Jon Theuerkauf

BNY Mellon – Automation and Innovation in Services Delivery

Just two years into automation, Jon Theuerkauf joins us from Intelligent Automation Chicago 2017 to take us through the persona around the initiative which, as you might guess is called 'Alex' - a tip of the hat to BNY founder Alexander Hamilton - meaning all of the bots are called Alex, with their own subsequent bot id code.

The institution is a large global enterprise and searched for a partner that would suit an intelligent automation ecosystem of that size and scope. Jon says that the enterprise doesn’t really have processes running on a desktop by desktop basis and while he and the team continue to re-engineer and standardize processes, they know what’s what within each process... Which is why when they dove into automation, they did so at the enterprise level.

Listen Now (or read the transcript below).

[Transcript]

From BNY Mellon, Jon Theuerkauf . But first some supporters to thank, and thank you for listening. This episode is supported by SSON. With over 100,000 members, the Shared Services and Outsourcing Network is the largest and most established community of shared services and outsourcing professionals in the world. SSON is a one-stop shop for shared services professionals, offering industry leading events, reports, surveys, interviews, whitepapers, videos, editorial infographics, and more. Engage at SSONetwork.com.

This episode is also supported by SSON Analytics: digestible, date-driven insights for shared services and outsourcing. SSON Analytics is SSON's global data analytics center. It provides shared services professionals with the global data insights you need through interactive maps, tools, and charts. Get headline industry statistics sent straight to your inbox. It's that simple. Explore the first layer of industry data for free now. Start a conversation to find out more. Sign up at SSON-Analytics.com.

Seth Adler : Welcome to SSON on B2B IQ. I'm your host, Seth Adler. Download episodes on SSONetwork.com or through our app in iTunes, within the iTunes podcast app in Google Play, or wherever you currently get your podcasts. Jon Theuerkauf. I've got an update for you, since we last spoke, which you're gonna love. My sister, brother-in-law, niece, and nephew got a dog. Guess what the dog's name is? Jon Theuerkauf.

Jon Theuerkauf : Awesome.

Seth Adler: Well, no, it's Alexander Hamilton though. They literally named the dog Alexander Hamilton.

Jon Theuerkauf: Well I think Mr. Hamilton would be very pleased. I hope it's a cute dog.

Seth Adler: Yeah, sure. I mean, that's what they tell me. I'm not a dog guy. Did you grow up with a dog?

Jon Theuerkauf: Oh, I had dogs and cats, snakes, turtles, whatever I could find crawling around and running stray I would pretty much adopt it. My parents let me have them.

Seth Adler: Yeah. Alright so you're a pet guy. I'm not like really a pet guy.

Jon Theuerkauf: Well that tells me something right now.

Seth Adler: Oh sure.

Jon Theuerkauf: We'll get you a pet robot.

Seth Adler: That's what I need. I need a bot.

Jon Theuerkauf: We'll get you one. Exactly.

Seth Adler: That's exactly right, but just further on Alexander Hamilton and then we'll stop. A video was sent to me of my niece sitting with the dog on day one. It wasn't a puppy. This dog is six/nine months old-

Jon Theuerkauf: Right.

Seth Adler: And so she's combing the hair on the head and it's curly hair, and she's just combing it straight up and my sister from behind the camera says, what are you doing? I'm combing his hair. Why? Well he likes it spiky. So there you go, that's Alexander Hamilton.

Jon Theuerkauf: He's a bit rad, you know. He likes that stuff.

Seth Adler: Exactly.

Jon Theuerkauf: He was cutting edge back then. You know 233 years ago he was cutting edge.

Seth Adler: And he's cutting edge now, right?

Jon Theuerkauf: He is. He certainly is.

Seth Adler: Let's start where we are. This ones about the automation that you guys are engaging in. The whole things named after, right?

Jon Theuerkauf: It is. It certainly is. When we started about 17/18 months ago after we had done a considerable amount of due diligence around a number of firms that had been forerunners in this. One of things that we decided is that we needed to create some type of persona around the initiative. Part of it was just as an inclusive nature of our change management-

Seth Adler: Okay.

Jon Theuerkauf: Initiative around this intelligent automation program.

Seth Adler: Okay.

Jon Theuerkauf: And so who better than to name it after our founder, Alexander Hamilton, who was a pretty innovative guy in and of his own right.

Seth Adler: I know.

Jon Theuerkauf: So all of our bots are named Alex and then they have a nice long string of bot ID's after them-

Seth Adler: Sure.

Jon Theuerkauf: So they're all unique in that way, but so you'll hear us refer to them quite often in presentations probably throughout this little chat you'll hear me talk about Alex.

Seth Adler: Alex, and that's who Alex is.

Jon Theuerkauf: That's who Alex represents.

