The Chief Digital Officer of professional services firm, Genpact, explains how the huge company uses collaboration to achieve high-performance results.
The bigger your company, the more important it is that every team member is on the same page. When you're as big as Genpact, with 90,000 employees and twice as many partners, then collaboration is a top priority. Sanjay Srivastava is well aware of the challenges. As Genpact's Chief Digital Officer, he is front and center at the effort to make sure the disparate teams and employees within the company are working successfully in a collaborative organizational culture, as well as offering a satisfying customer experience.
For Sanjay, there are three main factors that need a strong collaboration platform within a company. It starts with the idea of the business as a connected ecosystem that drives a collective intelligence. Then there's the concept of continuous learning and innovation that requires a collaborative framework to be successful.
Finally, there's the convergence of domains, the ability to pull people together from different disciplines, with different experiences, and across geographies and time zones in innovative ways. Specifically, it means having a kind of distributed collaboration: self-aggregating teams that come together to solve a specific problem, disband, then re-aggregate in a new formation to solve a different problem.
Collaboration plays a large part in working with customers outside the company, especially if your company offers business consulting services. That means having the kind of operational intelligence not just to be able to understand what outcome the customer wants, but to offer them a stellar customer experience in the form of seamless information sharing.
The concept of collaboration isn’t limited just to other people. The future of business collaboration is going to involve being able to utilize artificial intelligence (AI). According to Sanjay, AI is actually a paradigm shift in how we do business. With deep learning, computer networks can now learn unsupervised, even if the data is unstructured or unlabeled.
But AI in the current enterprise space doesn’t yet have a 100% prediction accuracy. You still need human expertise in data science and algorithms to develop and apply AI to produce results and predictions that are comprehensible and actionable by your customer.
Watch the video above to see what Sanjay considers the key factors in creating a collaborative environment for your company. That includes creating a work culture open to change and different ways of thinking about how to do business and finding people who have a different way of approaching a problem, as well as those who can act as effective translators across the different teams.
- Why is Collaboration Important to Genpact?
- Is Collaboration Strategic to Genpact?
- Historical Development of Genpact
- Culture Change and Collaboration
- Collaboration Challenges in Professional Services
- What are Metrics or KPIs to Measure Collaboration Success?
- Collaboration and Innovation
- Customer Experience in Professional Services
- Artificial Intelligence in Professional Services and Collaboration at Genpact
- What is Genpact Doing with Envision Virgin Racing and Formula E?
- How Do Data Scientists Work With Business Experts at Genpact?
- Advice to Business Leaders on Improving Collaboration?
This transcript has been lightly edited.
Michael Krigsman: We're discussing the impact of collaboration in one of the largest professional services firms in the world. Our guest is Sanjay Srivastava who is the chief digital officer at Genpact. We are live in front of an audience at Logitech.
I'm so grateful to Logitech for hosting CXOTalk today. I want to give a shout out to Logitech's chief information officer, Massimo Rapparini. Massimo defines--
Michael Krigsman: Massimo defines the concept of the innovative and transformational CIO. Tell us about Genpact and tell us about your role.
Sanjay Srivastava: We're about 90,000 employees. I run our digital business globally. We built a business around automation, analytics, AI, and experience, customer experience as a discipline. Then we use that to actually help large corporations transform their current business operations and help them move from today to the new version of who they want to be tomorrow.
Michael Krigsman: How does collaboration, therefore, fit into your business and to your employees working with customers? Let's start there.
Sanjay Srivastava: We work across large corporations. I've actually personally been a startup CEO and I've built small companies with very agile, small teams. You look across the ecosystem of small companies to large companies, I think two or three things stand out really clearly.
I think there's this whole notion of a connected ecosystem that comes and drives a collective intelligence. There's been so much work that's happened around collective intelligence, talking through, you know, the whole notion of it takes a village to bring up a child.
It's the same idea to get a product out to market; you get a concept into an industry. You drive transformation. It takes a very large body of knowledge to come together. We call it collective intelligence and there are many angles to this. We can come back and look at that, but that's one.
I think the second thing that's really important, you see this more in startups and, I think, increasingly now across a span of industries is this concept of continuous learning and continuous innovation. That requires a bed of connective tissue or a collaborative framework for that continuous learning, continuous innovation to happen.
