Building a future-ready tech organization: Vialto's blueprint for success

Transcript

[00:00:00] Kimberly Boyd: Welcome to Pragmatism in Practice, a podcast from Thoughtworks where we share stories of practical approaches to becoming a modern digital business. Hello, everyone. I'm Kimberly Boyd, your host for today's episode, and I'm joined by Sayee Bellamkonda, Chief Digital Officer at Vialto Partners. We're looking forward to exploring the remarkable journey of how Vialto Partners transitioned into a new technology and engineering organization and the strategies they use to navigate a complex digital transformation along the way. Welcome, Sayee. Delighted that you could join us today on the podcast.

[00:00:34] Sayee Bellamkonda: Thank you, Kimberly. I'm really excited. Looking forward to our conversation.

[00:00:38] Kimberly: Maybe to get us started, I'm sure a number of folks are familiar with Vialto, but it would be great if you could perhaps give a little bit of an introduction to the organization and also yourself and your role.

[00:00:51] Sayee: Absolutely. Thanks to you and Thoughtworks for this opportunity. I partner very closely with Thoughtworks as one of my strategic partners, so this is great. My name is Sayee Bellamkonda. I'm based in Phoenix, Arizona. I'm Chief Digital and Technology Officer for Vialto Partners, which is a global mobility solutions company. Originally, it was part of PwC, and a couple of years ago, we got spun off and we are part of a private equity company. We were super busy with our TSA exit from PwC, and now, completely, we stand as a global company. We are operating in 60-plus countries and providing global mobility solutions in tax, immigration, remote work, money services, business travel. 

[00:01:56] Kimberly: I have to imagine that's been an interesting past few years, which is what we're here today to talk about, that transformation, but a unique experience that I think many people probably don't get to go through in terms of spinning off an organization to its new separate entity. Perhaps some parallels for our listeners that oftentimes folks are trying to create a new organization or a digital organization while they're running their traditional legacy IT organizations, so perhaps some insights and learnings for folks who are considering that as well.

Maybe to kick off, you could tell us a bit about some of the initial challenges and strategies that were involved in setting up an entirely new entity. You're setting up the new tech and engineering organization while ensuring that you have continuity of service to your customers and maintain your regular day-to-day operations. That's a lot to juggle. Where did you begin?

[00:02:58] Sayee: In my 30-plus years of experience, I led many TSA exits. However, I have to tell you, this is the most complex and challenging TSA exit, and I'll explain why. There is a lot of learning for me and for the entire organization for a few reasons. One is as I was recruited for this role, we were just even before that day one exit from PwC, I have to start the organizational design and the charter and why that structure is that way and how it is going to help everyone's needs, so the journey started from there. Of course, some of very talented digital side, the product management and engineering people came from PwC into my organization, but other than that, I need to start from ground zero creating the structure and all.

I spent quality time in first understanding and listening from the rest of my peers, so the chief operating officer, chief financial officer, chief growth officer, and CEO and chief people officer, chief legal and compliance officers. After listening from all of them their expectations and what way Vialto is different, it's truly, truly a global company supporting 60-plus countries and so many thousands of clients across the world. I started with organizational design and then recruiting the right talent for these ones. I'm proud to say that in 45 days, I was able to recruit my entire organization, that is my next-level leaders.

[00:04:52] Kimberly: That's pretty amazing. 45 days, wow.

[00:04:55] Sayee: Yes. Thank you. Then from there, we went to the next level and all. First, I will answer your question broadly, and then I can go into the details. The TSA exit, the reason why it is very unique and most challenging in my 30-plus years of experience is, first of all, in PwC, various applications are designed, built locally in various territories and they were all hosted with different cloud providers and sometimes it was on-prem solutions and all, and various financial applications, various HR, document management systems, CRM, time tracking, all these things we had many.

During this TSA exit, we took a decision to bring them and unify and create more global tool set in the first attempt itself. It's not like first we'll get to the TSA and then we will do this unification later. We did it during the TSA.

