UiPath Forward 2024 | Brandon McCormick, State Street & Maneesh Bakshi, Deloitte Consulting LLP

Clips
News
More from UiPath Forward 2024

Brandon McCormick

Managing Director – State Street Alpha

State Street

Maneesh Bakshi

Senior Manager - Investment Management

Deloitte Consulting

Investment management automation sees momentum as AI-driven testing enhances efficiency

Investment management automation is accelerating as firms increasingly rely on AI-driven testing platforms to streamline complex processes and enhance operational efficiency across applications. That’s the case for State Street Corp., which has sought to utilize UiPath Inc. to streamline operations and enhance productivity.State Street’s Alpha platform is a front-to-back platform that utilizes a lot of different applications and services that use various different technologies, according to Maneesh Bakshi (pictured, left), senior manager of investment management at Deloitte Consulting LLP

play_circle_outline Alpha platform overview for investment managers

play_circle_outline State Street's Transition to Automated Testing: Improving Efficiency and Quality Assurance Processes

play_circle_outline Impact of automation on testing roles and career advancement

play_circle_outline Ideal future state of automation and generative AI integration

Info
Transcript

Brandon McCormick, State Street & Maneesh Bakshi, Deloitte Consulting LLP

Brandon McCormick

Managing Director – State Street Alpha State Street

Maneesh Bakshi

Senior Manager - Investment Management Deloitte Consulting

During a discussion on automation journeys, Brandon McCormick from State Street and Manish Bakshi from Deloitte Consulting talked about their use of the Alpha platform for testing. The platform combines various technologies for end-to-end testing, with UiPath being their tool of choice for its rapid evolution. Brandon's team focuses on testing implementations and partners with clients for successful testing journeys. They highlighted how automation allows testers to focus on problem-solving and improves quality. The importance of governance, skills, and commo... Read more

explore Keep Exploring

What is an overview of the Alpha platform? add

What are the responsibilities and challenges of the testing team in a large, complex project with multiple global teams and client implementations? add

What are some of the reasons why employees are excited about moving towards automation and using tools like UiPath? add

What does the ideal future state look like if everything works perfectly for State Street's journey of automation, including generative AI? add

bolt Powered by CUBE AI

Brandon McCormick, State Street & Maneesh Bakshi, Deloitte Consulting LLP

search

Rebecca Knight

>> Good afternoon everyone, and welcome back to theCUBE's Live coverage of UiPath FORWARD 2024 here in Las Vegas. I'm your host, Rebecca Knight alongside my co-host and analyst, Dave Vellante. We're talking about the automation journey for State Street in this next segment, and I'd like to welcome our next two guests, Brandon McCormick, managing director, global practice lead, Alpha Professional Services at State Street. Thank you so much for coming on theCUBE, Brandon.

Brandon McCormick

>> Thanks for having me.

Rebecca Knight

>> And Manish Bakshi. He is the senior manager, investment management at Deloitte Consulting.

>> Thank you.

Rebecca Knight

>> Thank you both so much. So, I'm actually going to start with you, Manish. Can you give our viewers an overview of the Alpha platform?

Brandon McCormick

>> Yep. It's a front to back platform that comprises a lot of different applications and services that use various different technologies. It's a combination of on-prem and cloud technologies coming together to offer investment managers a full view of what front to back servicing looks like. And that's where some of the complexity comes in, because you're dealing with so many different applications using different technologies, we needed a tool that kind of helps us test extremely diligently end-to-end through the spectrum. And that's where we had UiPath as our tool of choice. And the other thing I would mention is the product itself is evolving at a rapid pace. So it's again, prudent for us to continue to make sure that we're testing it diligently as it gets released to our clients because it is being used to manage a lot of money in the market.

Dave Vellante

>> So what are you doing here?

Brandon McCormick

>> Well, I run testing the most exciting job in the business.

Dave Vellante

>> Oh, interesting. We had an interesting conversation about testing-

Rebecca Knight

>> We did indeed. And how much it's changing.

