Artificial intelligence and blockchain are already being used in the banking industry. Two PNC Bank experts share their thoughts on where the technologies are today and where they are headed.
— By Sharon Ross
The work going on in the back office of PNC Bank is the present and the future of the integration of technology in the banking industry. The complexity cannot be overstated as people like Matthew Shenk, Vice President, Director of AI and Intelligent Automation at PNC and Christopher Ward, Executive Vice President, Head of Treasury Management Product Management and Operations at PNC focus on the best current use and the future applications of advanced and still emerging technologies like Artificial Intelligence (AI), machine learning (ML), and blockchain at PNC Bank.* These are transformative technologies destined to produce a new generation of in-house and global processes, and new consumer and business products. AI, ML, and blockchain are the continuation of the 20-year trend of digitization or technification of the banking business. The difference is that it will not take another 20 years for the newest technologies to radically change how transactions are managed by the bank, consumers and business clients take care of their financial affairs, and customer services are delivered.
Different Technologies Doing Different Jobs
The first thing to understand is that Artificial Intelligence, partnered with machine learning, and blockchain are different technologies that perform differently. As Matthew explains Artificial Intelligence, "It is information technology that enables us to build software programs that mimic the execution of human analytical and/or manual operations." There are many applications of AI, some of which are mimicking human intelligence to carry out tasks in a smart way like the human brain would process information, but some are not. The applications do not have to be programs that are intelligent in and of themselves. They are programs that just function and learn, like moving data around or completing processing tasks. The programs are performing manual tasks and do not require much intelligence. It is more along the lines of software memorizing a process. The program is written so that humans are not needed in the process.
Machine learning is a mechanism for generating a machine model that derives insights on top of data. In most cases, a program is written that essentially fits itself to the data in order to predict some predefined output. Though often used interchangeably, AI is a much broader application. Think of machine learning as a subset of AI. Chris adds that blockchain is "just a different way to store data that makes it possible for disparate entities to work together and share and store information in a more secure way than possible today."
AI enables one thing and blockchain enables another, but blockchain and AI could be used in concert with one another. The two technologies today are deployed separately, but in the future they may be deployed together. Blockchain is a distributed way of storing information that is certified in a particular way. AI is put on top of that information in order to do something with it. For example, AI-enabled smart contract technology fits on top of blockchain which enables the automation of business processes (i.e., configure contracts), and the validation and storage of transactions between parties. The Australian Stock Exchange is currently in the process of replacing their infrastructure with blockchain technology to make it easier to track stocks, bonds, derivatives, and other financial products that are traded on the exchange.
Competition and Cooperation
At PNC, Christopher and Matthew believe that AI and blockchain have interesting implications for banking industry processes. Banks exchange a variety of products and services all over the world, and AI potentially offers more security. AI could provide more security around global trade activities, trade letters of credit, invoicing, and a number of customer documents. Blockchain already enables the exchange of the documents by financial institutions in a secure manner, so AI could add another layer.
AI will help the bank execute more efficiently and effectively many of the same processes that are used today, like fraud detection, anti-money laundering surveillance, and monitoring of other operational-oriented processes. The technology will play a support role in these processes by automating activities, like the assessments of transactions generated by fraud detection models that human beings currently complete. Humans will be freed up to focus on analytical activities, such as analyzing machine-generated information and determining if it warrants reporting to regulators. AI can also help the bank potentially develop new capabilities and new products and services for customers or help employees of the bank offer insights to customers to help them be more adept at making decisions.
AI and machine learning, and blockchain use in the banking industry is still relatively new, creating a current scenario in which banks are moving at different paces of adoption. A lot depends on the primary business of the bank. For example, a bank in a large capital market will focus on blockchain because of the transaction security and visibility it offers. PNC is focused on money movement and cash transactions, so is not focused on a single technology. "There are things we will be doing from a competitive perspective, and things we will be doing to improve the bank infrastructure," explains Christopher. For example, consumers want to be able to use a single credit card or bank card around the globe and be able to go to any bank. Banks have a history of not talking to each other, and AI and blockchain will enable cooperation. It creates an interesting situation in which the new technologies will be applied to improve competitive status and promote cooperation.
First Mover, Fast Follower, or Laggard
The quest to be first mover in the application of AI in banking depends on the specific business problems addressed. The reality is that AI and blockchain can help financial institutions achieve their strategic priorities and transform customer experiences. Matthew explains, "In some areas, like automating processes, it doesn't really matter if a bank is a first mover, fast follower, or laggard in terms of AI applications. What matters is meeting strategic goals. Some may be ahead in automating processes but are likely to end up in a similar place." He further explains that the one area of difference is in the application of AI and machine learning in terms of data. The banks that leverage their data to derive insights about customers will probably create a sustainable competitive advantage. Banks that are first movers in developing products that improve customer experiences may have a meaningful differentiation strategy that enables capturing more market share and developing customer relationships that did not exist before.
