Optimizing platforms offers customers and stakeholders a better way to bank

When it comes to banking, whether it’s personal, business, or private, customer experience is everything. Building new technologies and platforms, employing them at scale, and optimizing workflows is especially critical for any large bank looking to meet evolving customer and internal stakeholder demands for faster and more personalized ways of doing business. Institutions like JPMorgan Chase are implementing best practices, cost efficient cloud migration, and emerging AI and machine learning (ML) tools to build better ways to bank, says Head of Managed Accounts, Client Onboarding and Client Services Technology at J. P. Morgan Private Bank, Vrinda Menon.

Menon stresses that it is critical that technologists stay very focused on the business impact of the software and tools they develop.

“We coach our teams that success and innovation does not come from rebuilding something that somebody has already built, but instead from leveraging it and taking the next leap with additional features upon it to create high impact business outcomes,” says Menon.

At JPMorgan Chase, technologists are encouraged, where possible, to see the bigger picture and solve for the larger pattern rather than just the singular problem at hand. To reduce redundancies and automate tasks, Menon and her team focus on data and measurements that indicate where emerging technologies like AI and machine learning could enhance processes like onboarding or transaction processing at scale. 

AI/ML have become commonplace across many industries with private banking being no exception, says Menon. At a base level, AI/ML can extract data from documents, classify information, analyze data smartly and detect issues and outliers across a wide range of use cases. But Menon is looking to the near future when AI/ML can help proactively predict client needs based on various signals. For example, a private banking client that has recently been married may ask their bank for a title change. Using the client’s data in context and this new request, AI/ML tools could proactively help bankers identify additional things to ask this client, such as need to change beneficiaries or the possibility to optimize taxes by considering jointly filed taxes.

“You have an opportunity to be more proactive and think about it holistically so you can address their needs before they even come to you to ask for that level of engagement and detail,” says Menon.

This episode of Business Lab is produced in association with JPMorgan Chase. 

Full transcript

Laurel Ruma: From MIT Technology Review. I’m Laurel Ruma and this is Business Lab, the show that helps business leaders make sense of new technologies coming out of the lab and into the marketplace.

Our topic today is investing in building great experiences. A number of people benefit from enterprise investment in emerging and new technologies, including customers who want better, faster, and newer ways of doing business. But internal stakeholders want the same investment in better tools and systems to build those fast and new ways of doing business. Balancing both needs is possible.

Two words for you: optimizing platforms.

Today we’re talking with Vrinda Menon, the chief technology officer of Managed Accounts, Client Onboarding and Client Services at JPMorgan Private Bank.

This podcast is produced in association with JPMorgan Chase.

Welcome, Vrinda.

Vrinda Menon: Thank you so much, Laurel. I’m looking forward to this discussion.

Laurel: Great. So, let’s start with how often people think of JPMorgan Chase. They likely associate the company with personal banking, ATMs and credit cards, but could you describe what services the private bank provides and how operations and client services have evolved and transformed since you began your role at JPMorgan Chase?

Vrinda: Sure. JPMorgan Chase indeed does far more than personal banking, credit cards and ATMs. The private bank of JPMorgan Chase is often referred to as the crown jewel of our franchise. We service our high net worth clients and ultra-high net worth clients across the globe. We provide them services like investment management, trust and estate planning, banking services, brokerage services, customized lending, etc., just to name a few. And in terms of what has transformed in the recent years since I joined, I would say that we’ve become far more tech savvy as an organization, and this is thanks and no small measure to new leadership as well in operations and client services. I think three things have changed very dramatically since I’ve joined. The first is culture. In my first few months, I spent a week doing the job of an operations analyst. And in doing that I started to understand firsthand the painful manual work that people were subject to and feeling like they did not have the permission to have things changed for them.

