct errors while cognitive robotics can achieve improvements in high volume functions such as look ups, data transfer, calculations and reconciliations.
Still, the foundation of any successful Artificial Intelligence strategy depends on understanding of how different tools work and how they can interact with humans and machines.
Data integration: the greater the volume of data, the greater the potential value of AI. Combining data effectively as well as enabling flexible connections between platforms is a challenge.
Data interpretation: different forms of AI can work with different types of data. Machine Learning has the potential work through alternative data sources. Since these systems are “trained” rather than programmed, the various processes often require huge amounts of labelled data to perform complex tasks accurately. An accurate selection and labelling are currently one of the biggest challenges in order to implement an effective data strategy and governance.
The ideal tech-and data-enabled organization of the future needs more than data. It is essential to understand how increasingly powerful tools, particularly those enabled by Artificial Intelligence can overcome the limitations of legacy technology. In the securities services industry, a speaker from a tier 1 institution at the Network Forum recently held in Vienna stated that in just one market, for one year only, there was six billion lines of data to analyse.
It is not surprising that the industry, as a whole, is expected to spend more on AI than any sector in the next few years.2 Furthermore, as the application of AI expands, regulatory requirements could also drive the need for more explainable AI models.
Harnessing the potentials of AI poses unique challenges such as the need of data governance and a number of soft factors such as culture, flexibility and leadership. AI has the potential to reshape institutions, market and the industry as a whole. Its application to business problem solving is growing and, at the same time, concerns about AI’s implications are rising: the impact of AI-enabled automation on the workplace, employment, and society represent some issues of concern as well as a challenge.
Firms of all types need to consider what AI can already do but also how it might develop in the future by creating a strategy around its potential benefits. While some firms are developing AI in-house, others are considering outsourcing significant portions of their AI development to external providers that can offer cutting edge machine learning as well as cloud-based analytic platforms.
Whatever the approach, it’s time to start understanding what is happening at the AI frontier so to position impacted organization to learn, exploit, and possibly even advancing the new possibilities. Timing is key.
1 - https://www.hsbc.com/news-and-insight/media-resources/media-releases/2018/hsbc-human-advantage
2 - https://www.mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/mckinsey-quarterly-2018-number-1-overview-and-full-issue