YTread Logo
YTread Logo

Automation Robotics in Banking | Exponential Finance

Mar 18, 2024
What we wanted to do in this panel and I present it very quickly was not only talk about

exponential

technology and its adoption in banks, but I also talked about what is real today, what are the challenges in banks and services? financial in general and what the limitations are. Also, there are some real life examples that our panelists are going to talk about very briefly today, from my right, jeff kuhn Adlon Zhu and Peter Spencer Jeff is executive vice president of the Bank of New York Mellon Bank of New York Mellon today is the The largest custodian bank, it has almost $30 trillion in assets under custody.
automation robotics in banking exponential finance
Jeff has led many innovative programs using cognitive machine learning in his store that are real today and I'm just going to talk a little bit about that. Adeline Xue is a chief marketing officer and she is in top-ups, she is a leading research and strategy firm focused on educating and advising Fortune 500 companies specifically on artificial intelligence and purchased technologies, and she will speak from a perspective research and strategic advice on what it is. what is happening in the market and last but not least, Peter Spencer who is a member of the board of directors of several startups and technology companies, one of which is Temenos, a large technology company with more than 2,000 clients for general financial services, Prior to that, Peter also led Deloitte's analytical and cognitive offerings of its clients served in a number of areas, very distinctive perspectives in a general

banking

advisor research.
automation robotics in banking exponential finance

More Interesting Facts About,

automation robotics in banking exponential finance...

I'm the moderator, my name is Rita Rajan and I run Cognitive Robotic Life Consulting offerings, so with that as context, let's dive into this for half an hour, hopefully it'll be an interesting dialogue with the first few questions you have more around to an explanation, maybe even a broad question before we dive into what's real today, where do you see it from an institutional standpoint? From a

banking

perspective, the financial services industry was headed in five to ten years, especially in the adoption of these technologies, so you know, ten years is a long time, so I'll stick to the five-year frame on the second one, but I think you already know with all the Regarding the previous speaker, which is a tremendous insight from our perspective in the large institutional banks, it is more about the adoption of technology, without a doubt, we are very, very focused on digitalization, on try to extrapolate the behaviors of our clients based on digital patterns. of managing liquidity, but a lot of it is really focused on how the technologies that we've been working with are adopted and now implemented and those can range from a much broader implementation of workflow technologies, types of machine learning technologies which I'm going to talk about a little bit later about a major program that we've implemented around machine learning that really saves us enormous amounts of money.
automation robotics in banking exponential finance
How are expenses driven? Leadership efficiency reduces risk. Improve customer experience. That's the kind of thing we're going to present. at the bank we are in the past we could have been and not in the past a couple of years ago we could have been more worried about whether or not Apple or Google will become the next big giant competitor I'm not sure they really want to embrace the regulatory environment that big banks have to deal with institutionally, so I think right now it's more about adoption versus disruption for institutional banking at Adalind. From your perspective, obviously, you've dealt with a lot of companies, you obviously research and analyze things as well, what is your opinion on where the bank is and where a financial services company is going, given that in the past we have been one of the biggest adopters of