Brandon McCormick

>> It is, yeah. It's definitely, and that's why we're here today, right? So my team we're involved in the implementations. We also do post-production support when regression testing is needed. But we have teams of people around the world that are deployed to all the different client implementations. As Manish said, these are big complex projects, multi-year, hundreds of people. Lots of change, lots of surprises along the way in terms of people learning as they go and uncovering more details as they start to pull back the onion more and more. And my team's responsibility with partnerships like Deloitte and Manish who leads that for us is to make sure that the platform itself is internally tested and that we partner with our clients to make sure that they're successful in the testing journey that they're on. They're going through an equal experience, meaning that they're having an operating model change, a very disruptive, big operating model change just as much as our firm is. So, as a partnership, we're co-developing the platform and we need to make sure that we're all successful in the testing journey.

Dave Vellante

>> So, Rebecca is a journalist by trade and writes a lot about the future of work and obviously as concerned about employment and shifts and so forth. And I said, "Well, you've never done testing," because they'll be happy to automate tests.

Brandon McCormick

>> They will.

Dave Vellante

>> I totally agree with you. I used to do testing and it's not the most exciting job in the world. But so what do you do with those people? How do you retrain them? How do they advance their careers?

Brandon McCormick

>> So, I'll tell you first and foremost, they are very excited. They don't want to manually test just as much as management doesn't want them to manually test. We all want to move towards automation. They want to be at the forefront of their job, of their profession, using the first grade tools, first class tools like UiPath and taking some of the mundane, repetitive things off of their plate and allowing them to go deeper and to really get into what they love to do, which is problem solve. They love to look at the different areas of the platform, work with the subject matter experts and really figure out, how do you break it? How do you challenge what was built to see if it will work or if it doesn't work? So that's what they love to do, and they want more time and more freedom and better tools to be able to do that. So they're not being displaced. I mean, a lot of people talk about that, but re-skilling, we're making sure that they have access to all the training that UiPath provides to learn how to do their job differently. And I'd like to think that pretty much everyone is excited about that.

Rebecca Knight

>> You're eliminating those administrative burdens and the tedium from their job.

Brandon McCormick

>> The profile of the role gets elevated so much on more critical and just more smart activities that you're not doing the admin or the mundane or the repeatable things that you are spending a lot of time doing. So, that offers people an ability to, again, like Brandon said, think about what can be done more or what we should be thinking about more rather than getting bogged down by execution-related things that the bots can do.

Dave Vellante

>> So Manish, this is an interesting time for consulting, which is a very human-led business, but of course you're injecting... there's more of an asset-based business increasingly, and there's more software and sort of a platformization. How are you thinking about, whether it's AI, automation, agentic, affecting your business?

Brandon McCormick

>> It comes with a little bit of that concern that, oh, it's going to eliminate the workforce. But at Deloitte, we hire the top talent. So we definitely don't want people wasting or spending time doing mundane things. They look at it as a challenge as well that if some of the things that I spend a couple of hours doing today gets automated, I can actually focus on high-value activities and tasks, and it's a very promising career direction for them as well. From that standpoint, I think it will definitely help us with getting a lot of this stuff done much faster. But it frees up our people to do things that they like doing.

Rebecca Knight

>> And the stuff that AI can't do, the human touch stuff. Yeah.

Brandon McCormick

>> That will always be there. You will need people who have lived through transformations, who experience these things, who bring their expertise to the table. It's the way you execute could get a little bit more easier or more streamlined because the bots can do that for you.

Dave Vellante

>> So, in talking to some of the folks yesterday, it took a while for the testing automation to really take off, and my understanding was that the ROI wasn't so obvious in the early days, but with gen AI it becomes more palpable. Can you help us understand that, Brandon?

Brandon McCormick

>> Yep. No, that's spot on. I would say that we are still in the early stages of that automation journey. So right now we are at the robotics stage. We're actually going to speak a little bit later about taking advantage of Autopilot and the generative AI capabilities that UiPath has to offer. But it took a little while to get up and running. We were lucky enough to have a centralized team within State Street that spearheads AI, machine learning, robotics. So, we were able to partner with them and tap into the partnerships that they had, one being UiPath and operate under their guidance and really go to them with our problem statement and ask them, "How can we solve this problem?" So we partnered with them. We have a strong partnership with Deloitte as well and some other service providers in the market and it's a slow start. Really what I tell people, you have to have your house in order before you can actually start doing automation or you'll just end up automating a whole bunch of random things that don't necessarily make sense and they may not even help move the needle at the end of the day. So, we spent a lot of time creating the governance structure of the team, making sure that we had good processes in place, making sure that we had the right people in the right roles, and really preparing for the ability to scale and doing it intentionally and not just sporadically.