Seth Adler: 17/18 months since podcast line knows no time, we're talking on August 1st.

Jon Theuerkauf: We are.

Seth Adler: As it were of 2017. So you've been in it for a year and a half. You've talked about the journey to find the right partner-

Jon Theuerkauf: Right.

Seth Adler: Whether or not you care to share who that is, talk about finding the partner. What you knew about your organization that you needed.

Jon Theuerkauf: Right. So looking back I still think we made the right choice.

Seth Adler: Good.

Jon Theuerkauf: I mean we're a large global enterprise and we needed a solution in the RPA space as part of our ecosystem to provide us with the platform for our intelligent automation ecosystem that would scale to the same size and nature that we have. We have enterprise wide processes. That's what we were trying to automate.

We are not yet interested or have the bandwidth to pursue desktop automation.

Seth Adler: Okay.

Jon Theuerkauf: Now maybe some day our ecosystem will need to include desktop automation because we have scaled all the mountains and we have automated all of our enterprise processes where we need to use other tools like RPA, ETA, and ETL tools and those types of things that will help us.

Seth Adler: Yeah.

Jon Theuerkauf: So you know we're pretty happy that we made the choice that we did and we are unabashed, we chose Blue Prism because of the nature that they have built ... the way that they built their platform and we've been very happy with that.

Seth Adler: So you were looking for a RPA/ERP essentially?

Jon Theuerkauf: Yeah, pretty much. We needed somebody ... and look their background these are the two fellas that started this in the kitchen were enterprise architects.

Seth Adler: Right.

Jon Theuerkauf: So who better? Their mindset was to build enterprise scalable, reusable solution sets that people on the front lines could actually build themselves. So here's the product, you build your own solutions.

Seth Adler: Yeah.

Jon Theuerkauf: And that fits very well into our mission and our value system that we have at BNY Mellon.

Seth Adler: What's interesting to me is that I've spoken with a number of folks from a number of companies both big, small, public, private and for the most part it seems folks are starting small. Folks are starting at the desktop level. It seems to me and please correct me if I'm wrong that you are taking this in the inverse.

Jon Theuerkauf: We did. We were not going to stress test things at a level that were not the norm within our bank. We don't have processes running on a desktop by desktop basis. This is a large institution-

Seth Adler: Right.

Jon Theuerkauf: We're one of the largest global financial institutions in the world.

Seth Adler: Sure.

Jon Theuerkauf: We're heavily regulated. We sit at the center because of the nature of what we do.

Seth Adler: Yeah.

Jon Theuerkauf: Our customers look at us to be very trusted and they count on us to always be there. So any automation that we need to apply needs to be able to scale to provide that trust and that service level.

Seth Adler: Sure, but some of your friends in financial services-

Jon Theuerkauf: Sure.

Seth Adler: Who offer similar kind of product and services are doing it the other way. How much of this has to do with the fact we've got a Six Sigma guy here so our processes are pretty optimized thanks on the way into automation.

Jon Theuerkauf: Well I wish I could tell you that all of our processes were as clean and as optimal. I've heard as even of yesterday at the first day of the conference here where some of my colleagues and friends have talked about they've been doing Lean Six Sigma for many years, their processes were already running pretty high. I'm assuming their considering it some high level of Sigma performance, very few defects-

Seth Adler: Sure.

Jon Theuerkauf: Very little variation.

Seth Adler: Sure.

Jon Theuerkauf: And so therefore they were looking for the next path to optimization, which then lead them to automation.

Seth Adler: Got it.

Jon Theuerkauf: Nice journey. We weren't there.

Seth Adler: Okay. But, that's why they call it continuous improvement Jon.

Jon Theuerkauf: Correct. So I will tell you that there are many who advise firms that are moving into this intelligent automation space that you need standardized processes. That it would be great. You can do all kinds of things with these tools. You don't have to optimize them first. You don't have to do any of these things.

I don't believe that you need standardized processes. I do believe you do need to optimize them before you automate them.

Seth Adler: Okay.

Jon Theuerkauf: Now we re-engineer everything before we automate them.

Seth Adler: Fine.

Jon Theuerkauf: And sometimes that re-engineering is with a big R and sometimes it's with a little r.

Seth Adler: Got it.

Jon Theuerkauf: But, standardization to me can mean many different things because does that mean that if I have three billing processes as an example, and because of acquisitions we may in fact have that.

Seth Adler: Sure.

Jon Theuerkauf: One of them may be a big one that takes care of if I look a 100% of the total billing that we do, one process handles 50% of that.

Seth Adler: Okay.

Jon Theuerkauf: Well if it's ... let's say it's 60% automated and it could use part of the intelligent automation toolkit that we're building to help them get another 20 or 30%, does that mean that I should go and standardize all the billing processes together into one before I automate it? That will take me a lifetime to do that.