But working off that and actually connecting with a lot more people at the ground who are doing these types of activities, we worked with them to make those changes and make them see light at the end of the tunnel. And then suddenly the demand for more change and demand for more automation started building as a groundswell energy with support from our partners in operations and services. Now, routine, repetitive, mundane, mind-numbing work is not an option at the table. It’s become a thing of the past. And secondly, what we’ve done also is we’ve grown an army of citizen developers who really have access to tools and technologies where they can do quick automation without having to depend on broader programs and broader pieces of technology. We’ve also done something super interesting, which is, over the past three years we’ve taken every new analyst in the private bank and trained them on Python.

And so, they’ve started to see the benefits of doing things themselves. So, culture change I think has been one of the biggest things that we’ve achieved in the past few years since I joined. Second, we built a whole set of capabilities, we call them common capabilities. Things like how do you configure new workflows? How do you make decisions using spreadsheets and decision models versus coding it into systems? So,  you can configure it, you can modify it, and you can do things more effectively. And then tools like checklists, which can be again put into systems and automated in a few minutes, in many cases. Today, we have millions of tasks and millions of decisions being executed through these capabilities, which has suddenly game-changed our ability to provide automation at scale.

And last but not least, AI and machine learning, it now plays an important role in the underpinnings of everything that we do in operations and client services. For example, we do a lot of process analytics. We do load balancing. So, when a client calls, which agent or which group of people do we direct that client call to so that they can actually service the client most effectively. In the space of payments, we do a lot with machine learning. Fraud detection is another, and I will say that I’m so glad we’ve had the time to invest and think through all of these foundational capabilities. So, we are now poised and ready to take on the next big leap of changes that are right now at our fingertips, especially in the evolving world of AI and machine learning and of course the public cloud.

Laurel: Excellent. Yeah, you’ve certainly outlined the diversity of the firm’s offerings. So, when building new technologies and platforms, what are some of the working methodologies and practices that you employ to build at scale and then optimize those workflows?

Vrinda: Yeah, as I said before, the private bank has a lot of offerings, but then amplify that with all the other offerings that JPMorgan Chase, the franchise has, a commercial bank, a corporate and investment bank, a consumer and community bank, and many of our clients cross all of these lines of business. It brings a lot of benefits, but it also has complexities. And one of the things that I obsess personally over is how do we simplify things, not add to the complexity? Second is a mantra of reuse. Don’t reinvent because it’s easy for technologists to look at a piece of software and say, “That’s great, but I can build something better.” Instead, the three things that I ask people to focus on and our organization collectively with our partners focus on is first of all, look at the business outcome. We coach our teams that success and innovation does not come from rebuilding something that somebody has already built, but instead from leveraging it and taking the next leap with additional features upon it to create high impact business outcomes.

So, focusing on outcome number one. Second, if you are given a problem, try and look at it from a bigger picture to see whether you can solve the pattern instead of that specific problem. So, I’ll give you an example. We built a chatbot called Casey. It’s one of the most loved products in our private bank right now. And Casey doesn’t do anything really complex, but what it does is solves a very common pattern, which is ask a few simple questions, get the inputs, join this with data services and join this with execution services and complete the task. And we have hundreds of thousands of tasks that Casey performs every single day. And one of them, especially a very simple functionality, the client wants a bank reference letter. Casey is called upon to do that thousands of times a month. And what used to take three or four hours to produce now takes like a few seconds.

So, it suddenly changes the outcome, changes productivity, and changes the happiness of people who are doing things that you know they themselves felt was mundane. So, solving the pattern, again, important. And last but not least, focusing on data is the other thing that’s helped us. Nothing can be improved if you don’t measure it. So, to give you an example of processes, the first thing we did was pick the most complex processes and mapped them out. We understood each step in the process, we understood the purpose of each step in the process, the time taken in each step, we started to question, do you really need this approval from this person? We observed that for the past six months, not one single thing has been rejected. So, is that even a meaningful approval to begin with?

