Leadership Insights: Data Migration

Transcript

Amanda Razani: Hello, I’m Amanda Razani, with Digital CxO, and I’m excited to be here today with Bikram Singh. He is the founder of EZOPS. How are you doing today?

Bikram Singh: Well, Amanda, thanks for having me.

Amanda Razani: Thanks for being on the show. Can you share a little bit about yourself and your background, and then a little bit about your company?

Bikram Singh: Sure. My name is Bikram Singh. I’m the founder of EZOPS. This is a company that is focused on helping our clients deal with all kinds of data quality and data sanctity challenges across the entire lifecycle of data. And what we help our clients predominantly do is, with a very flexible low-code, no-code platform, give them the ability – give the business users the ability and the power to actually manage the data themselves, without having to rely on IT, or even us, as a vendor. So they have the ability to transform the data, validate the data, set up any kind of rules along the parts of the process, set of reconciliation rules. And then we also give them the ability to use machine learning to predict, you know, why the breaks are happening, and how to sort of resolve them in an automated way. And then how do you report these breaks? So the entire continuum, the entire lifecycle of the data lifecycle, that is what we help our clients with; we are a FinTech company based here in the New York, New Jersey area. And we have about 175 team members spread across the world. And most of our clients include very large global financial institutions, and they use us for mission critical business processes. And oftentimes, the reason why we are used is because they’re either in some sort of a regulatory sort of a situation, or they’re looking to sort of improve operational efficiency, cut down costs, get more transparency, and things of that sort.

Amanda Razani: Fintech is a growing industry. So tell me – how you mentioned low-code, how is low-code a game changer for professionals, when it comes to this?

Bikram Singh: Historically, what has happened, Amanda is that, you know, anytime you wanted to build an application, or you wanted to, you know, add any kind of tweak to what was already out there, you had to rely on your internal IT resources. And internal IT resources in any organization are at a premium, because there is typically a long waiting list of people who are looking to sort of work with them; it’s a prioritization issue. It’s an issue of timing and things of that sort. And so oftentimes, a business – users, you know, would get into a situation where they weren’t really sort of getting what they wanted in the time frame that they wanted it. So as these local no-code type of platforms are coming in, what they are doing at a minimum is giving the business users the ability to start working and start writing very simple instructions in plain English, that get then translated by some series of steps on the back end. So for example, if I wanted to write a Python code, for example, that sort of pulls in a certain function and does a certain mathematical calculation, historically, you’d have to write you know that thing in a programming language and then sort of do it that way. Nowadays, what’s happening is through very intuitive UI UX frameworks, a business user can just sort of write in plain English, “I’m looking to sort of do this.” Think of ChatGPT, right? So sort of the, you know, the crude analogy, if you apply it to a low-code, no-code platform, you know, still holds true, which is write simple English commands that the machine can understand and then give you sort of the end result you’re looking for.

Amanda Razani: So that’s where that machine learning comes in. Much more efficient, and it allows less skilled individuals to be able to do these jobs.

Bikram Singh: Absolutely. So what has happened is, you know, oftentimes, you would need – even when you were sort of, you know, going down a digital transformation journey, or you’re doing any kind of big, sort of a data migration project, you have to rely on data scientists; you have to rely on fairly technical people. But that knowledge is now sort of getting democratized in the sense that tools like low-code, no-code platforms, using machine learning are now empowering the end users to be able to do a lot of the work themselves. Now, is it nirvana right off the bat? Perhaps not. But the journey is well along its way; the results are very profound. And I think they’re going to drastically change the way we operate as businesses going forward.

Amanda Razani: Absolutely, I agree. So I want to ask, what are some of the consequences of failed data migration, if it doesn’t go right for financial institutions?