Of course, the third one, increasingly the best ideas and the highest and the most transformative outcomes come from convergence of domains, a convergence of principles. You can almost think about it as the low hanging fruits off the table. When you actually converge two different things, machine learning and, let's say, financial analysis, it's really where you get the best outcomes.
How you drive this convergence of domains becomes really important. Of course, collaboration is again the underlying fabric. It's the connective tissue that brings these three things out.
Michael Krigsman: Why do you say that collaboration is the connective tissue that underlies these things?
Sanjay Srivastava: Well, increasingly, the world is changing. We are bringing in people from different disciplines and different experience sets in a common environment that has self-aggregating teams with distributed leadership. That's the new order of the world, right?
As an example, we as an employer no longer hire for skills. We hire for different things. We hire for critical thinking and we hire for continuous learning.
If you think about for one second, what we're really saying is that, underlying all of this, we're accepting the fact that the skills will change. The skills that we need will change over time. If you hire for one skill today, it's not really going to address the skill need for tomorrow.
Our thinking has changed around the talent that we bring in and how we actually then give them the components. Really, to be able to do all of that, to be able to have distributed teams or distributed leadership, to have this notion of self-aggregating teams that come together to solve for specific problems, disaggregate, move on, and then reaggregate in different ways to solve different problems.
You can't even get started on that without a connective tissue, without a collaboration platform. That's why we're really keen on this idea of how you actually get people together across geographies, across countries, across time zones, across customer bases and, most importantly, across disciplines to come together in innovative ways.
Michael Krigsman: This type of collaboration actually is central to the kinds of projects that you work on and is strategic to you going forward because it's what enables you to do the things you do for the core of your business.
Sanjay Srivastava: Well, we couldn't even get started without it. It's the first thing you've got to get in place. Some of that is culture. Some of that is tools and the fabric, the physical logistics of it, and some of that is just bringing the right people into the mix. All three of those components are critical. We won't get out of bed without it.
Michael Krigsman: I'm assuming that you've studied the nature of this kind of collaboration very carefully in order to package the components together in these ways that you're describing.
Sanjay Srivastava: The key for us has been to unpack each of those three areas. We think about collective intelligence very deeply, partly thinking about our position. We're coming into a client. We're helping understand their current scenario. We're helping them redesign and reimagine what their business looks like in the future. Then we're actually stacking up, building, and deploying a set of capabilities, technologies, people, processes to be able to make that new vision a reality, right?
Then in that environment, the notion of collective intelligence, the pieces of DNA and domain that we bring into play and the attributes from their clients' and our clients' perspectives have to come together in a really seamless manner. Out of that general set of ideas, we need to reimagine, to redesign what the engineered output needs to look like. That takes a significant amount of collaboration and this notion of collective intelligence becomes really important. We have studied it and so we've found some of the things that work for us and other things that we watch out for but it's an important thing to get right.
Michael Krigsman: Your business has changed over time. When was Genpact founded?
Sanjay Srivastava: We were founded about 20 years ago, so we've been a business that's been out 20 years and we've transitioned 3 times already in what we do. We're actually on the verge of a fourth transition.
We started as a business process outsourcing company and we take the work that was job number one for us, perhaps job number 10 or 11 for clients whom we served, and then we do it for them. We do it well. We do it with discipline, with rigor, with quality.
Over time, we got that to a science of processing, so we could actually take a look at a high-level metric of something that a large Fortune 500 company might report to Wall Street, and we break it down into its end-to-end process. We'd actually then be able to prescriptively change some knobs, some levers, and some dials to be able to get to where the outcome was.
We transitioned from that to being an intelligent operations player, which is to say, instead of doing all of the work manually, we started doing it with additional technologies. Lots of experimentation. Lots of trials. Lots of mistakes. Lots of learnings.
Through all of that, we really got it right and we became a leader in intelligent operations. That was stage two for the company. It took about five years to get there.
Now what's happening is we've become a leader in digital transformation, which is to say we come in with a set of technologies we've developed in a curated way of applying that that can be done in a replicated fashion across business problems that span different areas of the enterprise. We can do this in scale, so we actually just provide digital transformation services to large clients now.