Seth Adler: Yes it will, but how are you not then stuck in a world where the process that you're optimizing is not perfect, which makes the whole thing a kind of a mess? You know what I mean? Cross that bridge for me.

Jon Theuerkauf: So when you're looking at automation the beauty of the way at least that we're approaching it-

Seth Adler: Yeah.

Jon Theuerkauf: is that we do have the measures in place so we know how the process is running beforehand. So part of the re-engineering may simply be that we have to know how fast is it, what's your output, what's the defect rates, or what is your reject rate. Those types of things. The NIGA rates, Not-in-Good-Order rates. So we're certainly not gonna promote Alex to speed up production of bad of things.

Seth Adler: Right.

Jon Theuerkauf: That's not going to happen.

Seth Adler: Alright. So this is exactly where I was trying to get. Your point is you're such a Six Sigma guy that your point is, I know that we're not perfect, but I sure as shoot-

Jon Theuerkauf: Yes.

Seth Adler: Know exactly what's going on within the process. I know how it's not perfect.

Jon Theuerkauf: So we know.

Seth Adler: You're not being surprised by these processes.

Jon Theuerkauf: No. So we know when we put automation in and this is again why ... we're always a bit disturbed I think when we hear someone at a conference or we hear when we're interacting with our clients who may have a third party partner advising them. They'll say well you don't need to re-engineer, you can go ahead and do the automation. I'm not sure why they would advise them to do that.

Seth Adler: Right.

Jon Theuerkauf: Because you may in fact be ... the reason they make the argument and I understand because they say look, you're buying a license. It's just time. It's a robot. You're not paying. You've already bought the license you might as well use it, what difference does it make?

Seth Adler: Let's do it.

Jon Theuerkauf: And some cases that may or may not be true, but why would you not take the time because once you've fixed it, once you've removed unnecessary cues, as an example, in our case Alex only needs one cue to pull information from it. It doesn't need multiple cues that a human may need where you have multiple humans in a process and each human needs a cue as an example in some processes. Alex doesn't need that.

So I think, again, why would you not want to take the time to re-engineer to at least get as optimal as you can given the time constraints in your SDLC so that then whatever you're automating is going to be as good as you possibly can get it.

Seth Adler: Right.

Jon Theuerkauf: And also because look at the end of the day we're trying to do things for the betterment of our clients.

Seth Adler: Yeah.

Jon Theuerkauf: And if we can do whatever we can to reduce unnecessary wait time for them-

Seth Adler: Sure.

Jon Theuerkauf: Reduce any errors that may still get kicked over to them-

Seth Adler: Does the whole thing.

Jon Theuerkauf: Look it's really what we're trying to do.

Seth Adler: Totally.

Jon Theuerkauf: Now is it because of Six Sigma, is it because of Lean, is it because of automation? I don't really care. We are very simple thinkers in our group.

Seth Adler: Right.

Jon Theuerkauf: So we are trying to make things better, faster, and more efficient, and help reduce the risk for the overall environment within our bank, and also for the risk of our clients. That's why they come to us.

Seth Adler: Yeah.

Jon Theuerkauf: Because the things that we do, they're coming to us to help them reduce that risk. To take on some of the responsibility that they have and that's why they trust us, and we've been around for 233 years.

Seth Adler: And now you happen to be adding bots. You don't care that they're bots. You care that their improving process and serving customers better.

Jon Theuerkauf: Correct, I really don't. And, again, it's for us we're building a virtual workforce cloud base to help us take on work that for many, many years we've known that if we had another way of doing it, we would fix the underlying platforms that lie underneath the floor, right?

Seth Adler: Sure.

Jon Theuerkauf: Right. The big pipes, the big application.

Seth Adler: Mm-hmm (affirmative).

Jon Theuerkauf: But for a number of reasons we haven't been able to. It's too costly. It takes too much time. Even for simple things on some of these applications would require a great deal of investment and a great deal of time, and there's bigger priorities for the technology divisions. Hence, why this new ecosystem is being created, it's been such a windfall.

Seth Adler: Yeah.

Jon Theuerkauf: Because now you can move away from the way we use to fix them would be create Macros or usually what we call UDT's, user defined technologies, using Excel.

Seth Adler: Mm-hmm (affirmative).

Jon Theuerkauf: Access databases.

Seth Adler: Right.

Jon Theuerkauf: Very fragile, very fragile. And they solved the problems, which the operating problems would have to deal with, but it wasn't a very safe technology to really base financial systems in and financial institutions on.

Seth Adler: Why do you say so fragile?

Jon Theuerkauf: Because you don't have a control environment for it.

Seth Adler: Okay.