Dave Vellante

>> So, having your house in order means governance and skills and with a scale mindset. What a lot of people tell us is data, but this is not a data problem or is it another piece of the puzzle?

Brandon McCormick

>> So, it is a data problem. What I would use instead of data for us in our situation, the data would be the equivalent of the thousands of test cases that we have that cut across all of our clients, that touch all different parts of the platform. So if that's my data set, I want to make sure that that is something that is rationalized, normalized, it's reusable. So, it's not just test team A coming up with their approach on how to test something and then test team B having the same set of test cases, but they're different. So, how do you create that common set of test cases that act as your denominator that you can start automating and have reusability and get scale out of that? We also deal a lot with data. Data underpins our whole platform, so in terms of data that moves, investment data that moves from our front office applications all the way through to our warehousing and out to the client, we do a lot with data and that's a whole topic in and of itself, but that was my first reaction to your question.

Dave Vellante

>> So no, I appreciate that nuance. And so is it a matter of identifying the efficacy of the testing and the commonality across so that you can reuse them?

Brandon McCormick

>> It is.

Dave Vellante

>> So it's not a one-off test, even though it could be incredibly effective, but it's a one-off, so it's really not applicable.

Brandon McCormick

>> Exactly.

Dave Vellante

>> You would dial down the efficacy to get more applicability.

Brandon McCormick

>> Yup.

Dave Vellante

>> If we just go and automate everything that we currently do manually, we will run a risk of having a high-tech debt later on because we haven't gone through that rationalization process. That's why that broader strategic vision and how we execute becomes paramount because that tells us, spending time upfront is a lot more valuable rather than dealing with a lot of stuff that we just transformed and we have to get rid of later on.

Dave Vellante

>> So, that's what? A test asset rationalization exercise?

Brandon McCormick

>> Yeah, it's been .

Dave Vellante

>> What are the buckets of the... I'm familiar with application rationalization, invest, retire, et cetera. The same kind of-

Dave Vellante

>> It's similar, right? Because we're doing multiple implementations at the same time, all with a little bit different flavors. So, for us to understand the commonality and having that one library, which is common across all of them, or spending the time doing that is a valuable exercise.

Dave Vellante

>> And that tribal knowledge is presumably embedded within the testing team. And would Deloitte helped surface that and the map of it?

Brandon McCormick

>> Oh, absolutely. I'll speak as Deloitte's client. I'll speak, with their contribution to everything that we do, not just testing has been invaluable. And just to geek out on you for a minute on your last question. We mirror what the business does. We're testing on behalf of them to make sure that what they built based off of the scope that they gathered off of requirements is working properly. And so we always look to the stakeholders upstream of testing to understand how we construct our approach to testing. And in this situation, we have what's called a class of service families and service capabilities within each of those service families that our clients can buy. Picture it as a menu that our clients can shop off of. And so when we look to scale and automate, we're looking at which test cases align with each of those service families, the service capabilities within it. And then we have some more foundational work just in terms of transactions that are manager initiated by portfolio managers. Non-manager initiated, which is cash and dividends and things like that, that flow through the entire platform. And there's your treasure trove of stuff that you need to rationalize and make applicable at scale to all clients.

Dave Vellante

>> So you're aligning your testing protocols and patterns with the business value, which is critical. Because a lot of times the business gets the test that gets the results and they say, "What'd you test? Well, we don't care about that. I mean that's like 10% of the problem. What about this other 90%?" And it used to be kind of throw over the fence and almost like DevOps now, much tighter-

Brandon McCormick

>> Yeah. You have to be fully integrated with the teams.

Dave Vellante

>> Very cool.

Rebecca Knight

>> As you said, it's early days and it's a bit of a slow start. You had to do a lot of housekeeping even just to make sure you could start this journey, but have you seen any early successes, any wins to celebrate and get the team even more galvanized?

Brandon McCormick

>> We have, we have. Actually, I often go back to the first proof of concept example that we created with UiPath. And that was taking the most basic end-to-end transaction through the full platform. And we did a time study, it took about 12 minutes for a tester to manually do that. Logging into each of the different systems, doing all the different validation checks. When we turned that into robotics code using UiPath, it took four minutes. And it took four minutes, not because the bot couldn't do it faster, it took four minutes because the platform itself took that amount of time to process the full life cycle. Now mind you, that's not production, right? That's a lower environment, so we don't have full production support, full production horsepower. So that's where the four minutes comes from. But a huge win. And we run thousands of those types of test cases on every client implementation. And we often have to rerun those same test cases as things change just to make sure that we don't get different results from the first time it was run. So, that's the example I always point to.