[00:06:00] Kimberly: We're doing it all at the same time.

[00:06:02] Sayee: Exactly. Because of that, what we did is many finance, many HR, many time tracking, many document management systems, we moved to one global system. I can tell you very proudly that from that perspective, Vialto is truly global today and we are able to provide the global tools and platforms and provide more unified experience for not only our employees within Vialto but also for our customers. That is very important for us.

Now, coming to the operating model perspective, going from waterfall to totally agile across the company, and then ensuring that the team is coming along with this change moment. That's a big thing, actually. It's not just technology and digital organization. We need to partner with our business leaders and we need to come together in this journey. That's another massive change moment.

[00:07:07] Kimberly: Was that something you were socializing with them? Because you mentioned how you went around and got feedback from all your peers, and in that 45 days, you were also bringing on board the whole organization. I imagine the socialization of like, we're moving to this new way of working was a key part of those conversations.

[00:07:25] Sayee: Absolutely. That's very important because we all need to go in this change moment together. That's why we created, in fact, a digital product delivery model in which for each service line, there is a global head of that particular service line, whether it's tax or immigration or business travel and all. Then there is a head of product management, head of product engineering, all, they partner together as one unified team.

Starting with defining what is the North Star for each of the service lines and how do you bring that North Star down to a meaningful product roadmaps? What products are required? What products we already have? Some we need to enhance versus some we need to build from ground up kind of thing. All that work was done. As I said, in addition to moving from waterfall to agile, we also, because in PwC, many of these applications were built in the respective territories, we wanted to move from project mentality to product mentality. That was, again, another big change moment.

Now, we are further pushing that. We are not stopping our journey at product. We are going to platform centric. It's almost like there are two halves happening, project to product, product to platform.

[00:08:49] Kimberly: You're doing all the transformations. [laughs]

[00:08:52] Sayee: Exactly. Again, thanks to the rest of my peers, as well as my direct reports, this is why we made tremendous progress. Also on the infrastructure side, we have to build cloud infrastructure ground up. We decided to go with OneCloud. When we were migrating all the applications, we ensured that we are moving them to OneCloud. There again, there are two steps involved. One is, moving to OneCloud is one thing, and then optimizing that application within that cloud is, again, that's a second step. That was another massive accomplishment, I should say.

Then even on the automation side, we keep a special focus on robotic process automation part. We absolutely wanted to make sure that all the robots we built in PwC environment, we are transitioning them in a very careful way, because they play a very important role in our efficiencies and in our productivity gains and all. Again, standing up the automation platform from ground up, and again, ensuring the bots are transitioned successfully was great. We also wanted to make use of this opportunity to get rid of few things. This is where the real thing happens when-- Even in my personal case when I'm moving from one house to another house, that's a great opportunity to get rid of--

[00:10:29] Kimberly: Yes. If you haven't opened the box, you can get rid of it, right? [chuckles]

[00:10:33] Sayee: Exactly, so that is the approach we took. We didn't take, okay, let's move all 50 boxes here and I'll figure out later, right, instead of that. By doing that, we were also able to reduce 20-plus applications and we consolidated in this process before even we came onto the other side. Now coming to the people's side, I was also a little bit surprised to see that we had literally 30-plus software vendors who were building software previously. What we did is we identified four strategic partners, I don't want to use the word vendor partner, they are much more than that, they are four strategic partners. You are absolutely one of those four, Thoughtworks played very crucial role in this. With these four partners, we also ensured that we are bringing to one location in India, for example, that way they are not sprinkled all over the place.

That's one big change moment, moving from 30 to 4, we did this also part of the TSA exit. When we did that, we also created our in-house technology and digital center in Hyderabad in India. That way it's a very thin layer that partners with these four strategic partner resources together, so even our office in Hyderabad is called as Collaboration Hub. It's Vialto T&D Collaboration Hub. That way it is not meant for just employees or strategic partners or something, we all come together there and we collaborate and we build and we support our software. That's another big change moment.