Jon Theuerkauf: Now some institutions where I've been and you know I've been blessed to be able to work in a number of firms around the world and many countries-

Seth Adler: Right.

Jon Theuerkauf: All of them work with Excel. If you ever turned off Excel most banks would go, wait a minute what happened. And you know we'd all be scrambling around trying to figure out wait a minute how do we get payments out or how did we get trades out the door because their Macros, that were built into the process, nobody even thinks about anymore because they're invisible.

Seth Adler: Right.

Jon Theuerkauf: They just feed things in.

Seth Adler: Right.

Jon Theuerkauf: Your pivot tables are running and dragging in and doing reconciliations and then spitting things into the normal process flow.

Seth Adler: So baked in that I don't even know it's there.

Jon Theuerkauf: Right and that's the danger because we don't have that. They can't be seen. You can't record them. There's no audit trail.

Seth Adler: Right.

Jon Theuerkauf: None of those things that you would normally see in a major application that lies under the floor or even now with robotics, which is the beauty.

Seth Adler: Okay.

Jon Theuerkauf: With the beauty with RPA tools now, is that you always have an audit trail. I can see what Alex is doing when he runs from 10 o'clock in the morning til noon. I can go back and look at every transaction he completed, and the transactions he didn't complete, and where they went.

Seth Adler: Yeah and why.

Jon Theuerkauf: They went over to the human exception team to handle those.

Seth Adler: Before we get to exceptions, let's talk about the processes that you have going. So how many processes are we kind of tackling here?

Jon Theuerkauf: So we have about 60 projects completed now. We interface because a lot of the projects where we're working in the processes if there's multiple processes where the automation is touching and so we're over 100 plus different processes that we'll utilize some or all of the intelligent automation ecosystem that we've built into the project stream.

Seth Adler: Okay. I just want to check you on a word you said, completed. Why did you use the word completed?

Jon Theuerkauf: Because we have projects in the pipeline. There's 16 projects underway right now. There's 45 projects in our pipeline.

Seth Adler: Okay. So completed in that they are running these processes?

Jon Theuerkauf: They are running in production.

Seth Adler: Thank you so much.

Jon Theuerkauf: And there's about 252 bots that are in production or in what we call, live proofing, which means they're in the production environment and we trickling transactions through them, onesies, twosies, threesies, let's see how he runs. Let's test him and see how it goes.

Seth Adler: If we're talking enterprise wide, whatever ever you can share as far as give us an example of a process that we're talking about.

Jon Theuerkauf: Sure.

Seth Adler: That is at least sizable, right? Enterprise wide.

Jon Theuerkauf: Sure. So we do have some things we can talk about very specifically and I have at other conferences. So one of our biggest ones is our settlements process where we have the first process where we actually have Alex running or sitting side-by-side with people that are doing payments, settlements. They're settling trades or settling payments. When we have a very high STP rate, but because we do something like a 190,000 payments a month that get kicked out for exceptions.

Seth Adler: Okay, okay.

Jon Theuerkauf: Sorry 190,000 payments go through STP and then there's about, I forget, multi-thousands that get kicked out.

Seth Adler: Okay.

Jon Theuerkauf: On a monthly basis because those can be very complicated things like the system will want an extra check because this is a payment that is for settlement that needs to be done because it's going to a bank that we may not necessarily recognize on a normal basis. So we need to go do a special lookup.

Seth Adler: Okay.

Jon Theuerkauf: So instead of complicating and slowing down which is a good manufacturing principle, you don't want to slow down your assembly line-

Seth Adler: That's it.

Jon Theuerkauf: So we run about in the low to mid 90s STP on a normal ... making the normal payments, which are again huge.

Seth Adler: Mm-hmm (affirmative).

Jon Theuerkauf: We have an exception cue that comes down and we have teams that look up these payments. They fix payments where there may be errors. Someone may have gotten a SWIFT code wrong, as an example. An account number doesn't look right-

Seth Adler: Sure.

Jon Theuerkauf: For that particular payee. So what we've done is we've built Alex now to do those payments as well.

Seth Adler: Okay.

Jon Theuerkauf: So Alex now sits there, when something comes into the cue, Alex will go into the cue and just pulls them down just like the people do.

Seth Adler: So he's pulling down exceptions as well?

Jon Theuerkauf: As well. He does-

Seth Adler: Taking out exceptions and then pulling them back in.

Jon Theuerkauf: He doesn't kick them out.

Seth Adler: Okay.

Jon Theuerkauf: The STP system will kick it out and then Alex will go in just like if you were on the line and I was on the line, it would Alex in the cloud sitting there pulling them down and doing payments just like you and me. And the beauty of that is, is if we ever have a problem with, let's say SWIFT has a hiccup and for some reason SWIFT goes down for an hour, well that shortens the window that we have with the fed to get the payments out the door.