Questioning if that process could be enhanced with AI, could AI automatically say, “Yes, please approve,” or “There’s a risk in this do not approve,” or “It’s okay, it needs a human review.” And then making those changes in our systems and flows and then obsessively measuring the impact of those changes. All of these have given us a lot of benefits. And I would say we’ve made significant progress just with these three principles of focus on outcome, focus on solving the pattern and focus on data and measurements in areas like client onboarding, in areas like maintaining client data, et cetera. So, this has been very helpful for us because in a bank like ours, scale is super important.

Laurel: Yeah, that’s a really great explanation. So, when new challenges do come along, like moving to the public cloud, how do you balance the opportunities of that scale, but also computing power and resources within the cost of the actual investment? How do you ensure that the shifts to the cloud are actually both financially and operationally efficient?

Vrinda: Great question. So obviously every technologist in the world is super excited with the advent of the public cloud. It gives us the powers of agility, economies of scale. We at JPMorgan Chase are able to leverage world class evolving capabilities at our fingertips. We have the ability also to partner with talented technologies at the cloud providers and many service providers that we work with that have advanced solutions that are available first on the public cloud. We are eager to get our hands on those. But with that comes a lot of responsibility because as a bank, we have to worry about security, client data, privacy, resilience, how are we going to operate in a multi-cloud environment because some data has to remain on-prem in our private cloud. So, there’s a lot of complexity, and we have engineers across the board who think a lot about this, and their day and night jobs are to try and figure this out.

As we think about moving to the public cloud in my area, I personally spend time thinking in depth about how we could build architectures that are financially efficient. And the reason I bring that up is because traditionally as we think about data centers where our hardware and software has been hosted, developers and architects haven’t had to worry about costs because you start with sizing the infrastructure, you order that infrastructure, it’s captive, it remains in the data center, and you can expand it, but it’s a one-time cost each time that you upgrade. With the cloud, that situation changes dramatically. It’s both an opportunity but also a risk. So, a financial lens then becomes super important right at the outset. Let me give you a couple of examples of what I mean. Developers in the public cloud have a lot of power, and with that power comes responsibility.

So, I’m a developer and my application is not working right now because there’s some issue. I have the ability to actually spin up additional processes. I have the ability to spin up additional environments, all of which attract costs, and if I don’t control and manage that, the cost could quickly pile up. Data storage, again, we had fixed storage, we could expand it periodically in the data centers, but in the public cloud, you have choices. You can say data that’s going to be slowly accessed versus data that’s going to be accessed frequently to be stored in different types of storage with different costs as a result. Now think about something like a financial ledger where you have retention requirements of let’s say 20 years. The cost could quickly pile up if you store it in the wrong type of storage. So, there’s an opportunity to optimize cost there, and if you ignore it and you’ve not kept an eye on it, you could actually have costs that are just not required.

To do this right, we have to ask developers, architects, and our engineers to not just think about the best performance, the most optimal resilience, but also think about cost as a fundamental aspect of how we look at architectures. So, this for me is a huge area of focus, starting with awareness for our people, training our people, thinking about architecture patterns, solution patterns, tooling, measurements, so that we completely stay on top of this and become more effective and more efficient in how we get to the public cloud. While the journey is exciting, I want to make sure that as we land there, we land safely and optimally from a cost standpoint.

Laurel: And especially in your position, thinking about how technology will affect the firm years and the future is critical. Therefore, as emerging technologies like AI and machine learning become more commonplace across industries, could you offer an example of how you’re using them in the areas that you cover?

Vrinda: Yeah, certainly. And we use AI/ML at many levels of complexity. So let me start with the base case. AI/ML, especially in operations and client services, starts with can I get data from documents? Can I OCR those documents, which is optical character recognition? Can I get information out of it, can I classify it? Can I perform analytics on it? So that’s the base case. On top of that, as you look at data, for example, payments data or data of transactions, and let’s say human beings are scanning them for issues or outliers, outlier detection techniques with AI/ML, they are also table stakes now, and many of our systems do that. But as you move on to the next level of prediction, what we’ve been able to do is start to build up models where say the client is calling. The client has all these types of cases in progress right now. What could they be calling about in addition to this?