automation

and digitalization of some of the newest technologies and what is it?
automation robotics in banking exponential finance
Different on this, let's take a step back, many companies in the financial industry have to get all their data first and banks and financial institutions are one of the best prepared to use artificial intelligence and machine learning. I mean, no other industry has done it this well. As much data as we do in the financial sector. I have seen many companies, for example Visa, that began to organize their data. For example, five years ago was when they started a division or team that was solely focused on Cola collecting their data. in becoming a center of excellence for the different departments and now that the teams have grown to 80 people and that capacity can be used by the different divisions, then the regulatory division or the Cyber ​​Security Division all of them begin to turn to this set of data that has been cleaned and trained over time so that they can begin to improve and use machine learning in the other different sectors of their business, another more tangible side of consumer applications.
I've seen eyes, they all talk about BOTS thoughts, as we fundamentally know them, they may or They may not be artificially intelligent or intelligent many times they may just be dumb pipelines and decision trees. Well, we've all been on the other end of a customer service experience where the BOK keeps asking us the exact same question. We're finally human, please, but a lot of times, although those are just precursors and we're seeing, especially if you're touching consumers a lot on the banking side, a potential customer service for the banks is one of the huge costs of the center.
Well, it costs between five and twenty-five dollars or more to attend to each customer service ticket, which is a great savings because if you can reduce them by using an intelligent agent or many times in these times it is the cyborg model where an agent plus cognitive AI. With solutions you can now increase your revenue without increasing your costs and consumers care, so I've seen a lot of banks move in that direction on the consumer side. Companies like TD Ameritrade have a robot that Bank of America announced a month ago, so we have yet to do it.
Look at it, it's capital, a super innovative one, the space they have the Alexa-scale SMS ad bought there, we are funny for exactly so many, several hundred, yes, yes, and Peter, you, you played so much on the demand side to help customers adopt some technology like Now you are on the supply side and you also look at a number of technology companies. What is your perspective? Does it feel different than yesterday and the day before yesterday in terms of