Even as we do that today, we're sitting and thinking, "What's next for us?" For us, leadership in artificial intelligence becomes a really important cornerstone. As I said, we're not planning to get there tomorrow morning, but we are working on actually thinking about what the future of work is and, in that future of work, how does artificial intelligence change the game. In that change of the game, what is the role of domain, what is the role of data, and what is the role of AI engines? How do we materialize economic value at the intersection of those three things?
In some ways, the long answer but the short version of it is, we're a company in transition. We're transforming ourselves from a company we were to a company we're going to be and we're in sort of step three of this four-step way.
Michael Krigsman: There must have been various cultural changes. With respect to collaboration, how do you get a workforce as large as yours to shift from having their work assigned based on skills to now aggregating teams that come together and go apart, having to learn? How do you do that?
Sanjay Srivastava: I think it's super critical to have a goal and a vision of who you're going to be, not at the CEO level, but at the 89,936th employee level. It has to permeate across the organization. What you really have to do is not only come up with a vision for who you're going to be and what your value proposition is at the board level but find meaningful ways of actually getting it across the employee base.
I think the second thing for us that's been really important in this journey is getting the backbone in place. If you just think about IT systems, part of our business operates in a highly regulated industry where basically we locked on everything in your IT backbone for security for good reasons. Another part of the business where I have employees that are waking up at 3 o'clock in the morning and spinning off our actual server and writing some code and they need to know that they're going to get reimbursed the next day when they come to work.
How do you bring all of this cohesive, different workforce together in a cohesive fashion? The backbone for collaboration needs to come into play. That was very, very important for us.
The third lesson for us has been that you have to bring what I call bilinguals. In the world that we live today, it's insufficient to be a master of English or a master of French. You have to be able to understand both languages in a way that you can bring cumulative knowledge together.
For us, it's machine learning folks with supply chain individuals. It's artificial intelligence engineers with finance and accounting disciplines. The more we thought about bilinguality and the ability to bring people to multiple disciplines and understand enough of two, I think the more we found our way to success. I'd say those three things have been crucial for us in our journey to transforming ourselves.
Michael Krigsman: What are some of the challenges that arise as you're trying to basically accomplish, do business with these kinds of collaboration goals in mind?
Sanjay Srivastava: I'll tell you where you struggle. You struggle when you aren't thinking end-to-end. I'll give you an example. A lot of the work we'll do will take an end-to-end process and we'll want to transform it. The traditional way of doing it would be to take any process. Think about a large corporation that is in the business of selling something. They're going to actually sell it, which means there's a core process. There's an invoice process. There's a purchase order process. There's a billing process, an invoice process.
You sort of think about this as a code to cash process. Traditionally, when you first attack it, what you say is, "Well, let me just take this entire process and break it into its bits, into its components. Let's digitize every single component one-by-one-by-one." Then the reactivated version of those digitized components gives you the end view and, actually, that's wrong. It doesn't work that way.
When you think about transforming a business process, don't take it component-by-component into bits and bytes. You have to think about it end-to-end as a whole and then approach it very differently.
Michael Krigsman: How do you get people who are not used to thinking about it this way to suddenly have this kind of holistic view? That's really hard.
Sanjay Srivastava: It is hard. Diversity counts, so you need to have people with different backgrounds, different predispositions, different experiential curves, and bring them together in a self-aggregating team. That's really important.
You have to have a very clear articulation or understanding of what the end goal is so everyone can play on the same team. They need to have this fabric we were talking about earlier so people can communicate and correlate that they use the same vernacular. Vernacular is really important because you're bringing bilinguals together and you need enough of a commonality that people understand each other. Vernacular, the tools, and diversity are three important attributes of getting it right.
Michael Krigsman: I'm assuming dealing with the challenges is an ongoing battle, like with everything.
Sanjay Srivastava: It's an ongoing battle and the battlefront keeps changing because the edge of innovation keeps moving forward. You feel like you learned this, you got it right, and then the outside edge has just moved and you can't get there. It's an always evolving journey.
Michael Krigsman: Do you have metrics, KPIs, or measures around collaboration at all?
Sanjay Srivastava: I'll give you a really interesting example of something we did some time back. We started looking at email traffic within the company. We essentially built a permission-based database where we took certain elements of email messages, so the "From," the "To," subject line, et cetera, time, and started putting it into a big data infrastructure. Then we superimposed a data science methodology on top of that to really start seeing what the patterns are in the way people communicate.
