Modernizing Technology Through Low Code I Tech Trends I J P Morgan

Show video

[Music] welcome to tech trends tech trends is a podcast series provides perspective on the latest trends in technology finn tech and digital on today's episode we're going to talk about low code development what it is how it works the capabilities and benefits it provides and how it may change the business landscape going forward i'm aneesh bhamani chief product officer for commercial banking and joining me today is john napoli head of strategy governance and execution for the chief technology office here at jpmorgan chase john welcome to tech trends thanks for having me anish really pleasure to be here so john you and i both spent years in technology we're seemingly the only way to deliver software and platforms and processes was to employ you know skilled developers to write hundreds or thousands of lines of code uh flash forward to today and we're increasingly hearing more and more about low code development and citizen developers and how it's revolutionizing software can you explain for our listeners what exactly low code is and how it works yeah sure so so low code is really an approach to software development and process automation that requires little to no coding as the name would imply and the idea is pretty simple you try and leverage configuration based or more um or more graphical user interfaces to allow people to build enterprise technology in a way which in the past you would have to do through traditional kind of coding and so the other thing that really focuses on trying to put this power in the hands of what we call citizen developers so people who aren't necessarily trained as traditional software engineers so i use an example of kind of websites when i graduated uh university i remember when we used to build a website we'd have to go and code through html and you probably remember we would do style sheets and all these complex things now nobody really does that you can go to something like wordpress or go to a different website where you can use templates or drag and drop and you can build some pretty powerful websites without a lot of work including e-commerce type websites so it's taking that kind of approach and applying it to other type of uh scenarios or situations so basically you're abstracting away the complexity of writing code and letting people write whether pseudo code or other things like that to help uh generate programs right yeah absolutely so that's why the term is so much generic in the sense of low code it could really apply to a bunch of different uh use cases which we'll talk about a little bit later okay so let's let's pull on that thread a little bit right so it's clear how low code works now that we understand that what are some of the real tangible benefits of low code solutions yeah so a couple of different things so first it really empowers end users to be able to build things themselves right so if you think about an organization like ours it has over 250 000 individuals in it a lot of those different subject matter experts can actually then go and build their own enterprise solutions as opposed to just relying on enterprise technology also it does reduce backlogs for those developers to people who are in enterprise technology they typically are trying to work on some big boulders that are really strategic and then they also have a bunch of things in their backlog that are smaller and just things that need fixing or might specifically just address one specific area line of business by removing some of those other smaller items you can really focus on the big strategic tasks at hand making the biggest difference for the firm overall the other thing is it's pretty quick time to market so as opposed to having to go into a backlog for an enterprise technology organization if i can use enterprise technology and build something myself i can get it up and up and running quickly and i can actually iterate on it much faster the other nice thing is i'm the subject matter expert so as opposed to just always translating somebody else to build something exactly what i need i can actually iterate on a regular basis i can make tweaks if there's something i don't like as i build this and start to leverage it i can do that on the spot or much quicker the last big thing is it really does improve risk of control so if you think about organization there's a lot of manual processes or things that happen in an organization or leverage more traditional technology like excel or or access by putting this stuff in enterprise technology i can then govern it better with clearer tech asset management and also i can leverage that data in better ways in the future and even feed it into things like artificial intelligence machine learning models so a lot of benefits to uh to us driving this approach to the organization so basically you're providing you know these citizen developers whether they be in operations for the business or or whatever with a platform with which they can do all the things they used to do offline in excel or in one-off tools that had no controls into something that is sort of sanctioned and and approved but gives them the flexibility they need to do what they need to do to do their jobs right absolutely so can you talk a little bit more about uh some of the use cases that we're seeing for low code as i alluded to you can really anybody who owns a prosthesis can leverage this or anybody who wants a visualization or or needs to get their hands on information and so we're seeing as a result of that we're seeing use cases emerge throughout the entire organization and so a lot of the use cases we see are in the areas of things like case management so we think about our organization we have over 70 000 operations people i said before that we have over 250 000 overall they can all use it but those operations