One more thing I want to tell you is the reason why I'm calling this organization as technology and digital, we call it T&D, rather than calling as IT or something else. In the technology and digital, the T part is like our backbone of the organization that consists of infrastructure, network, end-user computing, all those things and then enterprise service management, help desk, all those things. Then enterprise applications, that is the finance related, HR related, marketing-related, and all of them. This is what is-- and then most importantly our cybersecurity or CSO of the organization, these are all part of the T. Now coming to the D side, maybe I should say it's a D square because it's digital and data, they're part of--

[00:14:24] Kimberly: They go hand in hand.

[00:14:26] Sayee: They go hand in hand. That's why when I say digital, it's almost, it's implied, so you cannot do that without the data part. That's another exciting story for us is we created this enterprise data platform team for the first time and we implemented our enterprise data platform in the first 120 days, actually. This is coming up with enterprise data platform, architecture, and ensuring that we are all bringing the right softwares together to create the enterprise data platform. When I created enterprise data platform in my past life, it took like three years. Then the second time it took 18 months. This is the fastest one where--

[00:15:13] Kimberly: It just keeps getting faster every time. You're going to have to go for the world record.

[00:15:18] Sayee: I think again, this is a very important point when it comes to AI, ML, and all other wonderful things we talked, to have that enterprise data platform is crucial. Good thing these days is there are many building blocks available. First time when I did, we have to build everything from ground up, versus now, the building blocks are available, and how you bring them and connect them and create this ecosystem is very important and very interesting, and I learned in every time I implemented.

I have to tell you, I utilized all my learnings from the--

[00:16:20] Kimberly: I don't doubt it, given the scope and kind of all the complexity of everything that you were just describing. I'm also conscious too for our listeners. Some might be familiar with what you mean when you're saying TSA, but maybe just a brief explanation of that for everyone.

[00:16:38] Sayee: The Transitional Services Agreement. Basically, when we say TSA, anytime when you are separating a company, this is what you need to take care of all aspects of separation. This is-- Literally, we had 50 work streams. When I double-click the word TSA, three-letter word, when I double-click, there are 50 work streams. The 50 work streams consist of, not in any particular order, I'm saying, this is creating new infrastructure and then migrating all the applications there, creating new network infrastructure, creating all cybersecurity tools and creating the document management system, CRM systems, and HR, finance, time tracking and ensuring that all these things, the previous data is migrated successfully.

We also need to make sure that we are getting the consents from the client what data we are transitioning. There are very tight checks and balances on PwC's side, what data they're transitioning. It's a massive, massive effort of very carefully how we untangle and bring this data onto the other side kind of thing. All these things together is what when we say TSA, and by the way and one thing which we all may think trivial, but getting the new laptops--

[00:18:09] Kimberly: Oh, that's some of the biggest pain points, I know.

[00:18:14] Sayee: Yes. Yes, and then mobile applications, and when you use your BYOD part. How the applications, everything need to come seamlessly, and then all email migrations, all the attachments, and everything. People were so nervous that, "Sayee, am I going to see my emails from Pwc or not? Do I need to do something?" Kudos to the entire T&D organization. How well they planned and executed this is.

I still remember that night I was in India sitting with my entire team here and we had another hub in San Jose, and then we are working on this and all my, the rest of the C-level executives, they are all really concerned. "Sayee, am I going to see my emails on my new laptop?" When we started on Friday night, I set the expectations. It's going to be there by Monday morning.

They were all thrilled and surprised and they said this is one of the best cutovers they have seen. Again, I want to take this opportunity to thank all my team members, and I especially want to call out where Thoughtworks played very important role for me.

Our core applications that we use, not only internally but also from my client's perspective, those things are supported by Thoughtworks as my strategic partner. They did amazing job in planning and, again, migrating the core applications as part of TSA exit to Vialto side. We could not have done without their support, so thank you.