exponential

and what do you see in four or five years, yes? I'm starting with the last question: does it feel different than yesterday?
Or even let's say 15 years ago dotcom and e-commerce because I think there are analogies with that transformation we went through. I think it's different because it's more real, but on the other hand, there are also certain attributes that I'll get to in a second that feel the same in terms of a lot of startups, fintechs, others, a lot, a lot of the which have great technological ideas, but shitty businesses, Weller and questionable businesses. model, I'm sorry, I've just been more polite, but business models are not sustainable and I think they will be a toss-up, but going back to, I guess the first question is kinder, in fact, it's ten years away but it's five years away and that's it.
You know. I can't think in three weeks most of the time, so I won't put a time frame on it and I think our previous speaker in terms of a vision describes a compelling vision. I think a lot of that makes great sense. These things are time, so when will we get there? I mean, if you think about what the forces are, technology is clearly critical and for much of this conference, but the regulatory forces will still exist, so the regulatory forces have an impact on what can be done and the timing around demographics, I think it's also been talked about in terms of changes in behavioral expectations etc, so trying to predict those for me is very difficult, so I'll just take an example in retail banking.
I've spent quite a bit of time in Europe, says the brick. I spend quite a bit of time there in my current role and they are talking about the banking community is talking about a regulatory Arnie that you demand called PSD to PSD to is basically saying that the data of the owner of the customers and therefore the banks They must make that data available as long as the customer gives consent to third parties, if they follow it well that means that the bank now for next year will have to try to establish API so that third parties can access customer data that are in the bank with the banks, which really has significant implications for third-party providers, whether it's fintechs, Google Apple, whoever, they can basically bring the customer management relationship to the front office II, the type of thing that have. with full access to the data again, the customer is given the customer's consent so they can add added value, they can add services, they can do huge amounts with it and this leads back to the cognitive, you know, everything that comes with the cognitive is PA, all these technologies, but you need yourself.
I need correct data, so once open banking takes hold it will change the landscape, so I think at least in Europe, where this started, I think what we will see is the model that the models will change and it will end with some banks. which can be just like utilities, you know they have banking products, they're highly regulated, they can be commoditized, they just provide, let's call it the back end in a crude term, you might feel, so you're going to have these third party providers, which is like The distributors add value and there can be a large number of those working on that, so there are two different types of models and a third model is what we would call true, pure digital banks that have open banking but were built from scratch. as a digital bank in terms of processes, technology, etc., so I think in terms of transformation, I don't know what 25 years or 10 years will be like, but I think in Europe at least you can see it. how that interplay between technology and regulation could give you a picture jutsu like that, and we're seeing some evidence of that today as well, especially on the retail side, where you have customers or customer onboarding that used to take four weeks now. have been reduced to three days, in some cases lemonade, which is a very public example of insurance in which a claim is resolved again and in a matter of four or five weeks, in seconds, the problem of the commercial proposal of the proposal of the client seems to be fundamentally different. and the only way to solve this is by adopting some of these newer technologies, one of the forces that you talked about a little bit, Peter and Jeff, let me ask you this question, one of the forces in this rise of The adoption of the cognitive machine learning was this explosion and unstructured data right in dealing with contracts, invoices, texts and a variety of unstructured information to really do an analysis of it and you, when we were talking, you mentioned a very relevant case example and its adoption and cognitive machine learning to solve this unstructured data problem, can you talk a little bit about that and talk about what is really real today and what is kind of an aided battle, so we are the institutional type of processing and serving larger? bank and we have thousands and thousands of customers throughout the history of the bank, so we have every contract variation you can imagine in multiple languages ​​in different tenures in different legal entities, etc. and this starts with the regulatory requirement around resolution and recovery where the Basically, if we were to go into resolution, the regulator wants to understand what are the potential contractual exposures that the bank has to its clients and establishes a whole series of different types of contracts that define contractual exposures, so we're actually engaging with a relatively young group. machine learning software company and we're using their software to, I mean, we digitize every contract, now we ingest the contract into the software, we apply rules and those rules can improve as you use the product, so the first step might be look for low ratings or key terms or paragraphs that define certain contractual responsibilities and as you review contracts you improve the rules and therefore your accuracy level could even go from 50 to 80% or 90% in terms of power We identify that information in the specific contracts and we don't actually centralize all the contract repositories, but we take a snapshot of each contract and apply these rules to it.
We now have over a million documents that we are applying to this set of rules. with this software and initially a very expensive project to comply with the regulation, but recently there was an issue around outsourcing requirements, what our clients' requirements are and whatWe have in contracts related to how we outsource business to other legal entities in the bank, which could have taken weeks, months and had a very low level of precision, it took us six hours to write the rules and scan our entire universe of contracts and have a opinion on this particular question, so we've really taken it completely on paper. based I moved it to digital I moved it to sophisticated OCR and I see our capability with machine learning and now I have a fantastic tool that we can use to extract the data that is in our contracts, not even for revenue purposes, so it's really very valuable. correct implementation, so I mean there are some common themes that are part of the proposal offer that Jeff talked about and that's how I significantly reject what exponentially proves the proposal to my client or in this case, that's the regulator .