Seth Adler: Sure.

Jon Theuerkauf: The beauty of it is we would normally have eight or nine of Alex and his colleagues on a regular basis doing payments. Well if we know that we've got an hour less, but the volume's is still the same, all I have to do is go up in the cloud, take some Alex's that are available, give them the model for payments, and bring them down. Now I might have 14 or 15 Alex's working and so the bench strength is there.

Seth Adler: So the bench strength and you're talking about doing more with automation, I haven't heard you say anything about saving FTE or anything like that.

Jon Theuerkauf: Mm-hmm (affirmative).

Seth Adler: All I've heard about is the front end. Adding productivity. Right? A better customer service.

Jon Theuerkauf: Sure.

Seth Adler: Better customer delivery. Is that all you're focused on?

Jon Theuerkauf: No. Look we're a bank and like any bank we have shareholders to take care of and so we are focused on making ourselves as efficient as possible. We have to balance that between our risks. So where we can't do things that are going to jeopardize the institution. I don't have the right to do that. There's only one person that ever does and that's our Chairman and CEO.

Seth Adler: Okay.

Jon Theuerkauf: They hold the keys to making those types of decisions. Right?

Seth Adler: Fair enough.

Jon Theuerkauf: So, yes, are we trying to make sure that we have the most productive workforce we possibly can, absolutely. Do we believe that there is a way of doing the work we're doing it today, yes.

Seth Adler: Sure.

Jon Theuerkauf: Yeah and does that mean that we'll need as many people doing the kind of work we're doing today, no.

Seth Adler: Okay.

Jon Theuerkauf: We're not gonna need. But, we're not gonna do the same kind of work. We're going to do other work.

Seth Adler: Right.

Jon Theuerkauf: And that's the beauty.

Seth Adler: Talk about upscaling part of the workforce.

Jon Theuerkauf: Right.

Seth Adler: And what you see there and maybe even what you've already done versus okay actually yeah, here's where we're gonna make some cuts. First lets kind of discover upscaling. Have you done that yet or where would that be?

Jon Theuerkauf: Look let me give you a couple of statistics that would probably help. In most companies in the United States, especially in finance institutions, you're running between about a 6 to 10% normal attrition rate.

Seth Adler: Okay.

Jon Theuerkauf: We all do.

Seth Adler: Sure.

Jon Theuerkauf: So if you look at there we're 55,000 people so that's 5,000 people a year that naturally attrit. Retirement, move on to better jobs, whatever reasons.

Seth Adler: One guy doesn't like Jon, you know.

Jon Theuerkauf: Right. So they're irritated and just walk out the door. That's a natural attrition so that is something that every institution goes through.

Seth Adler: Right.

Jon Theuerkauf: We couldn't ... right now there is no way I could automate 5,000 jobs out of the bank in a year.

Seth Adler: Impossible.

Jon Theuerkauf: It's just impossible.

Seth Adler: Right.

Jon Theuerkauf: So one of the things that we're very conscious of is look, we have to backfill these positions. Do we always need those positions? No, we don't. But, we need a large portion of those 5,000 jobs.

Seth Adler: Yeah.

Jon Theuerkauf: We need those skillsets. Now, some of these skills we're gonna automate. A lot of them we won't be able to. So I think the thing I always am concerned about when people ask about oh you're gonna get rid of all your workers, no because we're always going to need people, but we're going to need people doing things at higher levels.

Seth Adler: Mm-hmm (affirmative).

Jon Theuerkauf: We're already having people through natural attrition, retire-

Seth Adler: Sure.

Jon Theuerkauf: And move on and do things that they want to do with their lives.

Seth Adler: Sure.

Jon Theuerkauf: So we have to be able to be smart about that. If I can use automation to take care some of the lower level work, I'm still gonna have the higher level work, because clients are requiring us and wanting us to do more things, which requires us to come us with newer ideas, newer products, newer services. We're experimenting with, you brought up cognitive-

Seth Adler: Yeah.

Jon Theuerkauf: That requires a who different set skills that we may or may not have.

Seth Adler: Mm-hmm (affirmative).

Jon Theuerkauf: So we have to go acquire those skills in other places so that new opportunities are opening up. So we are looking at how do we use automation to help us create new opportunities. We are looking at it as the way that we take away the work. You know we don't want our people doing this work.

Seth Adler: Right.

Jon Theuerkauf: I mean this crazy stuff of taking and looking at one screen scraping off data on that one screen and entering it into another screen. Why would we ever want human beings to do that in the first place?

Seth Adler: It makes no sense.

Jon Theuerkauf: Correct.

Seth Adler: I'm with you completely.

Jon Theuerkauf: Right. So we do have examples.

Seth Adler: Okay.