[00:20:05] Kimberly: Very pleased to hear that that was your experience and glad we were able to deliver for you. There are so many directions and so many questions I could easily get into based on everything that you've been sharing. I think one that's very top of mind for me is, you just spoke to the breadth and depth of everything that was involved in setting up this new entity and transitioning everything over so there was continuity of business operations, that's a lot of change for an organization to manage. Even one of those things could be a really meaty change program. What did you do as a leader to help guide that organization through and manage and balance all those competing priorities and change programs that were taking place at the same time?

[00:20:59] Sayee: Very important question actually, because without thoughtful change management, nothing can be done. All these things like-- it's not just technically making a check mark saying, "We accomplished this." The change management side, there are multiple work streams of change management. Again, one is change management within T&D organization, technology and digital organization, because they're all working on 50 work streams and they all need to collaborate with each other.

For example, when infrastructure team is creating a brand new infrastructure, a new cloud here, all application teams, they need to move their applications there, and what kind of resources they need on cloud for them to get there. My infrastructure teams and my application teams, they need to partner very, very, very closely. That's one kind of thing. While they're doing, cybersecurity team is implementing all the tools required. Now, they need to bring them into the picture. There is a massive change management within T&D organization, how they need to partner with each other. That means they all need to know what the other person is doing, and when they are doing, and how that Lego blocks need to come together. That's one thing.

Now, the different level of change management is partnering with the service line teams. That is tax team, immigration team, business travel team, because whenever we are moving their applications to TSA exit part, that is to the new infrastructure, they need to know that because they need to prep their clients also, saying that, "Hey, this Friday, this change is happening and this is what the impact is going to be." If they need to log in differently or anything. Again, we minimized all those aspects. That's a different change management.

Now, third thing is, while all these things are happening, our practice team, they are absolutely ensuring that the clients are getting that complete confidence of their data, their applications are getting migrated with utmost care, and we are informing our clients on a regular basis. While all these things are happening, we were growing also. That's amazing, so kudos to our-

[00:23:30] Kimberly: Absolutely.

[00:23:30] Sayee: -growth team. It's not only losing our customers in this transition but also growing the business during that time. That speaks the volumes, actually. These are all the various types of change management. Now, crucial thing for this is communicate, communicate, communicate. Without--

[00:23:51] Kimberly: Then do it again, right? [chuckles]

[00:23:53] Sayee: Absolutely. Absolutely. We have something called Vialto Voice. Thanks to one of the leaders focusing on this is a weekly communication that comes, but many things, we don't wait for the weekly communication, but there are so many ways we were communicating, there were many standup calls daily. During the exit time, we also need to partner with our PwC teams because we are separating from their infrastructure. We are moving stuff from their Git repository to our new Git repository and all that. That's another level of change management that we need to collaborate with the PwC teams.

[00:24:47] Kimberly: One thing I want to go back to is you were talking about thinking about this almost like you would with the move. You didn't take all the boxes with you, you used this as an opportunity to simplify your business from a systems perspective, and I know that's something lots of other technology and digital leaders are also grappling with on a regular basis. Is there anything that you can share about what worked well in that approach and being able to simplify and rationalize your systems that maybe others can take away who are going through a similar challenge?

[00:25:26] Sayee: Absolutely. It is a complex process and it is not going to happen overnight like a switch. What we did is, first of all, we created a landscape of all applications we are bringing into Vialto. That itself, you may think it's a easy thing, but getting the complete inventory of all applications we are bringing and then what technology stack that is and who is the business owner and who are the users. Some are more localized to one country, some may be for the region, some are global applications, and some are for certain clients, some are for all clients, so there are many permutations in combinations in this. We first created that kind of inventory.

We also paid special attention on the cost of running these systems in the previous place. It's almost like before we transitioned, I wanted to make sure I have a baseline of information. Even that was not easy, we have to create that. This is human resources costs, infrastructure costs, any license costs, so all these things need to be, a catalog need to be created. Then I have the baseline. Then by working closely with our service line leaders, we also thought about which ones we can consolidate, which ones we can decommission, which one it's going to be a time-bound after which you don't need that application anymore.