people a lot of them are working day to day on these cases that they're using outlook and email they're using sharepoint that might take potentially weeks to to get an answer on something so they're doing handoffs which aren't very effective by leveraging a lot of this technology we've actually seen we can reduce a lot of this work to be what used to take weeks might be hours or in some cases even minutes and so overall in the local workflow space which is there i'm very passionate about right now we're seeing about a 30 productivity improvement and we're seeing about a seventy percent straight through processing improvement that being said uh you know i i'd spoken before about kind of low code being more of a generic term there are a lot of different areas within the industry and within our organization we apply this so some of the different places are process and task mining so process and task mining is really powerful you can actually look at you can actually scan your your logs you can actually um record your screen and what you do throughout a week and you can then as a result of that end up with an explanation of what are the processes that happen on a regular basis where do most of the the cases flow through and where might you have exceptions in many cases those exceptions take a lot more time we also are using robotics so i know last year you have a session on robotics and that's been really successful in the firms we have thousands and thousands of bots now within the organization and that's been great in terms of automating repetitive tasks that people traditionally had to do on a regular basis so what you do is you create a bot and that that bot almost logs on to your system as if you're an individual and we'll take these repeatable tasks and allow you to do it 24 7 which with a lot less defects as part of it we also see this being applied to a lot of document understanding so you think about all the contracts and all the different documents that a firm like jp morgan has to deal with on a regular basis leveraging these tools to and things like natural language processing to read through those and help assist in terms of identifying areas where you need to focus has been very powerful we also we think about data so an organization like jp morgan um given given our our breadth of relationship with households in the united states we have a tremendous amount of data that we can use to really improve our customer service so getting that data in the hands of the people in the front office or in operations who want to leverage it having some of these tools data transformation tools is very powerful especially when you then feed it into what we call data visualization tools which is also with the glow code so you think of if i have a good set of data as opposed to putting it in excel i can actually really do some pretty creative analysis through some of these data visualization tools and then share that analysis with shareholder with stakeholders and then last but not least again an error i'm really passionate about is the automation of manual uh workflows so that's a really big one where we're seeing a tremendous uptick now throughout the whole organization and so if i think about jp as an example we've already saved over a million hours of work by doing this build it out which is covers thousands of processes that being said we have thousands of more we focus on so it's really an interesting area that we'll focus on so john as you mentioned we had uh shefali shah on a previous episode talking about intelligent automation and talking about the range of almost a spectrum of robotic process automation or rpa at one end and machine learning at the other end there's a full spectrum of things along the way right one of the things that um we talked about that episode about robotics is when you put in a bot you you really are are almost putting in a temporary solution to give you the time and the free up the capital to invest in what we might call sort of more long-term or proper automation around that how does low code fit in that do you think about low code solutions being more of a permanent solution or more of a stop gap until you can do some of the more sophisticated things out there yeah i i think as we progress in technology we are doing more complex things so i really do see low code being integrated across uh not just you know broadening the spectrum or universe of people that can leverage enterprise technology but also impacting the way that traditional software developers do their work because they're doing more complex things unless you need more abstraction and so there's a lot of providers now that are even starting to incorporate that that being said a lot of traditional coding is still needed so in many cases whether you're doing integration work uh whether you're doing something really complex you can't just rely on the configuration or they use graphic user interfaces so it's that hybrid or combination but some of the basic stuff you know why do you need to put it all by hand when it's something you can do in a much easier way the other thing i'm seeing is again more on the nutritional software engineering side there are there are different providers who allow you to graphically do what you would do in a similar kind of framework from a software engineering perspective so similar constructs just again doing through a little bit more dragging and dropping and you could always get the underlying codes to do the more complex things and then in those cases they will actually allow you to deploy on top of their agents or their virtual machines and then that all gets deployed through containers which can be leveraged you know any of these new uh kind of paradigms would be on the cloud or internally in different places to be hosted so it really is going to change a little bit the way that uh software