Bikram Singh: The consequences can be very, very severe. So if you think about it, financial institutions – their bloodline, it all has to do with data quality; as the amount of data on a daily basis continues to explode, the quality of data becomes even more important and sacrosanct to be sort of captured in that high standard. So as a result, what is happening is if you do not, if you’re not migrating data the correct way, or if you’re not looking at your data quality standards, or if you’re not really sort of, you know, putting in good policies and procedures around how you manage data, you are going to be at a competitive disadvantage. You are likely going to become, you know, sort of the central focal point for regulators, because your business processes are gonna get affected and they’ll break down, your client service levels are gonna get affected. So overall, as a financial institution, the biggest currency that they have is trust, trust that they have with regulators, trust that they have with the consumers and their end clients. And all of that is predicated on making sure that the decisions that the institution makes as a financial institution is based on quality data. So if quality data is not going to be possible, because of a variety of reasons, then obviously, you know, it has profound consequences for these institutions.

Amanda Razani: So then how do these institutions make sure that they go about planning this whole transformation and migration properly?

Bikram Singh: Sure, oftentimes, you know, Amanda, what has happened is, historically, a lot of time is actually spent, and a lot of excitement is there in terms of picking up a particular application or a window in terms of where the data is going to get migrated to, and not enough time is spent understanding what the actual process is going to look like, from a project planning standpoint. There is also sometimes a gap that exists between people who are making these selections and planning these kinds of big projects around data migration – the understanding and cataloging the data that exists within the enterprise. The different sources of data, the types of data formats, volumes; how you sort of go around maintaining the availability of data, when you are sort of upgrading it, whether you’re doing it internally within the enterprise, or if you’re sort of migrating it out to, let’s say to the cloud, there are a lot of considerations that have to be kept in mind. You know, as they say, you’re better off sort of measuring three times, and just cutting once as opposed to continuing to sort of go down these, you know, these paths where, yes, you pick a window, the project starts, and that’s when you have all of a sudden realized, if I had just thought about this a little bit more carefully at the beginning of the project, you avoid a lot of disasters down the road.

Amanda Razani: Yeah. And so in regards to that, what are some of the other challenges that they face? And what are some of the best ways to get around those roadblocks when they hit them?

Bikram Singh: Sure. So some of the challenges have to do with, you know, this understanding the variety of data that exists within the enterprise, right? You have structured data, you have unstructured data, and within those sort of two big buckets, you know, what are the types of data does a financial institution hold? They will have consumer data, they will have data for, let’s say, if you are a bank, and you have a capital markets division, you will have a lot of data around your positions, what you’re holding. For example, you will have a lot of accounting data, you will have a lot of consumer data, as I mentioned earlier. So, understanding and profiling and cataloging the type of data that you have, both on the structured and the unstructured side – that is, number one. That’s very important. Number two is just making sure that the way that you are accessing that data today, whether it’s stored on a database, or you’re pinging an application to get that data, or if it is stored manually in spreadsheets, for example, because reality is, even in 2023, a big chunk of how we operate, even in these large financial institutions, is still by using Excel. So data sometimes is sitting in Excel, right? The point I’m trying to make is that between the numbers, that variety of data, the sources of data – understanding and cataloging that, because when you’re doing a migration, you want to sort of have an apple to apple comparison in terms of how you access it today, what you’re getting when you’re getting it, and when you sort of migrate the data. So that part is very important. Equally important is understanding the reconciliation because if you’re moving within, say upgrading from one version of a system to another version of the system, or if you are going from one application to an entirely new application, you want to make sure that the data that has now migrated over is actually reconciling to what you had previously. So that reconciliation function becomes increasingly important as you think about data migration. And the reason for that is you cannot operate with a different set of metrics. Once you migrate it, the metrics are still largely going to be the same. The metrics are predicated on making sure that the new processes and the new applications are also resulting in the output of the new app, instead of the migration is sort of equivalent to what you had previously. Now. The reason why you do that is probably because you know you are looking at more in applications, you’re looking at access, access accessibility, you’re looking at the quality, but the data reconciliation is at the core of where and when you sort of start thinking and planning about these data migration platforms. Another part of this is also the the human angle, the people who are involved in sort of migrating data or planning data migrations; they also have to understand the type of data or the context in which the data is getting used. So they have to be, either they have to have that background within the business division of the financial institution, or they need to understand or involve and get the participation of the people who use that data on a daily basis. So that human element is very important. You just cannot rely on technologists, to say, “I am doing this thing today; just move it to another application. And, I just have a turnkey sort of way of looking at things going forward.” You need a very collaborative approach between business users, IT users and vendors as well. So it’s a team sport; data migration is a team sport.