We created a application rationalization roadmap. In that we have, what are the things we are going to do in the first six months, first one year, then 18 months and two years kind of thing. In that, 20 things we identified as part of TSA itself, we didn't want to bring them. What I mean when I say bring them means we have the source code. We got the source code and everything, but the applications are not active and they're not used by anyone, so technically they're decommissioned. 

Now, I have to tell you, very, very important pointer for everyone is whenever you think of application rationalization the most important topic that comes along with that is reusability. Focus on reusability is absolutely critical. We started creating what are our reusable assets, catalog of reusable assets. Now, it's interesting, there are various levels of reusable assets. Again, when you double-click on reusability, let me give you an example.

Design system. We want to make sure there is one design system using which we are going to build all our future products/platform. That way our experience, the customer experience, they should not feel like they're working with the 10 products behind the screen.

[00:28:34] Kimberly: Feels unified.

[00:28:35] Sayee: Exactly. The design systems absolutely need to be reusable. There are many enterprise components, obviously, they have to be reusable. Like when I say document management system, it's one system everyone should use. One dashboard system, everyone should use that. Then like one single sign-on system that how everything need to use. Some are enterprise level, you may say, Sayeed, that's pretty straightforward, but then when we come down, questionnaires are the OCR capabilities, how we are going to read from structured or unstructured text and all, translation. These are all capabilities we should build one time and reuse.

There should not be one translation capability for tax and one for immigration, one for business travel. Those are the things we created a very crisp catalog of all reusable assets. That also helped us how we are going to consolidate few things. That's why I wanted to call out the importance of reusability whenever we talk about operationalization. 

The other part I want to tell you is, while the first part of moving these applications to OneCloud, I'm super excited about that accomplishment. However, as you can imagine, we cannot stop our journey there because we need to optimize these applications to this new cloud.

[00:30:17] Kimberly: Just getting everything to the cloud is not the end game.

[00:30:20] Sayee: Exactly. Now, optimizing these applications, the second point I want to tell everyone, maybe this is quite obvious for all of you, keep an eye on your cloud costs because many people think that, "Okay, I'm moving to cloud because I want to reduce costs." Answer is yes and no, but cost--

[00:30:40] Kimberly: If you don't do it right, it can cost more actually. [chuckles]

[00:30:42] Sayee: Exactly. It's almost what I need to tell my three kids every day to turn off the lights, turn off the fan when you are not using, otherwise the meter is running. Right?

[00:30:55] Kimberly: That's right.

[00:30:56] Sayee: The cloud optimization is a very important work stream for us, and that is helping immensely too.

[00:31:03] Kimberly: I'm assuming a lot of that upfront work you did is what allowed you to be successful in maintaining that business continuity and maintaining that customer focus while meanwhile, everything was going on behind the scenes. Can you talk a bit about what you viewed as being instrumental in maintaining a high level of business operations and customer centricity?

[00:31:40] Sayee: Absolutely. It is very important to think of the business continuation in all these aspects. The 50 work streams I said when we were doing this, every work stream, there is a special focus on what kind of impact it can cost, how long the impact is going to be, and who will be impacted, and during that impact time, if they need help, is there a plan B, how they can get their work done kind of thing. We have a massive, massive plan for each of these things.

Again, asking these questions. Is there impact? How long? When are you going to do this? Who will be impacted? What are the things they can do during that time, what they cannot do? The things that they cannot do, can they do differently during that time? That kind of plan was created for each whether it's application migration or data migration and all the aspects. I strongly recommend that, ask those questions and then come up with that.

Now, the other thing I'll tell you, we also need to think of what if that won't go correctly. When I'm moving my core applications on Friday night from PwC infrastructure to Vialto infrastructure, on Saturday morning, everything should be up and running. If there is a problem, if they need little more time to do something, what am I going to do? Am I going to switch back? I'm going to provide them the continuation on the other side? Those are the what if kind of questions we have to ask, because what I tell my team is one of the words I allow in this exit is identifying risks.

