Is there a place for robotics and AI in investment banking?
Retail and corporate banking already benefit from the use of robotics in processes and this should naturally offer similar benefits to investment banks in terms of efficiency savings, new opportunities and costs, but is this necessarily the case?
What are robotics?
The use of automation is nothing new in financial services, and banking like other industries clearly understand the benefits of technological innovation and importance of digital transformation with Chief Digital Officers forming part of the C suite in most organisations. Robotic Process Automation (“RPA”), relies on the integration of software automatons to existing software without the requirement to perform IT transformation or create new coding. Robotics, often referred to as artificial intelligence (“AI”), essentially denotes gathering and processing of knowledge by automatons, effectively synthetic intelligence. They operate through processing large volumes of data, analyzing patterns and identifying trends in order to make decisions.
Rather than with traditional IT solutions which rely heavily on coding and logic, AI can operate via cloud technology where information/ intelligence is stored and can be accessed via search functions. Vast volumes of data can be interrogated in this way, significantly more than through coding language.
Middle and back office operations benefit significantly from the removal of manually intensive tasks and processes around reconciliations, settlements and operations which involve voluminous data handling E.g. mortgage portfolio or data tape analysis. In the case of The Bank of New York Mellon, this can be exemplified where robotics is being used to perform account reconciliations, and significantly in trade settlement to match trades which are unsettled.
This process of matching would normally have been performed manually by individuals searching through the customer’s database and trying to reconcile the unmatched trade tickets against external database tickets.
“It takes a human five to 10 minutes to reconcile a failed trade. A bot can do it in a quarter of a second.”1 -
according to Doug Shulman, senior executive vice president at BoNY.
This introduces scalability as the same approach can be replicated across similar requirements in other businesses. Also, the removal of this more manual task from humans more client centric work can be focused on.
Opportunities in investment banking
Front office businesses are already making use of robotics in decision making tasks for clients. Goldman Sachs for example use “Kensho”, a financial robotics technology provider, to recommend trading strategies for clients in response to search based criteria and questions. Operating via a search bar, analysts and sales people are able to at a click of a button expediently yield analysis and strategies by entering the leanest of information. The bank clearly envisages long term benefits evidenced by itself having invested $15mn in Kensho. 2
Robotics benefit the client through expedient delivery of information and recommendations and from the bank’s perspective by enhancing their client offering, promoting technological innovation, scalability, and efficiencies from a time saving perspective for investment decisions which could be more lucrative. Clearly, considerable workload and functionality has been removed from the front office though there are still opportunities for the sales people and analysts to be involved in the process, however, the roles will become more technologically oriented with data science expertise playing a bigger part. Robotics will be changes to work demands for humans and their skillsets will have to change. The changes do not remove roles in developing sales, there will still be a need commercially with respect to front office demands and requirements to generate new business.
Algorithmic trading has been employed in many investment banks (for some time now) and is an obvious example of AI being exploited to execute significant volumes of trades at pre-set intervals and limits at incredible speeds. Whilst this has resulted in the notable loss of headcount in terms of pure traders, it has enabled the creation of roles for programmers and data specialists.
Robotics and AI can form part of the ever increasing remit of the Chief Data (and Digital) Office coverage, and can create additional roles in data science, software management, technology and marketing as a few examples.
Regulatory and compliance divisions gain from the synthesis of large volumes of E-mails, reports, communications and documentation to identify trends and when limits have been breached for example. Assimilating information in this way introduces scalability in terms of time saved by FTEs reviewing legal documentation.
Resistance from regulators?
There is still scepticism and some adversity from the regulatory bodies as it is hard for them to conceive that such fundamental and sizeable decisions are being made via an automaton rather than a human and the risk that this potentially poses. There are questions considered on the reliability of information furnished by robots to investors, in the eyes of regulators there is perhaps scepticism on the use of robotics to determine investment decisions due largely in part to the inability to effectively understand fully how this is being managed.
“If the regulator fully understood what the computerised trader was doing, it wouldn’t be legal” 3 – Eric Hunsader from US data firm, Nanex.
There are also legal considerations regarding the decisions being taken by robots and how they can be policed. Further, is the credibility of information being disseminated to customers more reliable than those provided by seasoned and experienced individuals? IBM Watson is a provider of robotics and is being used to deliver investment advice in wealth management, offering scalability in terms of products potentially such as residential mortgages, but is this necessarily a solution for investment banking clients wishing to make investments into large potentially million pound portfolios without engagement with a human?
Robotics and AI have a role in investment banking clearly open for significant expansion beyond reconciliation and manual processes in support functions. Robotics offer further advantages to significantly expedite and enhance decision making and analysis of information. There are impacts on human employment in terms of reduction in man hours required to perform these tasks, though there are also opportunities to be gained through the creation of new and changing functions. Specialisation of skills in data science and corporate technology are areas which will definitely be in significant demand and the CDO will have an increased role in facilitating these changes in investment banks. Removal of more onerous/ time intensive tasks also enables more focus on commercial activity in the front office to develop new business.
Beyond the financial outlay another consideration and drawback is the risk of robotics providing investment advice. As with investment advice from a human, the recommendation is not a guarantee, just because an automaton has delivered the recommendation it doesn’t necessarily mean it is correct, the usual risks obviously still apply with the investment. Just as with a child, you still need people to teach and provide the inputs to develop and if the quality of teaching isn’t strong then the output is less beneficial, so expertise is required to provide this education. Additionally, the loss of human interaction for some investors may be discouraging, however, not moving with technology means you risk getting left behind.
One thing is clear, not everything is able to be performed by robotics, and investment banking still has a need for core functions such as legal, structuring and engineering of deals and in trade negotiations, so there is solace that humans still offer value, for now at least….