engineering's done in the future and again broadening that universe to a lot of people when the past couldn't actually do software engineering okay so let's look at that a little bit and let's look at the other side of the coin obviously you and i both grew up in um in software engineering and computer science right and we've traditionally had season developers who have traditionally been your code warriors right what does low code mean for developers and how should companies that are exploring low code solutions think about their developer roles i mean are people not going to need you know these armies of developers anymore in my opinion the the organ the world in general is getting more technical right so there's there's very few business cases these days that don't require some technology component to it so it's not about less software engineers but what it is what this does this allows the software engineers to focus on those higher order problems so in myspace specifically we're actually focused on things like artificial intelligence machine learning which we discussed we're focused on things like quantum computing and then you know in the future things like um virtual reality and augmented reality there's a lot of complex things that are being done now that being said not everybody has a phd in some of these different focuses so as much as we can abstract out some of that complexity it helps right so we're actually thinking about how do we use low code uh for artificial machine learning we've actually deployed some things in our investment bank that allows traditional quants or analysts to actually leverage that so it's more about how do we assist and augment the existing software engineers and replacing it in any way okay great um now with any new technology um there's always risks involved right so um can what are some of the potential risks associated with low code and what should companies that might be looking to deploy this technology be on the lookout for yeah no so so anish you know you and i know in the past if we were to say we were going to give allow anybody within the organization to put their hands on enterprise technology we'd be very concerned and we worry about having these shadow i.t organizations that sometimes grow in the business and we worry about the issues it can cause so things like performance issues or potentially have outages that result into our operational systems and we can't have that right so in the past we were very controlled around this um so so things have changed these offerings have matured significantly so what they have allow us to do is one give give the power to the individuals who are out there in the front office in the business but also allow us to monitor that and facilitate that and orchestrate their actual development so just because you sit in the front office you're a citizen developer doesn't mean that you you don't have to do all the things that a traditional software engineer would have to do what we the thing that's a little bit different is because we know exactly who these users are because we're we're working towards a model that supports them we orchestrate a lot of this in an elegant way that facilitates a lot of stuff and hides that from that so they can really just focus on providing the business value but we do track and make sure we have the proper guard rails in place including understanding everything going on from a tech asset management perspective making sure they have the proper entitlements in place making sure that people aren't seeing as an example data that they're not supposed to be seeing but that's all part of the operating model and how this all works in paradigm yeah and actually you know you could think about it actually improves controls from what you have now because you think about the volume of you know sharepoint sites i remember when you're trying to get our arms around all the user-developed tools that are out there macros or models or things like that at least now we have a framework in which to sort of put everything right yeah yeah okay let's talk about scale right so in a firm like jpmorgan chase as big as it is um scaling a new approach to technology is not always the easiest thing in the world especially now as you're talking about engaging tens of thousands of people that are not technologists in this right how do you scale low code in an organization the size of jpmorgan chase and and what lessons have we learned that might be useful to others yeah so it's definitely a federated model uh so the idea is i could go tomorrow and hire 600 developers you can do uh you know can do visualization can do data manipulation can do the automation of workflows but that doesn't really make sense as a paradigm it kind of defeats the purpose so the idea is to get the power in the hands of these citizen developers that are out there in the business who want to leverage newer technology versus having to rely on somebody else so how do we really go about that you built a federated model and so if you think about it you have a center of excellence that sits within the business i'll call them the product and they're there to answer questions and to kind of um and to evangelize what this offering is so if somebody in the business comes in with a problem they can say okay here are the different kind of solutions you have here's how to think about it and we'll help facilitate the process now they're just really the tip of the spear in the first point of contact but then when you get to actually saying all right now i understand i want to build something within each of our lines of businesses or corporate functions we have centers of excellence that exist that are almost power users are experts in this so what they can do is they can really help you to uh help you to sharpen you along the process and guide you so the idea as an example