Many people may think risk is a bad word, but I encourage strongly team to think about risk, which is a potential issue, before it becomes an issue, what all can go wrong? Then not just identifying risks and then coming up with the mitigation steps for each of them. The risk mitigation is important. Also prioritizing them by understanding the risk exposure. May not be not all risks are costing--

[00:34:05] Kimberly: Equal.

[00:34:06] Sayee: Right. We have to understand the risk exposure and prioritize them, and then tackle them. Because some places I need to have a small crew to take care. Some places I need to put a massive crew to ensure the mitigation plan is taken care of. 

We also decided to inform at every milestone when there is a massive migration is happening, it's not like, okay, we started and we finished. In between there are five steps, and we made sure we are updating our stakeholders. Hey, we completed one out of five steps, two out of five steps, three out of five steps. When they need to validate things on their side also, and give me thumbs up saying, "Sayee, we are looking good." That kind of collaboration is important, so that's why I just wanted to connect the dots here.

[00:35:06] Kimberly: Yes, absolutely. Thank you for doing just that. I think one other, well, not just one other thing, there's plenty of other things. We could probably chat about this for a couple hours easily, but you also mentioned and talked about it today already, the move from being application-centric, but that then meaning it's the shift to being not just product, but really, platform and enterprise data platform on top of that as well. Can you talk a bit now that's up and underway, key reflections on those transitions, and I think what learnings and value you feel like making those investments and moves as part of setting up Vialto has brought to the organization?

[00:36:02] Sayee: Absolutely, Kimberly. This is a place where I want to tell you what went a little bit wrong in my past experiences and how I incorporated those learnings here, if I take enterprise data platform. Many companies, what they do is, first of all, they will try to build everything by themselves. It's going to take a long time. It's going to cost a lot. You may miss a very important opportunity. This is where there is no reason for us to build everything ground up. We have so many tools available, and then picking the right ones and ensuring that they're all coming together will help immensely. That's one thing.

Second, in the beginning stages of big data days, so what people used to think is, "Oh, just bring everything you have. We'll decide later what we want to do with that." Because of that, a lot of garbage was dumped into the data platform.

[00:37:00] Kimberly: Yes, that data can get really messy when you're dumping everything in there.

[00:37:03] Sayee: It is. It will become a swamp, and no one is going to do anything with that. Then they will say, "Oh, we are not using because the data is not clean and all. That's why the data quality is very important. That's a special consideration we kept it. The other very important thing, even more important than the past two things is, this time, we went by understanding what use cases we want to enable. Based on that, we identified what data sources, what data need to be fed into the data platform.

It's not like, let me get everything into the data platform for two years, and then I'll decide what I am going to do with that, versus, we started with use cases that we collected from our service line leaders and all, and now for each use case, what data they need, and where is the data, and what is the quality of the data, and then ingest that data by applying the data privacy rules. That's very important, because what data it is okay to bring and what data we cannot bring based on our client contracts and all. That's a very serious consideration for us.

Also, building the right level of data authorization rules is very important for us because again, who can see what data, not only geographically, but also the client's data and all. We take that extremely, extremely seriously. All these things need to be thought through in the beginning itself. Those are the examples of keeping an eye on bring more use cases-based thing, and second is, don't try to build everything by yourself just because you can build. We, technology people, we would love to build, but our strategy is build, buy, partner. All three are important for us, so when to use which one, that's what we did. Then the data quality part, what I said. If you don't pay attention to the data quality upfront, then you are going to spend a lot of time later.

[00:39:24] Kimberly: I love that. Beginning with the business need initially, and then everything flows from that I think is a great rule of thumb.

[00:39:31] Sayee: A thing I want to tell about the data platform is, many companies, they decide to build one data platform for their enterprise applications, like for finance, HR, operations, marketing, and all, and then one they will use for their service lines, like the data products for tax or immigration, and business travel and all. In our case, we built this platform keeping in mind both the needs. We are using one data platform for enterprise analytics, as well as data products for the service lines. That way, the cost of building and maintaining this platform and all, it was kept in mind, it was built very efficiently.