Jon Theuerkauf: I'll give you one quick one now that I've set that up.

Seth Adler: Good.

Jon Theuerkauf: So in one of the projects, a settlements project, we were able to reduce the number of people that actually need that because Alex is now sitting in there. We've actually taken some of the people from that job and they now work for me.

Seth Adler: There you go.

Jon Theuerkauf: They work on the RPA team.

Seth Adler: There you go. So they're managing that process.

Jon Theuerkauf: So they actually are in my team now building bots or in our control room.

Seth Adler: Yeah.

Jon Theuerkauf: So because they know how the processes work, they understand the bank. We're teaching them about intelligent automation and we're teaching them more specifically about how to manage and control RPA when it's in production.

Seth Adler: We're gonna talk about cognitive in a second, but while we're right here, you mentioned the fact that I can't 5,000 jobs in one ... that's ridiculous in one year. What kind of timeline though do you see generally? Not BNY specific-

Jon Theuerkauf: Yeah.

Seth Adler: Of that kind of bigger attrition starting to kind of rear its head. If right now we're not there, we've got upscaling, we've got people retiring, you know it's gonna kind of plus one, minus one, whatever. Where does that start to kind of may be change in your point of view?

Jon Theuerkauf: I think that is a long way down the pipe.

Seth Adler: Long way being three years?

Jon Theuerkauf: Oh no, no, much-

Seth Adler: Long way being 10 years?

Jon Theuerkauf: 10 years at least.

Seth Adler: Okay.

Jon Theuerkauf: I think, again, and trust me I don't put myself in the same league by any stretch of the imagination as Musk or Gates-

Seth Adler: I got cha.

Jon Theuerkauf: Or any of those guys who have now ... they're throwing up the red flags and sounding the sirens of and alarms about this automation robotics, cognitive AI, as one of the most dangerous things we have ever had.

Seth Adler: Sure.

Jon Theuerkauf: You've got Stephen Hawking's and a group of people from MIT saying that there should be standards and things around this, you should be limiting this, and all kinds of things. Look-

Seth Adler: Well Lee's working on the standards, right?

Jon Theuerkauf: Yeah so, again, whether ... and I guess that the Robocalypse view, right?

Seth Adler: Okay, sure.

Jon Theuerkauf: That there's going to be this terminator, Skynetty, what's gonna happen with us technology.

Seth Adler: Why do all those smart guys think that way? Is it because they don't have a kind of high level job at a financial institution that reports to the street. Of course, but my point is you're in the kind of regular person world, if you'll allow it. Whereas Elon Musk is a little bit of in his own kind of Stephen Hawking kind of those guys world. Is that why or?

Jon Theuerkauf: Look I'm a big believer that the future is a window into what reality is going to be.

Seth Adler: Sure.

Jon Theuerkauf: And fiction gives us that future view, right? So if we're futurist, right, they kind of show us what the reality is going to be.

Seth Adler: Sure.

Jon Theuerkauf: Fiction to me gives me the view of the future so you start connecting those dots.

Seth Adler: Because what other example do we have-

Jon Theuerkauf: Correct.

Seth Adler: Than HAL from 2001.

Jon Theuerkauf: Correct. So tomorrow I talk about my view of what the future is going to me.

Seth Adler: Do you mind just covering it now Jon since we're recording.

Jon Theuerkauf: Well as long as you don't play it before tomorrow.

Seth Adler: I promise not to.

Jon Theuerkauf: But, I do think that would it be possible to create the worst imaginable scenario, sure.

Seth Adler: Yeah.

Jon Theuerkauf: Is it possible?

Seth Adler: Yeah.

Jon Theuerkauf: But, I'm a firm believer in the human factor.

Seth Adler: Okay.

Jon Theuerkauf: And I'm a firm believer in ethical automation.

Seth Adler: Okay.

Jon Theuerkauf: And I'm working on a paper in my head right now, which is where most things start for me and then the notes will start in a framework about ethical automation.

Seth Adler: What are we getting at there? Understanding that you're in the beginning of the kind of creating what you think, but what are you talking about?

Jon Theuerkauf: It's really about why we wouldn't do that anyway. Why would we want to build something that we've already seen at the movie theater.

Seth Adler: Aren't we smarter than that as a society.

Jon Theuerkauf: Correct.

Seth Adler: Is your point of view?

Jon Theuerkauf: Correct.

Seth Adler: Are we?

Jon Theuerkauf: And I do think that you know the reason I believe that you get the Elon Musk's and the Stephen Hawking's and the Bill Gates's and others who will sound these alarms because they also know that that's the way you get people to react.

Seth Adler: I see.

Jon Theuerkauf: You create and as a psychologist one of the things that you have to do is people will stay in a point of homeostasis. You'll be in this I won't move until you give me a burning platform.