in our investment bank if there's something you want to automate around and making it up around an onboarding or kyc you can go to the center of excellence within the the investment bank and they will help to explain to you okay here's the training you need to do we're now going to provide you access to the system and all of that and then underneath that within the technology and this is where i'm focused you have these platforms that support that so you want a tech a enterprise technology platform to ensure it's up and running all the time you're dealing with the the contracts with the vendors um where there are outages you make sure that you address them you make sure you have the scale and you have separation of a functionality where need be so it's really that federated model where it's the tip of the spear with a firm wide center of excellence then the federator line of business centers of excellence and then the underlying technology platform which really makes it work so for anybody who's really interested in building this within an organization there's a few kind of tips i always think about so one is you have to make sure you have the right controls framework in place so a place like jp which is very risk-averse we always make sure that if we're going to open this up to a much different people that we know exactly what's going on and we make sure that this is proper oversight and guardrails end to end you also have to make sure you have the right support model in place you can't just let people have at it you really do need these centers of excellence whether in the line of business or in technology that can help facilitate this and answer questions for people because they will have questions over time as they mature they'll get much better and it'll be a virtuous cycle that continues to grow uh but in the short term i mean a lot of people are new to this you also have to think about go to market activity so this is an education for people there's a lot going on they have their day-to-day and they don't always know the tools that they have at their head so going out and talking to different organizations about here are some of the things you can leverage to do your job better is important so we spent a lot of time educating people on what low code or no code is and is it firm and then also this is not something you could do without enterprise technology so i talked about supporting the platforms but also a lot of these different things you're leveraging enterprise tools that require integrations um that are a massive scale and in some cases if actually the thing you build is still successful if you leverage firm wide they might decide that you know what that's an area where we actually want to take and build with an enterprise technology tool so that conversation with enterprise technology is constant and you have to make sure that partnership's there yeah that's great and i like the way you think about sort of the you know the pyramid if you will right at the base you have the technologies building in the platform the same way you might be building a platform for customers just their customers happen to be internal around this then you have called them like the power users that are sort of training and providing consulting services out to the tens of thousands of people that are out uh uh that are up building these things so it's not a total free-for-all with uh everybody just building whatever they want in this environment right 100 percent of niche so john let's uh we'll look into the future right if you look ahead three to five years what do you see as the future for low code solutions you know where are we going with this yeah so i gotta say i'm seeing exponential growth in in just the areas we're focusing within the firm and so i'm really excited about uh the market in general and i see it exploding and so if you talk to uh industry experts one of the things they're saying is by about 2025 you'll see about 70 of software engineering is going to be leveraging low code in some way and again so for me it's great is part of that is expanding the universal people who can code secondly it's also affecting existing software engineers and the way they do their jobs in a certain way and so if you look at the market in 2020 it was about 13 billion dollars in terms of size and by 20-27 the estimations are it's going to get up to about 65 billion so at least the the industry experts are saying it it's going to grow significantly um that being said you know on the front lines again i see it and it's it's not just the use cases for existing solutions that are growing in terms of low code but it's also the breadth of different use cases that you can apply low code to so i'm super excited about the space you know it's different every single day we talked about potentially applying it to aifml which we're starting to do here um so i'm really excited and i see a lot in the future and something that everybody should be paying attention to yeah it feels like it opens up whole new areas that we can focus on right so it's not a an either or sort of thing with traditional development but as you said you can focus traditional development on the much more technical uh use cases and things like that and then this on much more of the the more traditional but more widely applicable solutions that people need yeah absolutely and i know we're working closely with your business around a couple different cases that we think are going to be you know really good kind of baseline to really grow from there well john thanks very much for joining us today and for sharing your thoughts on this really interesting topic really pleasure to be here thank you so much for having me and to all our viewers thanks for joining us remember if you like this episode you can subscribe like or share tune in next time [Music] you

2022-08-01

Show video