The client expressed sentiment about something that they were not happy with two weeks ago. Is it likely that they’re calling about this? Can I have that information at the fingertips of the client service agent so they can look at it and respond as soon as the client asks for something? And think about the next stage of evolution, which is, the client came to us and said, “Change my title because I just got married.” Typically, in a transactional kind of activity, you would respond to the client and fix the title from let’s say, Ms. to Mrs. if that’s what they asked you to do. But imagine if when they came to do that, we said to them, here’s 10 other things that you should possibly think of now that you said you’ve got married. Congratulations. Do you want to address your beneficiaries? Do you want to change something in tax planning? Do you want to change the type of tax calculations that you do because you want to optimize now that you’re married, and you and your spouse could be filing jointly? Again, not that the client would choose to change those things, but you have an opportunity to be more proactive and think about it holistically so you can address their needs before they even come to you asking for that level of engagement and detail.

We also exploit a lot of AI/ML capabilities and in client onboarding to get better data to start to predict what data is right and start to predict risk. And our next leap, I believe strongly, and I’m super excited about this area of large language models, which I think are going to offer us exponential possibilities, not just in JPMorgan Chase, but as you can see in the world right now with technologies like ChatGPT, OpenAI’s technologies, as well as any of the other publicly available large language models that are being developed every single day.

Laurel: Well, it’s clear that AI offers great opportunities for optimizing platforms and transformations. Could you describe the process of how JPMorgan Chase decided to create dedicated teams for AI and machine learning, and how did you build out those teams?

Vrinda: Yeah, certainly. At JPMorgan Chase, we’ve been cultivating the mindset for some years now to think AI-first while hiring people. And we also leverage the best talent in the industry, and we’ve hired a lot of people in our research divisions as well to work on AI/ML. We’ve got thousands, several thousand technologists focused on AI. For me personally, in 2020, during the first months of the pandemic, I decided that I needed to see more AI/ML activity across my areas. So, I did what I called the “Summer of AI/ML,” and this was a fully immersive program that ran over 12 weeks with training for our people, and it was not full-time. So, they would dial in for a couple of hours, get trained on an AI/ML concept and some techniques, and then they would continue that and practice that for the week.

Then we had ideation sessions with our users for a couple of weeks and then a hackathon and some brilliant ideas came out of it. But when I stepped back and looked at this whole thing and the results of it, a few months later, I realized that many of the ideas had not reached the final destination into production. And in thinking a little more deeply about that, I understood that we had a problem. The problem was as follows, while AI is a great thing and everybody appreciates it, until AI becomes ingrained in everybody’s brain as the first thing to think about, there’s always going to be a healthy tension between choosing the next best feature on a product, which is very deterministic. If you say, add this button here or add these features using conventional technologies like Java versus game-changing the product using AI, which is a little bit more of a risk, the results are not always predictable, and it requires experimentation and R&D.

And so, when you have a choice of incremental changes that are deterministic and changes that are more probabilistic, people tend to take the most certain answer. And so, I decided that I needed to build out a focused, dedicated team of data scientists who were just going to obsess about solving problems in the space of data science and embed them across the products that we were building. And now the results are starting to show for themselves because the work they’ve done is phenomenal and the demand on them is growing every single day to the point where I’ve grown the team and the value that they’re providing is also measured and visible to the broader organization.

Laurel: So, in JPMorgan Chase’s client services, customer experience is clearly a driving force. How do you ensure that your teams are providing clients, especially those high-net-worth private clients that have high expectations of service with services that then meet their banking and account management needs?