[00:40:18] Kimberly: Yes. The whole mental model I've had for our conversation today, and maybe it started with the moving analogy was, you're building a house, and it's kind of your dream house, you're getting to design based on all the past experiences you've had living in other houses, and you've decided what you're going to bring. You're not doing everything custom, some stuff you're bringing in off the shelf for the house, and the end result is a beautiful structure that's designed to last for the organization.

Well, Sayee, I think no conversation in this day and age is complete if we're not talking about AI, [chuckles] in some shape or form. How has the role of generative AI or other emergent technologies played in this entire transformation that you've shared with us today?

[00:41:10] Sayee: You're absolutely right about no conversation is complete without talking about AI these days. You take any analysis results or anything, that almost 70% to 80% of the CEOs if you ask them, give me your top three to five strategic priorities, AI is certainly part of that. The reason why I want to expand the board-- we come to GenAI, but I do want to start with AI and machine learning first, and then get to GenAI.

This is one thing again, it's like many companies, they may be in the quite early stages, and they don't know where to start, but there is this lot of push on AI. For all of them, I want to tell you, you are not alone in this situation. There are many companies, they are doing many things great, and now they're trying to step into this new frontier kind of thing. My suggestion to those of you, by the way, the reason it's even a very passionate topic for me is as we speak, I'm currently pursuing my Doctor of Business Administration in artificial intelligence and data sciences.

[00:42:28] Kimberly: Oh, amazing, wonderful.

[00:42:30] Sayee: Thank you. Next month is my defense.

[00:42:33] Kimberly: Okay.

[00:42:34] Sayee: Hopefully, I'm this close. Now, coming back to the reason why I'm saying AI and ML, in general, before you go to GenAI is, first of all, again, start by closely partnering with your business leaders. This is not some technology initiative, or it's not some fancy thing for me. It is, how is this going to help my business leaders? We need to start from there, because having a powerful use case will always help making it into your reality.

Now, you may have all kinds of use cases. Now, then comes the question. You need to be carefully in prioritize them. This is where we apply something what's called Pain Gain analysis, basically. For each use case, the level of pain, level of gain, and based on that, you're putting into your--

[00:43:24] Kimberly: How does that equation balance there?

[00:43:26] Sayee: Exactly. It's a two-by-two matrix. When you put that and you can decide which use cases first you want to tackle. This includes, do you have a clearly articulated AI strategy in your mind? Like, why are you doing this? Is it just because others are doing, I want to do, or how this is going to help you in your company's overall strategy. That's very important, having that data point in your mind.

Then you need to understand that, do you have the data required for that? Without having data, again, there is no point of talking about this. Then do you have the talent required? Again, it's a combination of our in-house Vialto talent, as well as our strategic partners like Thoughtworks and others, how they can help us. 

Let me tell you, I conducted like two-hour session to the rest of the C-level executives on what is AI and ML starting from basics kind of thing. That shows the level of importance in the company, and also, thanks to the rest of the C-level executives that they really wanted to understand what it is and how it can help Vialto and how it can help our clients. We did that kind of data awareness starting from top-down kind of thing. We did that.

Now, after you have the strategy and data governance part is important, and then, people part, we talked, and then comes the infrastructure. Do you have the right infrastructure? How you are going to do that is important. The culture is absolutely important because whenever you try something new, you need to prep people, that excitement is important. Because many times when people don't understand the power of AI and ML, they may think, it's a threat. Their jobs are going to be impacted kind of thing, versus it's totally opposite, actually.

[00:45:53] Kimberly: No, it's making life easier. [chuckles]

[00:45:55] Sayee: [chuckles] Exactly. Now where they spend their time is going to be much more intellectually challenging and how it's going to help. Whether it is descriptive, predictive, or prescriptive, this is a awesome space, but again, we need to change management. Going back to your previous question, we need to bring the teams along with us. That is the thing. Now, my humble request to everyone is, while GenAI is a hot topic, you do think of AI/ML, in a more holistic way in which GenAI will come absolutely as one of the important building block in that. 

