Artificial Intelligence, Distributed Ledger Technology, Internet of Things. Are they truly game changers?
We are in the midst of a period of dramatic innovation and change in transaction banking, perhaps like never seen before. There are a number of catalysts for this, including accelerating globalization, growing regulatory demands, and the emergence of the millennial generation in the workplace, to name just a few. But probably no single driver of the current wave of innovation is as significant as the accelerated pace of technological change. New technologies have always emerged to transform or even displace prevailing state of the art technologies and processes, but what seems to be different now is the pace of it – the number of new technologies, business models, and players that are quickly emerging and evolving with the potential to transform transaction banking.
If it is true that necessity is the mother of invention, then what might the current pace of change say about the transaction banking business? Perhaps we have not been keeping pace with our clients’ evolving needs for the solutions and client experience they want. Let’s not lose sight of the fact that banks and other operators have provided functional, reliable, and relatively resilient services for many decades. But in the last decade, maybe even longer, our necessary attention to growing regulatory, risk, cybercrime, and compliance priorities has diverted some of our attention away from innovation. With many of these risk and stability priorities now in good order and ongoing, there is a renewed focus on growth and innovation, with adoption of new technologies as a key lever for success.
But which of these emerging technologies solve a real problem and have true potential to be transformative for our business? As Head of Innovation for BNY Mellon Treasury Services, my team and I are charged with evaluating new ideas, technologies and potential partners with a business lens − to assess the applicability and commercial viability of these opportunities in our payments and trade businesses. So it is from that perspective that I examine here three of the most exciting emerging technologies/platforms: Artificial Intelligence, Distributed Ledger Technology, and the Internet of Things.
Artificial Intelligence (AI)
There are many definitions of AI, but the generally accepted distinction from ordinary software automation like Robotic Process Automation (RPA) is that in AI applications, “the machine learns.” That is, it gets better at performing its function as it observes the outcomes of its recommendations or actions. While various applications of AI have existed for decades, public awareness of AI and its transformative potential has increased dramatically more recently, at least in part due to the growing visibility of IBM Watson in media and commerce. There is also growing fear of the technology and its potential impact on the workforce. While there is validity to most concerns, there is also a lot of misunderstanding of what AI technology can and can’t do at this time.
At this point in time, and likely for the foreseeable future, AI applications require significant structure: the machine needs to be programmed to know where to find the appropriate data, how it is structured, and what task is to be performed using the data. It then provides recommendations or performs prescribed tasks as programmed, based on analysis of the data. Current AI technology cannot perform unstructured, complex thought processes. Human-like “Artificial General Intelligence” (AGI), which would include the ability to contextualize information and apply common sense and reasoning, is many years away. Nevertheless, AI offers a number of exciting opportunities to transform treasury, payment and trade processes today.
AI has already been applied in the form of chatbots, robo-advisers, and even physical robots in consumer banking applications. Similar solutions will eventually be found in more complex and higher value functions, possibly including various treasury management functions such as cash positioning, liquidity management, payment channel optimization, invest/borrow decisions, hedging strategies, etc. These functions require a lot of data and are highly complex to program for, so it will take some time to develop them with the level of sophistication necessary to automate processes and optimize outcomes to maximal client benefit.
More near term, BNY Mellon is exploring the potential to leverage AI to improve our Office of Foreign Assets Control (OFAC) sanctions scanning process. Current technology tools for this process create a high percentage of false positives among flagged items, requiring manual review and delayed processing for payments that turn out to be appropriate for straight through processing (STP). We believe AI technology can be applied to recognize some of the common reasons for unnecessarily flagging, and make appropriate recommendations to the review operator for handling. The tool would learn over time based on the eventual actions taken by the review operator, thus gaining higher levels of confidence in its recommendations. We believe such a solution would improve the efficiency of our payment process, increasing STP rates and, thus, client satisfaction as well.
As noted, AI is already a reality in select banking applications, and we believe this technology holds great potential to transform the transaction banking business including both payments and trade.
Distributed Ledger Technology (DLT, aka blockchain)
While the concept of distributed ledgers was first proposed in the 1960s, it was not until the creation of Bitcoin and its blockchain technology underpinnings that the technology developed wide scale interest and exploration of its transformative potential for a number of financial processes. What started (in the case of Bitcoin) as an attempt to disintermediate central banks, large financial institutions, and financial market infrastructures (FMIs), is now being explored by those very institutions (and many others) as a means to eliminate unnecessary and duplicate processes; reduce costs; and increase efficiency, transparency, and resiliency. Rather than put them out of business, the technology may become an important part of the infrastructure that will continue the viability and relevance of these institutions for the foreseeable future. However, it has also spawned a number of new competitors, especially in the payments space, thus creating some new marketplace dynamics: for consumer payments certainly, but likely also eventually for commercial payments.