I respond in hours instead of rights, etc., etc. Adeline Peter, are you seeing really exponential forces, drivers or business value emerging in the market today through technology adoption? and other students basically say why does it take so much time to be able to do something for my client or my client and can I reduce that by 80 90 100 percent or improve the quality 80 90 hundred percent since some other financial services are limited , we have to be precise, we can't afford to make mistakes and so on, so just to respond I think you know one thing and again this is not new news but as in the case of technology innovative, innovation in general, but just in technology, there are new business models, new products and services that can drive us things that we can't even convince the day that we can do it, we know, so there is all this kind of new business model, new services, in addition to doing what we currently do more efficiently, at greater speed, on a larger scale, etc.
I think from what I've seen in a lot of financial services there's been more, let's call it that, meaning it's accelerating rapidly, reducing costs and things more efficiently in terms of new business models, we've certainly seen some which not much and I think again the previous speaker showed a vision of some of what's to come in terms of what's real and what I've seen, you know, like most of the people here who have worked in financial services , certainly in the retail sector. retail banking or even you know the world, imagine retail brokerage, there has been a huge absorption of analytics, advanced analytics leading to cognitive insights, machine learning, etc., the ability to I think it's better to call customer intimacy at scale so you know the chat box. it's natural language processing, I mean, there's a lot of almost point solutions when you're on the ground and I think Jeff rounds out all of this from your point of view by making this contact and point solutions that a lot of the larger institutions that are interested are getting. more comfortable and developing capabilities, they are not necessarily going to do something that is completely transformative and maybe the issue of disruption is still there, but I think by reaching for the low-hanging fruit showing a real return because most institutions do look for a business case, we're seeing adoption and again, we can talk about retail and I've mentioned some of those things, but I've also seen it, you know, in the capital markets, I've seen it, I mean wealth management .
Roble Advisors, which I don't think are particularly examples of cognitive because they're really more template-based investing, but they're highly automated and they use, let's call them, you know, newer technology. I think there's still a lot of potential involved in management, particularly around true cognitive. What we're seeing today, I think maybe a couple of points from Lee Intuit. I wanted to highlight one where you talked about having to invest a lot and before you had the system, you could work six hours and come up with an answer, so we definitely see something that we see if I also mentioned earlier that large companies have to gather their data and, Therefore, it requires a substantial investment of time, effort and dedication, and we are going to do this for a long time. term, that's one point, the second thing I've seen is that many times it's a single problem for which you hire a supplier or present a solution for your problem, that problem was understanding all the different contracts, there are so many parts of a business in financial institutions you can use exponential AI anywhere, but what I have seen most successful is for a champion within the institution to find a problem or find something they can better iterate on and be that champion throughout the process . bank and then once they start the pilot, once it seems to work, then I see, I see it being distributed.
I mean, we're all pretty. I would say we are risk averse, right? If it's working, why change it? We have so many regulations. supervision over us, therefore I see that it mainly takes a large investment, an internal champion who believes in the long term and then pass it through the bank, let me return to one of the points you correctly mentioned: risk aversion is a point versus. The services have been heavily regulated and rightly so, so they are also one of the highest in the most demanding sector in terms of accuracy and doing things right because of the implications of not getting things wrong.
The examples that the three, if we talk about, generally involve newer technologies. Smaller companies that are somewhat experimental in their capabilities in taking on a particular problem, it seems like that's forcing financial services companies to develop new muscles, almost right, so what do you see in large companies especially, Assuming the disruptors aren't going to completely disrupt? It's about embracing these kinds of exponential technologies in terms of new skills, new muscles, new capabilities, what's changing tomorrow, that you actually bring in today and open it back up to the panel, but Jeff, do you want to go, I mean, I think Look, we because of the security concerns because of the role we play in the financial markets around the world will very cautiously partner with other companies large or small, but certainly, if it is a larger company, there are certain assumptions when Regarding that you have the proper care and controls and structure within that large company, in this particular case we could partner with a smaller company that is a software company, but we are not going to run any data through that company, we will have control of the source code, we will have control of it when it is new.
New versions of an application are deployed to production through our test environment, so we initiate smaller enterprise solutions for controls within the enterprise. If you look at one of the topics that the previous speaker also talked about, which is kyc, I know. Many of you know that we could think about what are the solutions that small businesses come up with, but our preference would actually be part of a consortium and we are honest, you know that we work with the consortia that exist for institutional kyc and Ultimately , we envision a consortium being the provider because first of all the regulator holds each bank responsible for kyc liability even if they use a third party and we are providing a lot of our customer data to these consortia so we are very concerned for security around this, so we think through every company we partner with through a very, very rigorous third-party vendor review process that is reviewed annually or even more frequently if services change or if a company changes and recycles. and yeah, go ahead, so I was in Thailand with Adalind, okay, so let me enlighten you with one thing you mentioned.