This is an easy tool to it, but what we found out is that you start getting these spokes of the wheels of the networks, if you will, where ideas originate and I'll send something to you. You'll forward it to someone else. That person will forward to 100 people and so you can spot the centers of these new ideas, new initiatives, and new thinking in the way that emails actually permeated through the organization.
When we started correlating that to our internal performance management system, which is still the old world performance management system, you know, once a year we sit down, do this assessment, et cetera, what we found is the correlation between one to the other was very high. In other words, we realized that there are new ways to think about spotting emerging leaders, spotting issues that will become concerns well ahead of when you would otherwise know with traditional methodology. All of that comes in the back of using a new backbone, a foundation of data, a new backbone of AI and analytics to analyze that and a new approach to thinking about how you measure performance, how you spot for trouble.
Michael Krigsman: Can you boil it down to specific KPIs relating to collaboration or do you focus more on the outcomes of that collaboration, whatever you're trying to achieve?
Sanjay Srivastava: We're very fixated on business outcomes and we think collaboration is one of the ingredients that need to come together. In the end, we measure outcome performance. On the sideline, as a company, we actually measure, audit, and publish the business impact we drive our clients. We've done that for years and we take pride in how that's progressed. We realized there were other components that make that whole.
Michael Krigsman: Collaboration is the enabler of business outcomes. It's one of the enablers that make the business outcomes possible. It's part of the process.
Sanjay Srivastava: Yeah. We think of this as the underlying fabric across this notion of collective intelligence, this notion of continuous innovation, and this notion of convergence of domains really come together. If that connective tissue doesn't exist then these three things don't really come together and they don't operate. But when it comes together well, then we get the magic of all three.
Michael Krigsman: The measurement is of the business outcome.
Sanjay Srivastava: Correct.
Michael Krigsman: You're not measuring the fabric itself.
Sanjay Srivastava: Correct.
Michael Krigsman: You're measuring the results.
Sanjay Srivastava: Well, we're not measuring at the level of -- we're not scientifically measuring and reporting for every single thing, right? But it's very obvious when you look at teams that perform really well that collaboration is actually working well. You can sense it. You can pulse it. You can feel it. You have the dialog and discussion with the team members and what's working and what's not working. It all comes through.
We do a pulse survey. We've been doing this for the last few years now, so we do an NPS on our own employees, a net promoter score ranking, which is to say, "Would you invite a family member to work for the same company?" Then we rank it and then we try and understand why is it that you would or you would not. Through that, we capture a number of aspects in our collaboration, the backbone of the infrastructure that promotes the collaboration. That's the level at which we've done it and we found it very helpful.
Michael Krigsman: Then the measurement, in a way, it's two things. One, you have the indirect measurement that you were just describing and then the other, I guess you could say, it's one of those things that when it's good, you see it and you know it.
Sanjay Srivastava: When it's not right, you can feel it. When it's not right, when there's something wrong, you can feel it very quickly. The reason you can feel it quickly is because all the large results are broken down into iterative, smaller, small bite wins, right? Because we're agile, because we iterate quickly, you can realize there's a problem early in the cycle and then you try and jump in and understand why it is. If it's collaboration, then you address that. If it's something else, then you go and address that as well. Yeah, at the outset, the two boundaries, you got it right.
Michael Krigsman: Sanjay, you've been talking about digital transformation and you've been talking about innovation. Where does collaboration fit into enabling transformation and innovation? Is it something different? Is it just part of the same fabric that you've been describing?
Sanjay Srivastava: I think innovation and digital transformation are very similar. It's the same idea. We find that this notion of connected ecosystems, which isn't just to say that it's the people from different disciplines that are connected but actually the underlying data sets and building a foundation for that data as a platform upon which to work is a super important part of any digital transformation project we do. That's number one.
I think the second thing we've learned is this notion of experience, the customer experience, and journey mapping as a way of defining the true north because too often when you do digital transformation, you start from the foundation of saying, "We've got to take cost out. We have to improve this number. We have to take that number there." Actually, the right way to drive the decisions that need to be made on a day-to-day basis is to use one true north compass.
For us, we've come to realize that journey mapping and customer experience is that right true north. That allows us to think end-to-end. That allows us to think comprehensive and composite. That allows us to truly reimagine the end-to-end as opposed to automate bits and pieces of that and bring that to life.