Seth Adler: Yeah, boil the frog.

Jon Theuerkauf: Oh by the way your house is on fire, oh I guess I better get up and move now.

Seth Adler: Right, yeah.

Jon Theuerkauf: Oh, by the way, robots are going to take over the world. Well I better do something about that now.

Seth Adler: Right. I'm gonna call fire. It's my responsibility as a really smart person to call fire right now so that at least people start thinking about this.

Jon Theuerkauf: Correct. So if for nothing else I respect them for saying look guys, we actually at some point will be capable of building what Arnold Schwarzenegger was on the TV.

Seth Adler: Right.

Jon Theuerkauf: And what Skynet allowed all those bots to do.

Seth Adler: Right.

Jon Theuerkauf: I don't doubt that at some point in time. Is it 10, 50, 100, 250 years away? That's all probably going to be possible.

Seth Adler: Right.

Jon Theuerkauf: Right.

Seth Adler: Is it happening?

Jon Theuerkauf: But, why would we want to do that?

Seth Adler: Also and your point of view being someone that is in the space doing the thing, it's not soon.

Jon Theuerkauf: Yeah, I don't see us ... it'd be hard for me to fathom that and I've seen the numbers go from 10 to 20 years mass reductions of eliminations. I'm like trying to figure out, given what I know, and look I'm not the smartest guy or the sharpest tool in the shed, but given what I see with this and a lot of it has to do with the nature of the unstructured data that we have to deal with and the tools that we have right now don't really work well with a bunch of unstructured data. It requires a lot of data crunching to get these things to work.

Seth Adler: To start.

Jon Theuerkauf: Right.

Seth Adler: Yeah.

Jon Theuerkauf: So now you've got to bring in big data analytics to crunch data or to re-orient things. You still have a lot of regulatory barriers that prevent things to be cloud based where you can dump unstructured data, create API's to go in an extract out the structured nature for you to process things.

Seth Adler: Within the world of finance.

Jon Theuerkauf: Right. So you still have a lot of things to overcome.

Seth Adler: Right.

Jon Theuerkauf: It's not easy to overcome those things. Is it possible? Absolutely.

Seth Adler: Right.

Jon Theuerkauf: The technology is there and will continue to be improved to become available to us.

Seth Adler: But, the regulatory framework has to change before anything really changes, and good luck with that type of deal.

Jon Theuerkauf: And very much so.

Seth Adler: Right.

Jon Theuerkauf: And, again, there has to be, in my opinion, what I'm seeing I think there is going to need to be proof cases ... and I don't blame them. Where the feds in all countries are going to have to get comfortable that these things will work.

Seth Adler: Yeah.

Jon Theuerkauf: And you won't blow things up.

Seth Adler: Right.

Jon Theuerkauf: I mean look we've already seen flash crashes in the stock market, right?

Seth Adler: Yup.

Jon Theuerkauf: Because of algorithmic trading.

Seth Adler: Yup.

Jon Theuerkauf: Well you can imagine if you continue that out, you could see all kinds of crazy things happen utilizing automation and algorithmic or cognitive being more and more advanced.

Seth Adler: Yeah.

Jon Theuerkauf: So the fed, meaning all federal regulators, they do have a little bit of history at this stuff and they don't necessarily like everything that they see.

Seth Adler: Right.

Jon Theuerkauf: So if you begin to extrapolate that out, at least in financial services-

Seth Adler: Yeah, no I got you.

Jon Theuerkauf: You can go across others, you could begin to see where they're saying well wait a minute, are we, should we.

Seth Adler: Well I didn't love the flash crash either by the way you know.

Jon Theuerkauf: We had one two years ago in the month of August.

Seth Adler: That's what I'm saying.

Jon Theuerkauf: On a Friday.

Seth Adler: Yeah.

Jon Theuerkauf: For no reason.

Seth Adler: Yeah, exactly.

Jon Theuerkauf: I mean the market drops like 20% or whatever it did and a lot of the big stocks dropped 20 and everybody likes well why did they drop?

Seth Adler: Yeah.

Jon Theuerkauf: There's no reason for them to drop.

Seth Adler: Within the same day and back up on the same day.

Jon Theuerkauf: Yeah.

Seth Adler: And by the way after that, I was out. Just so you know.

Jon Theuerkauf: Yeah.

Seth Adler: Just so you know and so finally on the cognitive, where are you guys with that next step? I like to call intelligent automation as opposed to dumb automation, which is RPA with all do respect, right?

Jon Theuerkauf: Well no I think and, again, if you look at Lee Coulter's work and I think they've done a fantastic job, right? And we need standards.

Seth Adler: Right.