I'm on the temos board and we sound like an advertisement, it's not meant to be that way, they are commodities. large provider of wealth management and core banking software and I think they have a pretty well designed system. What they have done is build what they call a market to which they have opened APIs and work explicitly with fintechs so that a market of fintechs exists. which is part of the ecosystem provided by terminals for their clients, so it opens up that, in addition, from the point of view of the banks, some banks, with regard to their core banking, are trying to renew themselves and will speak with a terminal or a competitor we are afraid about doing that, but what I have also seen and I think it is an interesting way forward is that we can't do that, we can't move fast enough, that's not going to happen, so some banks will say , you know, we will do a pure digital bank we will open a pure digital bank or we will brand it completely differently Bank Leumi and Israel have done this, they hit the ground running, it is a pure digital bank and they came to a term loss and ten losses they have a digital bank product that they use that and over time that bank grows and that one shrinks, so there are two different strategies: one is to renew, rebuild within, it is like moving the proverbial battleship in motion, the second is that We didn't really touch that too much, but we built a new one and built. that one goes up and that one goes down so there are a couple of different macro level trading strategies playing adilyn and the last word already as you mentioned a lot of these startups or companies are startups in the FinTech space that have the OCR languages , NLP, sensing, so of course there are small companies that may not always have their systems in place, so what I've seen is that the sales cycles for a company selling to a large financial institution can sometimes be one year, two to two years, so if your financial institution looking to partner, you may need to start talking to companies and anticipate that these implementations will be done within two years to speed up this onerous process, which is which many banks need, for example, even Bank of America. have implemented a new procurement system where they focus a dedicated amount of procurement team resources to help these emerging technologies get through their procurement process faster, so they have dedicated calls with legal counsel who are used to working with startups and big banks they have dedicated themselves to. team members and people to help push a project forward, so that if a company is looking to use these technologies, other than your own, the team doesn't have to wait years before these companies clear all the regulatory hurdles, so I think We're coming right to the end of time, so let me try to summarize a little bit of what we heard.
What I heard from the panelists was a financial services company. Certainly banks believe that the future will define the adoption of These exponential technologies are going to grow at a rapid pace, they are driven by exponential business benefits and the technologies that are used for that. Some examples that you prompted around retail and the institutions that drive that I also heard is the entry point for more complicated cognitive machine learning. It is a one-time problem, it is not a large scale industrial solution, I have a problem as a regulatory example, we are solving that problem in general and that is where things are going well and I also heard today that there are many limitations that prevent the mathematics adopt that and they're trying to work on a long-term vision to solve that problem overall.
One last question, if I may, to the group of people who are looking from the banking perspective of financial institutions to how to develop this capacity. but what would be your recommendation to start with because it's hard to get away from the BAU perspective and there's an engine that's working overall and we've heard other speakers talk about how exponential adoption generally requires almost a different build outside of the BAU structure. which is very, very difficult for a financial institution, yeah, so how do you think about this entry point to say you know what I know, this is necessary?
This is how we start and I will open it, it will be you. You know the observation around Bank Leumi, if you know how to read or listen like a Clay Christensen from Harvard, there is the fact that you cannot, you cannot change large companies from within because it is as if there is a cancer and then all the antibodies They prevent it and stifle innovation. I think it is a challenge, certainly in the banks I have worked for it is not impossible, but it is difficult, so we take specific solutions and there are a couple of things because I have been responsible for many things. of large change-oriented projects and for me the key is that you have identified specific resources, you have a problem, the problem is regulatory or other examples, there could be a massive operational problem, you know we are failing in X percent of the operations.
How do we actually analyze the problems that occur based on the study of the data or how do we have an efficient solution with flowso that if an unstructured email comes in saying I have a problem with my corporate action, it automatically goes to the corporate action group, so we take a regulatory or efficiency issue, we address it, we go around it and we actually we make sure there's a defined team that includes program management resources, subject matter experts, lawyers, compliance people, so we have to like to show that that topic is Basically, I love addressing it.
Any final thoughts on where this is going, first Adilyn and then you, Peter, so if you're a senior member and an executive team, I would recommend getting your data and even hiring an AI director. Yes, that's the term today, if you're not talking about how it's different from a chief data officer and if you're not in that kind of role in the head of a business unit or your own department that we've mentioned. multiple times data data data data being the oil of everything exponential, they make sure that you have the data collected within your regulatory department, you have the million different documents organized, you have a data warehouse or a system depending on the type of system you we're using, so when the time comes to implement one of these solutions, at least have that oil ready to fuel whatever growth you need, and Peter, I don't have anything incremental that you can add.
I agree with the two previous speakers on a personal level. On a personal level, I found the course on machine learning from Coursera Andrew Inc. I'm sure some of you made it a great recommendation, just a purely personal recommendation, on an individual level, how can you understand anything about machine learning? and correct cognitive concepts, but institutionally I agree, but with these two fantastic ones, I hope this has given you a little bit of experience, their challenges and the learnings that are happening as well. We have the panel that wanted to thank you all for being part of this session and thank you to the panels

If you have any copyright issue, please Contact