Michael Krigsman: Why the emphasis on customer experience? Typically, when we talk about customer experience we're saying a product but you don't sell products; you sell services. Tell us about the customer experience dimension.
Sanjay Srivastava: That's a great topic because most of us have grown up thinking about experience in the context of a product. The iPhone is a great example of it.
We don't build products. We actually transform services. We work with clients who take their supply chain mechanism. We take their financial crimes outfit. We take their consumer banking relationships. We take the way they do processing in the insurance industry. That's a service. That's an end-to-end service. We digitize that and we transform that.
In all of that, to really get to the right endpoint, you have to be thoughtful about how does the end customer of that journey experience it and reduce the friction in that journey and enhance the value and experience in that journey. It's super critical because, in the end, the reality is, in digital transformation, what you're really trying to go after is a differential value proposition. You're not really just trying to take costs out of a system.
You trying to take cost out of the system is a given. It's going to come anyway if you digitally transform.
Michael Krigsman: Especially today.
Sanjay Srivastava: Especially today because you put automation in; you get a lot of returns on it anyway, so that's a given. That's no longer an ask.
The ask now is, how do I drive a better top line? How do I differentiate myself in the minds of my customers? How do I create a journey or an experience for clients that is sticky and people want to come back to? On the back of that, how am I more competitive? How do I manage my regulatory compliance better? How do I manage my drug safety better?
I think the whole business has shifted to driving better experiences and using that for top-line growth and long-term directional changes in the business. It's become a true north and we work with boards all the time now in helping their companies, mostly Fortune 500 companies, transform themselves.
I have to tell you the two things I get asked the most about. One of them is experience and the other one is AI. Those are the two of the biggest questions that come through and it's just because that's where I think the rubber meets the road.
Michael Krigsman: Sanjay, when we talk about customer experience, most of the time we're thinking about products and product companies, but you're a professional services company. Isn't there a song that says, "What's love got to do with it?" What's experience got to do with it? [Laughter]
Sanjay Srivastava: [Laughter] Yeah. That's a great question. I tell you, you're right. Most people who think about experience in the context of a product. That's not the case. The experience is really critical in the way you think about a service.
I'll give you an example. By the way, we see this across clients. We see this across boardrooms. What we find when we talk to boards of Fortune 500 companies, there are two questions that come up most all the time. One is experience. The other one is AI.
In your personal life, you probably use some e-commerce site to order a variety of things. I do a few and what you typically find is most of these e-commerce sites will have this recommendation engine. There's AI in the back of it that's predicting what you are mostly to buy. When you log on and get onto it, it'll say, "Here are ten things that we think is a great fit for you."
The reality is, today, the accuracy levels are pretty low, so two of those ten things you might want to buy, then you probably would never touch. What happens when you change experience to the point that you make the AI much more predictive and now nine or ten things I can recommend you're likely to buy? Of course, the first takeaway is, you've just logged onto the site. You were going to buy something else and you see these nine things being recommended or ten and they're exactly what you wanted. It's like a timesaver for you. It's just a fantastic experience. That's great, but that's not enough.
What's happening on the back of that is, now the e-commerce site has the ability to actually change its business model because now, instead of waiting for Michael to log on, on Saturday evening, and get this nine out of ten things right and then be able to order it, how about if I just ship you ten things? You open your door Monday morning. You step out and the first thing you notice is a box on your doorstep. You look inside the box. There are ten things. Guess what. Nine of those you would have ordered anyway through the week. You take those nine out and you leave the one that you didn't want. You put it back on the doorstep.
What happens after that? the company, the e-commerce company now is in a different business model. They're pre-shipping the things you're most likely to buy and they're focusing on the reverse logistics of bringing that one thing back. That accuracy keeps improving.
Just in that example, if you think about it, I've changed the experience. I brought it in a different experience for you. In so doing, I've actually changed my fundamental business model. That's why experience is such an important thing to focus on because it changes the playing field. It gets the company on a journey of transformation well beyond the bits and bytes you could do today.
Michael Krigsman: Let's talk about the role of AI. I know you've been chomping at the bit to talk about that. It's a very important part of what you do. Let's weave that into this process now.