Amanda Razani: And the human application is still important.

Bikram Singh: Absolutely.

Amanda Razani: So we’ve talked a lot about, or we’ve focused entirely on financial institutions, but this is an issue for any business that relies on data, correct?

Bikram Singh: That’s right. That’s right. And if you look at it, you know, there are other industry verticals, where there is even more explosive growth. As far as data is concerned, compared to financial institutions, if you look at the healthcare and the pharmaceutical industries, just in terms of, you know, the amount of data that they deal with; this is part of either clinical studies or just the day to day management of their businesses – tremendous amounts of data are getting produced on a daily basis. Same with the supply chain businesses, right? I mean, post COVID, if you look at how things are sort of evolving, now there is a hyper focus on understanding exactly what type of data do you have access to? Where is it being stored? The quality of data and things of that sort?

Amanda Razani: Absolutely. So looking at AI, when we’re talking about AI, where do you see AI going in the future? How do you see it, changing things five years from now?

Bikram Singh: I’m gonna type that question in ChatGPT, and see what it has to say, Amanda. But my answer, I think it’s going to have a profound impact in terms of how we operate going forward. A lot of the experience that people have just in terms of institutional knowledge, and their intuition, and just their experience, a lot of that is actually going to be rendered moot, unfortunately, because reality is, human intuition is increasingly going to be digitized. And, you know, things that we think that we know, and we do it a certain way, because you know, that’s how we trained ourselves – all of that can be done using machine learning going going forward. Barring a very few exceptions, like sales and relationship management, I would say, a large chunk of jobs, and within those jobs, you know, the functions within those jobs are probably going to get automated using machine learning. If you think about it, any programmer today, any earliest junior programmer today, the type of code that they’re writing, or what they used to, can be easily replicated using technologies like GPT. So what is then left is sort of the higher value sort of knowledge kind of activities in that chain. And even that – so, as an architect, if I have to sort of say – here’s sort of my tech stack, and this is what my value add is that I can design; the tech stack, the reality is even cheaper, you can even sort of replicate that – it’s just a matter of time. So what you’re really sort of good at is then understanding and defining instructions that the the AI algo can sort of follow, right? So that knowledge is going to be important. But it’s a question of time, right? How long are we able to sort of keep our edge compared to these technologies? I say that gap is going to start shrinking year over year pretty quickly. And so for anybody who’s going into college today, my advice would be: Definitely pick up some computer science classes, and definitely understand what the world is going to look like, because when kids graduate a couple of years from now, even the ones graduating now are going to face a very different world compared to the one that you and I grew up in.

Amanda Razani: Absolutely. And what you were talking about – being able to get the AI to do what you want it to do – I’ve been hearing this buzzword “prompt engineering.” That’s kind of a new thing is knowing how to prompt the AI.

Bikram Singh: But the reality is, you know, over a period of time, even that prompt engineering part itself is going to be moot right? Because if you give it enough prompts, it’ll figure out how to do that as well. It gets trained on data, right? So if there is enough data, enough prompts over a period of time, it will get better at that too.

Amanda Razani: Very true, and just get smarter and smarter over the years. Thank you for sharing your insights. Is there anything else you want to share with us today?

Bikram Singh: Just that we live in very interesting times. Automation is becoming, you know, central to every organization, every financial institution. And we here at EZOPS – we feel that we have built a very nice platform. And for anybody who’s interested in understanding what is it that we do, I strongly encourage them to visit our website at www.EZOPS.com and thank you again for the opportunity to be on your show today.

Amanda Razani: Thank you so much. Have a great day, everyone.