The reality is that AI and blockchain can help financial institutions achieve their strategic priorities and transform customer experiences.
Customer interactions with banks are becoming digital first interactions through websites or mobile devices. One of the major transformations already taking place is the use of AI and machine learning to improve the responsiveness to customers accessing the bank through digital interface. Historically for banks, the middle and back office response times have been slower than desired. AI and machine learning will help PNC Bank develop a more seamless, faster, and transparent experience for the customer. As Christopher sees it "The technology is going to enable better customer service because AI can find a problem or automatically fix a problem for customers. Blockchain's job will be to help with all the documentation. Together, AI and blockchain will take the friction out of complex customer transactions."
Two Buckets of AI
At PNC Bank, Christopher thinks of AI as being in two buckets. One is the robotics side which is the mechanization of human activities to achieve greater efficiency and effectiveness. A machine can take a 30-minute job and complete it in two minutes. AI can look at a customer email, notify the customer it was received, start processing the case, and help find a resolution. AI can see patterns with customers as to how they are doing things and make recommendations to improve efficiency. The second bucket is the processing of data which is likely to mean assessing millions of pieces of data to do things like minimize risks around fraud or regulatory non-compliance.
One of the risks concerns the models themselves. The models put in place must be stable, consistent, and secure. They must be well-controlled and well understood. For example, when using machine learning, the developers must have a deep understanding of what is driving the model. Do they know the right data is used to derive decisions; the application of the technologies is right; and biases are not formalized in the data and processes? However, robotics does minimize an important risk – human error. An AI model can do the work of multiple layers of checking currently completed by humans.
Data is the foundation of all the new technologies, but Matthew explains that just collecting data is not enough. "Data is important, but we must also understand the context around the data and the business problem we are trying to solve using that data. The massive amount of data the bank has today must be accumulated in a way going forward that is oriented around the customer and the customer relationship." It is about deriving insights to help consumers and wholesale customers better manage their financial lives.
The Role of Technology Providers
What can suppliers, especially innovative diverse suppliers, expect going forward and how can they benefit from the technological transformation taking place in the banking industry? Christopher believes the real opportunity lies in preparing better solutions for regulatory compliance. The financial services industry uses the term 'regtech' for the sector of vendors making solutions around banking regulations. The other major opportunity is developing technologies that make currently cumbersome systems, like procure-to-pay, easier.
Christopher and Matthew have been focusing on the need for vendors to be more agile rather than following the traditional path of delivering software releases once a year. An agile vendor regularly delivers new products that take advantage of the new technologies to create a whole new way of doing things. Machine learning, for example, is excellent for finding patterns and abnormalities that a human cannot detect. Using AI and ML to prevent problems like fraud is mission critical for PNC bank. A vendor simply cannot be in the fraud space and not be applying AI and ML, and this applies to any new tech-based problem solving solution.
PNC Bank is always looking for women- and minority-owned suppliers that can help the company meet technology needs with value-added innovative products and services or that can supply a variety of non-technical goods and services. In addition to state-of-the-art technology suppliers, PNC Bank is always seeking diverse vendors who can provide corporate services, lending/data services, marketing services, employee-related services, realty services and banking services. There is a well-designed Supplier Diversity Program website that provides a wealth of information on how the program works, the types of products and services sought, program participation requirements, and a number of informative resources that includes upcoming events. Diverse suppliers are encouraged to visit the website where they can find the Supplier Diversity Program links and information. (visit: https://www.pnc.com/en/about-pnc/corporate-responsibility/supplier-diversity.html.)
Getting it Right
In January 2019, PNC became one of the collaborators in a group consisting of Aetna, Anthem, Health Care Service Corporation, and IBM. The collaboration is working on designing and creating a network using blockchain to improve interoperability and transparency in the healthcare industry. The goal is to develop an inclusive blockchain secured network in a shared environment. Healthcare companies using the network can build, share, and deploy solutions that drive digital transformation. The healthcare industry has been striving to develop a system that improves patient care by enabling sharing of critical data, but progress has been slow. "Blockchain," says Christopher, "is the foundation of the health utility network because data can be stored in a way that gives permission for two parties to exchange information, know a transaction is complete, and know the transaction is successfully written in the blockchain."
The key to successful use of any technology now or in the future at PNC Bank is whether it is safe, sound, and secure around customer transaction. As wonderful as AI and blockchain are in terms of creating new opportunities for new products and services, and enhanced customer services, no assumptions can be made about safety, soundness, and security. Vulnerabilities need addressing. Controls must be put in place. Technologies must be as secure as possible. They must also be deployed in the right way with the right controls and with the right access rights and to the right parties. The people like Christopher and Matthew who have undertaken these enormous challenges are to be applauded for their willingness to do what most of us could not do – make machines work with us and not just for us.