Vrinda: So, we obsess over customer experience starting from the CEO down to every single employee. I have three tenets for my team. Number one is client experience, the second is user experience, and third is engineering excellence. And they know that a lot of us are measured by how well we service our clients. So, in the private bank specifically, in addition to reviewing our core capabilities like our case management system, our voice recognition systems, our fraud capture systems, all of that, we continuously analyze data received from client surveys, data received through every single interaction that we have with our client across all channels. So, whether it be a voice channel, whether it be emails, whether it be things that the client types in our websites, the places that they access, and our models just do not look at sentiment, they also look at client experience.

And as they look at experience, the things that we are trying to understand are, first of all, how’s the client feeling in this interaction? But more important is client one and client two and client three feeling the same thing about a particular aspect of our process, and do we need to change that process as a result, or is there more training that needs to be provided to our agents because we are not able to fully satisfy this category of requests? And by doing that continuously and analyzing it, and back to the point that I made earlier, by measuring it constantly, we are able to say, first of all, how was the experience to begin with? How is the experience now and after making these changes on these training programs or these fixes in our systems, how is that experience showing? And some of the other things we are able to do are look at experiences over a period of time.

So, for example, the client came to us last year and their experience based on the measurements that we did was at a certain level, they continue to interact with us over a period of months. Has it gone up? Has it gone down? How is that needle trending? How do we take that to superb? And we’ve been able to figure out these in ways that we’ve been able to prevent complaints, for example, and get to a point where things are escalated to the right people in the organization, especially in the servicing space where we are able to triage and manage these things more effectively because we are a high- touch client business, and we need to make sure that our clients are extremely happy with us.

Laurel: Oh yeah, absolutely. And sort of like another phase or idea when we’re thinking about customer experience and customer services, building a workforce that can respond to it. So here we’re going to talk a bit about how we promote diversity, which has been a tenet of your career, and you currently sit on the board of the Transition Network, which is a nonprofit that empowers a diverse network of women throughout career transitions. So, at JPMorgan Chase, how do you grow talent and improve representation across the company? And then how does that help build better customer experience?

Vrinda: Sure, that’s a great question. I certainly am very passionate about diversity, and during the past 15 years of my career, I’ve spent a lot of time supporting diversity. In my prior firm, I was co-head of the Asian Professional Network. Then subsequently for the past three years, I’ve been a board member at the Transition Network, which is all about women in transition. Meaning as they grow out of their careers into retirement and into other stages of life, how do we help them transition? And then here at JPMorgan Chase, I’m the sponsor for what is called the Take It Forward initiative, which is an initiative that supports 15,000 women technologists. JPMorgan Chase, as you know, does a broad range of activities in the area of diversity across all kinds of business resource groups, and we invest a lot of time and energy.

But specifically, for the Take It Forward initiative that I sponsor, it plays a key role in helping these 15,000 women technologists continuously enhance their leadership skills, grow their technical skills, build their confidence, develop networks, learn from senior sponsors and mentors, grow their careers, all of which makes their work experience very enriching. When I hear things like I’m motivated, I get new energy interacting with these senior women, I trust my personal power more. I’m confident to negotiate with my manager for a better role; I feel confident that I can discuss my compensation. It makes me really happy, and especially when they say I stay in JPMorgan Chase because of Take It Forward, it brings tears to my eyes. It’s really one of the most amazing volunteer driven initiatives in this organization. And a lot of people pour in passion, energy, time to make it succeed. And this initiative has won many awards as well externally.

I strongly believe all of these efforts are critical as it changes when people’s experiences change and when they’re happy, what they do becomes that much more effective. And it changes how we work internally, how we present ourselves externally and game changes our business outcomes. I’ve seen that in problem solving meetings. When you evaluate risk and you bring in people from diverse backgrounds, some are more risk averse, some are more risk taking, and you suddenly see that dynamic play out and the outcome becomes much different from what it would’ve been if you didn’t have all those people in the mix. So overall, I strongly believe in this, and I’ve seen it play out in every single firm that I’ve ever worked at. So that’s my take on diversity and how it helps us.