Jon Theuerkauf: Our intelligent automation work does include RPA, RDA, and it'll include a lot of things. And it will also include cognitive and a number of those things so it's really more of a continuum up and, again, it allows for as you move up the ecosystem and you begin to bring cognitive in which gives you more ability to make more complex decisions and deal with more complex data sets.

Seth Adler: Yeah.

Jon Theuerkauf: So we are, not my team directly, but in our bank we have projects that are using Watson right now.

Seth Adler: Okay, okay.

Jon Theuerkauf: So we're testing to see. We have other cognitive projects that are using combinations of big data tools and true cognitive platforms so that we can figure out whether all this data we have actually will benefit us in other ways. But, cognitive for us is relatively new and I give the guys that are, again, much smarter, much more switched on, out in our Silicon Valley Innovation Center. We have a great team out there. We've got great teams at other innovation centers that we have within the bank that are really turned on to this stuff and they're very good at it.

Me, my guys, we're down in the trenches dealing with RPA tools, host the tools, ETA/ ETL tools-

Seth Adler: Right.

Jon Theuerkauf: Smartform technology. You know the things that help my operating partners. We're helping out with the risk team now, our broker dealer services teams. So we're really focused on the things that we can do today and somewhere down the road we'll partner up and meet up with the cognitive team and see what we can do with them as well.

Seth Adler: I got it, but it does sound like they are also in kind of simply testing mode.

Jon Theuerkauf: They're very much. Our Watson project just started literally a couple months ago.

Seth Adler: Okay, alright.

Jon Theuerkauf: So they're just getting started.

Seth Adler: Yeah.

Jon Theuerkauf: And we've got some teams, the Quant teams, that we've had for a long time in our wealth division that do a lot of things in our asset management division where they have been doing quant analysis and quant trading for a long time. They're now looking at how can they, one, use RPA to help them with something for Robo-advisor-

Seth Adler: Sure.

Jon Theuerkauf: They're also looking at it and bringing in more and doing more cognitive type of analytics which, again, it's just another evolution of their quant analytics that they're doing today.

Seth Adler: Continuous improvement.

Jon Theuerkauf: Sure, absolutely.

Seth Adler: Alright so Ethical Automation by Jon Theuerkauf, right? And we feel kind of positive about the human experience, right?

Jon Theuerkauf: Yeah, I do. I think that in all of this, I have faith in the human race. I have faith in my bank and my employees.

Seth Adler: Right.

Jon Theuerkauf: My staff and I have faith in my colleagues at the bank as well. At least for us, the decisions that we make, the tough decisions as well, those that will change things in our workforce, that change things in our process structure, that change things in our technological structures, all those decisions that will have a future and allow the bank for the future to be more future proof. To be able to be resilient and continue on for another 233 years, God willing, and all those things I think are gonna be very, very important. And I think a lot of it has to do with the fact that we have to trust each other.

Seth Adler: That's it.

Jon Theuerkauf: And technology is great, but it's still only part of the famous triad of people, process, and technology.

Seth Adler: Just another tool.

Jon Theuerkauf: Just another tool.

Seth Adler: Just another different tool by a different name, right?

Jon Theuerkauf: It is, very much so.

Seth Adler: So last time I asked you the three final questions. I won't ask you the first two by I'll try to ask you the third question and you can give us the same song if you'd like.

Jon Theuerkauf: Sure.

Seth Adler: But, on the soundtrack of your life Jon, and by the way go back and listen to the other interview if you want to hear the two answers, but on the soundtrack of your life, one track, one song, that's got to be on there either the same or different your choice?

Jon Theuerkauf: So am I to pick another song?

Seth Adler: It's your choice.

Jon Theuerkauf: Well I think it's probably ... most people who know me and they actually asked me when the song came out many, many years ago if Bon Jovi actually saw me riding down a highway one day and pulled over and wanted to have a chat so I think, "It's My Life", would probably be it. A pretty classic one for me as well.

Seth Adler: There you go man.

Jon Theuerkauf: Along with, "Free Bird". You put those two together and it's pretty spot it.

Seth Adler: And you're fine. You don't need anything else.

Jon Theuerkauf: No, no, I'm pretty good with that.

Seth Adler: Jon Theuerkauf, man, thank you so much.

Jon Theuerkauf: I appreciate it, Seth.

Seth Adler: And there you have Jon Theuerkauf. We've trying to make things better, faster, and more efficient, and help reduce the risk for the overall environment, for our bank, and for our clients. Very much appreciate Jon Theuerkauf's time. Very much appreciate your time.

Stay tuned.

Jon Theuerkauf

Unethical Automation: The Act of Doing Nothing

SSON Podcast with BNY Mellon's Jon Theuerkauf

FIND CONTENT BY TYPE

SSON COMMUNITY

ADVERTISE WITH US

JOIN THE SSON COMMUNITY