Rebecca Knight

>> Yeah, absolutely.

Dave Vellante

>> And the other thing we heard the other day was that in the early days of testing, you could only attack, let's call it 25, 30% of the problem. But today with AI and greater automation, you can go after a much bigger piece of the pie. So, the business case has becomes much more compelling.

Brandon McCormick

>> The repeatability, of course, is a big aspect of it, right? But the coverage part is tremendous as well, that if there are X multiple permutation combinations you can do with different asset types for a simple thing like a trade, there could be 200, 300 different combinations that the bot again could do in minutes rather than again, a person trying to do it in hours or something.

Dave Vellante

>> And the quality's got to be much, much better.

Dave Vellante

>> You take out the operational risk or any of that out of the equation completely.

Brandon McCormick

>> Well, and the quality to that point, if you're allowing the machine to do that work, then the person who actually has all that knowledge, that tribal knowledge, they get to actually apply a level of scrutiny that they might not be able to do if they're stuck in the weeds. So, it elevates them to a new view that allows them to think about things differently and really do that value-add work that Manish about. And I'd also say that Autopilot, I mean, we're really excited to get our hands on that. So to take those testing inputs and let Autopilot tell us what the testing landscape looks like, to get that coverage that we might not fully get today, that's game changing.

Dave Vellante

>> Can you explain that in a little bit more detail? I don't quite understand how that's going to extend the value.

Brandon McCormick

>> Yeah. As Manish said, robots can do things faster and they can do more things at that greater scale. And so if we take an engine, I'll call it like Autopilot, and if we give it the same inputs that our human test team takes today and we tell Autopilot what those inputs are and how to think about them and we give it whatever other information it needs with our desired outcome, the tool, the engine can then bring all that to life and capture all the permutations that could take a human hours, days, months to do and do it quickly.

Dave Vellante

>> So, you prompt that with natural language.

Brandon McCormick

>> Yeah.

Dave Vellante

>> Maybe shape it, fine-tune it.

Brandon McCormick

>> And we just got a preview of a future product release where a tester is using natural language communication to the engine. I use that term loosely to say, "Hey, I'm going to test for client X and I'm going to look at this aspect of the test and can you tell me what we've done on other deals?" And then the bot and that whole engine pulls forward what's applicable as opposed to that person having to do all that themselves.

Rebecca Knight

>> So, I've got a question for both of you about if everything goes to plan, what is the ideal? As State Street embarks on this journey of automation, including the generative AI too, what does the ideal future state look like if everything works perfectly?

>> You want to go first?

Brandon McCormick

>> Yeah, let me take this side. So, the vision that we often talk about as a team includes that autopilot. You take those same inputs, feed them into a tool called Autopilot and engine, you create that testing landscape, Autopilot actually creates the test code for all the test cases that need to be run. The bots, which we actually have running in all of our client environments today, will actually take those test cases, run them in an automated way, and it will return the results back to our central utility platform, Jira, which we use to manage all aspects of every implementation that we do, including the development work. And then the icing on it is if the bot finds a defect, it can raise its own issue in Jira and it can bring forward all of the relevant information that's needed for the person who actually has to figure out what went wrong and how to fix it, to fix it. And it can tell you where it came from. It can tell you what requirement it was derived from, what solution it ties to, what application it ties to. And so all that can be done to free up our team to do so much more on top of that. .

Rebecca Knight

>> Excellent. Well, Brad, you can feel your hunger, Dave.

Dave Vellante

>> And this is coupled with, like I said, the product is evolving at a tremendous rate. One of the best-selling points of the product is that it's highly customizable. So, we expect to continue to see that evolution happen. That's why our ability to kind of repeat some of this work and run it at a pace that we want to run it gives us, again, a great opportunity to do this work in a much robust way.

Rebecca Knight

>> An excellent note to end on. Manish, Brandon, thank you both so much for coming.

Brandon McCormick

>> Thank you for having us.

>> Thank you for having us.

Rebecca Knight

>> I'm Rebecca Knight for Dave Vellante. Stay tuned for more of our live coverage from UiPath FORWARD. You're watching theCUBE, the leader in enterprise tech news and analysis.