Ripple has emerged as a new competitor to the tried and true correspondent banking model for cross-border payments, and continues to expand its network globally. Some other companies have launched cross border payment services that leverage Bitcoin as a bridging currency in their process, thus leveraging DLT as part of their payment process. These developments provided at least some of the motivation for SWIFT to self-assess its process and technology, leading to the launch of SWIFT gpi, which has already produced impressive improvements in the timeliness and transparency of cross-border payments. While SWIFT has not leveraged DLT for its infrastructure at this point, it has shown an interest in considering its potential, leading to a proof of concept that explored a DLT solution to enhance the nostro account reconciliation process.
Given BNY Mellon’s focus on institutional clients and flows, and our role as the largest custodian in the world supporting various FMIs, our current focus on DLT very much aligns with use cases that are relevant to our clients. Many of the world’s FMIs are moving to DLT-based platforms, including NASDAQ (live, private equities), Australian Stock Exchange (ASX), SIX Swiss Exchange (SIX), and Depository Trust & Clearing Corporation (DTCC), while others are exploring it. As these exchanges bring their asset clearing and settlement processes on ledger there is the opportunity also to bring the corresponding financial settlement on ledger, thus creating atomic transactions – where both the asset exchange and payment happen near simultaneously, or neither happens. This could dramatically reduce settlement and counterparty risk, while also reducing collateral needs and costs and improving liquidity management. BNY Mellon has been a participant in the Utility Settlement Coin (USC) initiative, a consortium of 17 banks and FMIs along with Clearmatics (a DLT technology company), that is looking to create digital tokens representing the top fiat currencies and backed on a 1-1 basis by cash deposits at the respective central banks. We believe these utility settlement “coins,” exchanged and tracked on a shared permissioned ledger, can be the means of financial settlement for atomic transactions. We are very excited about the transformative potential and risk reduction opportunity that USC could provide for this and other potential use cases.
There certainly is a lot of hype around DLT, and many of the applications being explored will fizzle for lack of viability or simply because DLT is not the optimal solution to a particular problem or inefficiency. Some of the challenges still to overcome include further development of standards, interoperability, adoption or network effect, privacy concerns, and regulatory uncertainties. However, as demonstrated above, DLT has already been launched for a number of financial market applications, and intense exploration continues for additional high-potential use cases.
Internet of Things (IoT)
Even newer than AI and DLT, is the emerging Internet of Things. As sensors and actuators have become smaller and lighter, they are increasingly being integrated into all sorts of devices as we continue the transformation to a digital economy. These devices also have internet connectivity for communication with owners, businesses, and even other devices, creating a wide array of new opportunities to monitor and control these devices remotely, and even autonomously. There are examples across many industries, from healthcare to manufacturing to transportation to financial services, and many others. And not just for consumer applications − like refrigerators that autonomously replenish consumables − but commercial applications as well, like elevators and freight trains that can predict maintenance needs and self-schedule for such.
Many of the current or envisioned examples include some form of eCommerce, with a service or good provided and a payment to be made for it. Also, IoT is spawning new business models or changing existing models, for example, paying for capacity on a per-use basis rather than owning a depreciating asset, whether a laundry machine or a commercial jet engine. This means new payment and trade flows, financing models, and financial processes, new ways that our clients − whether they be corporates, banks or other institutions – may need our help.
At this point in time, most IoT eCommerce activity is consumer oriented, with cards as the primary rail for payment (including P2P apps). But commercial IoT applications are proliferating, and these will require more sophisticated payment and trade solutions including better services and information for back office financial processes. While it might be early for IoT’s impact to be felt by our clients, we are monitoring this technology platform to understand its transformative potential and implications for transaction banking.
Reality, Not Hype
The three technologies highlighted here – AI, DLT, and IoT – certainly do generate a lot of inflated and, in some cases, unrealistic expectations. But we believe they are very real, albeit with varying expected timeframes for widespread adoption. AI is easiest to deploy, with the least external variables to overcome. IoT is also easy to implement and scale on a use case by use case basis, although countless potential applications will likely emerge over many years. DLT will take the longest to become widely applied, given the network effect that is required for its benefits to be fully realized, and this will take time as standards, interoperability and legal/regulatory hurdles are overcome. But given its potential, we believe there will be strong motivation to get it there.
And a final point: these technologies will likely converge for some use cases, and in some instances already are. DLT can be the book of record for IoT transactions, providing the audit trail and immutability of data that may be needed in some use cases. For example, proof of how many flight miles have been logged on a jet engine; this data needs to be accurate, immutable, and viewable by regulators to ensure transportation safety. Some of the trade use cases that are already being explored combine IoT (temperature sensors in refrigerated cargo containers), DLT (tracking and recording goods through various shipping stages and routes), and AI (taking actions based on potential problems that may emerge with the goods due to inappropriate storage/shipping/handling conditions or timeliness). This will likely have an effect on trade documents and related payment flows. Other convergence opportunities will no doubt emerge. This is just the beginning.
The views expressed herein are those of the authors only and may not reflect the views of BNY Mellon. This does not constitute Treasury Services advice, or any other business or legal advice, and it should not be relied upon as such.
©2018 The Bank of New York Mellon Corporation.