Laurel: Well, it certainly is important work, especially as it ties so tightly with the firm’s own ethos. So Vrinda, looking forward, how do you envision the future of private banking and client management? As we see emerging technologies become more prevalent and enterprises start to shift their infrastructure to the public cloud?

Vrinda: As I mentioned earlier, I see the next set of emerging technologies taking the world on a super exciting ride, and I think it’s going to be as transformational as the advent of the world wide web. Just take the example of large language models. The areas that are most likely to be first disrupted will be any work that involves content creation because that is table stakes for a large language model. As I expand that to my work, the rest of my work, client services and operations and many other areas that require repetitive work and large-scale interpretation and synthesis of information, that’s again, table stakes for large language models.

Expand that now to the next evolution, which is agents that are now the emerging technology with large language models. When agents are provided with a suite of tools and they can use reasoning like humans to decide which tool to execute based on input, that’ll game change the whole thing I was talking about earlier on workflow and task execution and operationally intense activities in the organization. When I look at myself as a software developer in the areas of code generation, code testing, code correction, test data generation, just to name a few, all of those are going to be game changed. So not just the work that our users do in the private bank, but the work that we do as technologists in the private bank. A lot of that is going to game change dramatically.

And then you add on the next level, which is problem solving. Large language models are continuously being trained on all subjects ever known to humans. And that for me is the most fascinating part of this. It’s like hundreds of thousands of brains that are working together on a diverse set of subjects. So, imagine a model that’s been trained on a domain like medicine or aerospace or defense, and then trying to bring all of that brainpower together to solve a problem in finance. That truly to me is the ultimate gold standard of problem solving. We talked about diverse people in the room coming with different experiences but imagine models that have been trained. You suddenly have breadth, depth, range of diverse knowledge that could never have been contemplated at that scale.

And in order to do all of this, obviously one of the key underpinnings is the public cloud and being able to spin up compute as quickly as possible to do complex calculations and then spin it down when you don’t need it, which is where the public cloud becomes super important. So, all I can say in conclusion is I think this is an amazing time to be in technology, and I just cannot wait to see how we further step up our game in the coming months and years, and things are moving almost at the speed of light now. Every single day, new papers get published and new ideas are coming out building on top of some of the exponential technologies that we are seeing in the world today.
 
Laurel: Oh, that’s fantastic. Vrinda, thank you so much for being on the Business Lab today.

Vrinda: Thank you so much, Laurel. It’s my pleasure. I really enjoyed speaking with you and thank you for your thoughtful questions. They were super interesting.

Laurel: That was Vrinda Menon, the chief technology officer of Managed Accounts, Client Onboarding and Client Services at J.P. Morgan Private Bank, who I spoke with from Cambridge, Massachusetts, the home of MIT and MIT Technology Review overlooking the Charles River.

That’s it for this episode of Business Lab. I’m your host, Laurel Ruma. I’m the director of Insights, the custom publishing division of MIT Technology Review. We were founded in 1899 at the Massachusetts Institute of Technology, and you can find us in print, on the web and at events each year around the world. For more information about us and the show, please check out our website at technologyreview.com.

This show is available wherever you get your podcasts. If you enjoyed this episode, we hope you’ll take a moment to rate and review us. Business Lab is a production of MIT Technology Review. This episode was produced by Giro Studios. Thanks for listening.

This content was produced by Insights, the custom content arm of MIT Technology Review. It was not written by MIT Technology Review’s editorial staff.

This podcast is for informational purposes only and it is not intended as legal, tax, financial, investment, accounting or regulatory advice. Opinions expressed herein are the personal views of the individual(s) and do not represent the views of JPMorgan Chase & Co. The accuracy of any statements, linked resources, reported findings or quotations are not the responsibility of JPMorgan Chase & Co.