Now, GenAI per se, it didn't start yesterday or the day before yesterday. Thanks to ChatGPT and Gemini and all, like everyone is talking, but the deep learning and neural networks, it started from 1950s. The thing is the resources we have, the cost of computing power or memory, or the cloud and--

[00:47:05] Kimberly: Makes it so much more accessible to everyone now.

[00:47:08] Sayee: This is really like turbocharged what we can do now. Now, the large language models is amazing, amazing things. I will tell you, we are all still scratching the surface here and the kind of agents you can create on this GenAI perspective, how it's going to help our productivity perspective. This is amazing, amazing opportunity in front of us. Again, whether it is a basic, you're using Microsoft Word or PowerPoint or Google Tools, whatever it is, you know how it is helping. Now, I can explain, I want to create a PowerPoint in five slides to do this, this, this kind of thing. Boom, I have a starting point, and then of course, I can modify and change from there and all. This is where the copilots that's coming into the picture is phenomenal.

We are still in, again, the implementation stages. The way we approached GenAI is first thing first, again, protecting our clients' privacy data. Privacy is very important for us. We came up with the employee policy of when to use publicly available GPTs and when not to use. What kind of information you can use? That's very important because we absolutely don't want to use a publicly available ChatGPT kind of thing to search some PAA information there kind of thing. 

Now, we also created a private GPT within Vialto. This sandbox is absolutely safe, secured. In this, when I'm loading, for example, 10 documents, and then I am doing some intelligent search on top of that to summarize or to bring the insights from that and all, we are making sure all the key information is scrubbed in that. You are not showing any personally identifiable information there in this one. This is where you need to think of these aspects from the beginning.

[00:49:35] Kimberly: Yes. When you're designing.

[00:49:35] Sayee: As I said in GenAI, when I take it to the two parts, one is publicly available GenAI is what you can do, what you should not do, that we need to be very clear. Second, we created that private, safe, trusted GPT within Vialto, where this information is not going back to train their algorithms. That's very important because when you go to the public sites and when you are searching something, they are using our data also to train the models. Whereas here, it's publicly- it's a private, ChatGPT kind of thing, so the information stays there. They are not using that to train the model. 

Ensure there are right employee policies and you are educating your employees so that we-- These are powerful tools, but we need to be careful. Again, we are using them with the utmost care and responsibility.

[00:51:03] Kimberly: Being thoughtful about how you set it up and making sure you have all the dimensions in place or at least thought through before embarking on that journey, that's very good practical wisdom. Sayee, thank you so much for today's conversation. It was really quite impressive to hear what was involved with the transition, creation, and transformation of the technology and digital organization at Vialto. Kudos to you in really accomplishing so much with this effort in a relatively short amount of time.

Thoughtworks has very much enjoyed being a partner on that journey with you. Best of luck in your doctoral defense coming up here soon. I'm sure that'll be an exciting time.

[00:51:54] Sayee: Thank you. Thanks, Kimberly. I really, really enjoyed our discussion and your thoughtful working questions. That was amazing. I also want to take this opportunity to pass your congratulations to my entire T&D Organization, which is a combination of our employees and our strategic partners together, and the rest of the C-level executives in the team. Without their understanding and support, we could not have made this progress.

They are all very digitally savvy, and they all want to make sure that we are helping not only our internal community but also our clients every single day. Thank you, and looking forward to partnering with you.

[00:52:44] Kimberly: Yes, likewise. Thank you so much. Thank you so much for joining us for this episode of Pragmatism in Practice. If you'd like to listen to similar podcasts, please visit us at Thoughtworks.com/podcast. Or if you enjoyed the show, help spread the word by rating us on your preferred podcast platform.

[00:53:12] [END OF AUDIO]