Sanjay Srivastava: It is a very important part of what we do. We do work for clients and the future of work is going to change with artificial intelligence. We think AI is not an advancement in computing. It's a complete paradigm shift. We're no longer telling computers "what if/then" and programming it to do that at a speed faster than humans and do it autonomously across the world, et cetera.
We're actually now, instead of that, we're telling the computer this is the right answer, this is the wrong answer, here's the question, and you figure out how to program yourself. That very simple change in design philosophy has a very profound impact downstream because it allows us to automate the last mile that most enterprises have not been able to in the last 10, 20, 30, 40, 50 years of automation be put in.
As that last mile now gets automated, the work of the future ends up being different. Obviously, we do work, as do most of us. It's something that it becomes really important to us.
As we've sort of double-clicked on it, what we've found is that, to build artificial intelligence systems and to deliver economic value in any situation, certainly in the enterprise, you need to be able to take engines, which are AI engines, and contextualize that. We call it goal orientation machine learning. We use distillation techniques and actually data science. We use contextualization in the way we automate different things. These words of goal orientation, contextualization, and distillation are all concepts that come of understanding the domain, understanding the subject that you're trying to automate or to digitize, that understand the handshakes the processes will have on both sides.
This bit about how do you make AI productive in an enterprise really comes back to understanding the domain, being able to use the data in meaningful ways, and then leverage large AI platforms that are increasingly commodity platforms available in the industry. For us and for our company, that is just the center spot of where we want to head because we can use the domain expertise that we have and help our clients to be able to materialize value from AI systems which, by the way, they're going to get significant benefits on because it automates the last mile, so it's big for us.
What's really relevant in the world of AI, tying it back to where we started the conversation on collaboration, is the fact that AI systems, at least in the enterprise today, don't get you 100%. I can get you to a prediction accuracy of 80%, 82%, 83% off the bat but you don't get to 100%, which means you need a human in the loop to address the pieces that today AI doesn't address. Then you have to actually learn from those exceptions and be able to tune your AI models to be able to progressively get from 82% to 84% to 86% to 88%.
It's really important to understand that AI is actually changing the future of work but, to make AI happen, you have to have collaboration in a way that human in the loop is actually involved in the cycle and continually tuning and accommodating for things like model drift, data drift, and so on and so forth.
We're very passionate about AI. We find a lot of potential in applying that into client environments. We're cognizant that there is a human in the loop component that is critical to getting it right.
Michael Krigsman: A core part of this collaboration in relation to AI, it seems to me from what you're saying and also from what I hear from other companies, is you've got the combination on the one side of the domain expertise so the people who are experts in the business, but then, at the same time, you need to have domain expertise in data science, in algorithms, in how to develop and apply AI. This becomes a driver of a new kind of collaboration that needs to take place.
Sanjay Srivastava: That's exactly right. This notion of bilinguality becomes really important because you need people from different disciplines. I often go into enterprise clients and the first question I get is, "Can you do a quick assessment and give us a sense for where we are in the AI journey?" As you do that, the next question that comes up is, "Well, how many data scientists do you have?" or, "How many AI engineers do you have?" I actually almost never ask that question because, for us, what's most important is, how many people do you have that understand enough of AI and understand enough of your domain so they can bring the two things together? I think that's really critical to get right.
Michael Krigsman: You have been working very closely with something called Formula E Racing. I don't know if too many people are familiar with it, but it's basically taking Formula Racing, Formula 1 Racing, which I think everybody knows, has heard the term, but with electric cars.
Sanjay Srivastava: Right.
Michael Krigsman: That bears directly on that conversation we were just having about business domain expertise with AI, data science, data expertise. Tell us about that.
Sanjay Srivastava: It does. Formulate E, I think, just quickly, is the next evolution of Formula Racing where it's based on electrical engines. It's a really interesting event not only because it's the future of mobility but, actually, the demographics and the people that are participating is a much broader cross-section and much more engaged.
What we do with Formula E racing isn't just helping the company you support to win the race and the data science isn't just about finishing first. Actually, the algorithms we're building are going to end up impacting the future of mobility because they're the same algorithms and principles that we use today that we develop today that will be used tomorrow to make electric cars safer, to make mobility and transportation more cohesive and more evolutionary.
There's just a higher purpose goal that when I find my teams getting engaged across different projects when we get involved in purpose-driven projects, you can see a very different level of engagement. We are looking at drivers, we're looking at cars and engines, and we're looking at the skill of making decisions at high speeds on dangerous roads, and then we're looking at data science and artificial intelligence to predict things, to be able to play out different scenarios and make better decisions. Bringing those two disciplines together has just been a great, fun exercise. It's a meaningful indicator of how and what enterprise work we do as well.
Michael Krigsman: How do you bring together racing teams? These racing teams, just to make it more concrete, these are racecar experts. They run races. You've got drivers and you have engineers. Tell us briefly about the racing team first just to set the stage.
Sanjay Srivastava: Sure. Yeah, absolutely. Just by way of background, this is a different race. Most races, you think about as a race where you have a fixed distance you need to cover and then how fast you get from A to Z. This race is a little different because what's happening is you're racing for a specific amount of time and then, based on how you use the energy, you get a little bit of a boost to it.
Really, what that translates into is predicting how many laps you will have to drive because you have a fixed amount of energy. All cars start with exactly the same amount of energy.
Michael Krigsman: In the battery?
Sanjay Srivastava: In the battery, that's right. It's an electric car battery, so you have the same amount of energy and the decisions you have to make is, how do you drive in a way that you can get the maximum distance with the battery amount you have. You can go really conservative, you can go slow, and so you can go as long as possible or you can be very aggressive and then your battery dies out.
The decision you have to make is, look, if you end the race with 20% of your battery remaining, actually, you haven't done well because you could have used that battery and gone farther ahead. If you end the race because you ran out of battery and you couldn't get to the endpoint, well, that's a disaster as well. This game is about predicting accurately how many laps you have remaining at any point in time.
Now the reality is, a second or two seconds or ten seconds before anyone can predict how much lap is remaining because you can see it, right? But you go back 15 laps before the end of the race and being able to predict exactly that you've got 15 laps remaining is really important because you can make real-time decisions on who you want to accelerate and overtake, when you want to pull back, and all those sorts of things.
Now, with that backdrop, think about two teams of people. There's a set of people that are experts in driving. If you tell them, you know, "Take the curve sharp, turn here, and overtake this person," they'll do that really well.
Then you've got a set of people that are sitting there using AI engines to predict what's the competitive driver going to do. What have they done last time? What can we predict about their behavior next time?
What happens to the friction in this road because we're racing as we did in Berlin on, basically, a runway at the Tempelhof Airport? Basically, the friction on that road is very different from what you would experience on a normal road because it's a runway surface for aircraft.
How do you take all these permutations, combinations of weather, temperature, friction, road conditions, and traffic and then use that to project how many laps do you have remaining? You've got people that really need to get that right and you have people that can use that information and make that decision. They have to operate seamlessly with trust, with integrity, with communication.
It's not dissimilar to what we do with other enterprise projects except, most enterprises, if you think over the life of a company and the work we do, what happens in a year gets done in 45 minutes in a race. We have the ability to run that cycle in 45 minutes, learn from it, come back the next time and run it again in 45 minutes. That view, that 45 minutes is what we would normally do in a large enterprise over a year, so it's a really good way for us to learn and iterate quickly.
Some of the biggest things I think we've had to focus on is, how do you get two sets of people with different backgrounds, different disciplines, frankly, different demographics, personal dispositions to come together on a common objective and collaborate in a way that they can trust each other to make quick decisions? That's just has been really great exercise.
Michael Krigsman: From a collaboration standpoint, actually, how do you do that? I have to imagine that on the Genpact side, which is the data science side, okay, they're nerds. They're data scientists and programmers. On the racing side, they're racing geeks. These are really different profiles.
Sanjay Srivastava: Yeah, they're very different. I think that difference is the beauty of what they can do together. That's the whole reason why this works.
To answer your question, I think of some of the things that have to come together as the glue to make it happen. This ability to have trust and trust doesn't come from an implied trust or assumed trust. It comes from this iterative process of trying, experimenting, seeing the results, and then building your confidence around it.
Michael Krigsman: Especially in something like this with a race.
Sanjay Srivastava: That's right. Exactly right.
Michael Krigsman: It's sort of all or nothing.
Sanjay Srivastava: That's right. That's exactly right. You have to deliver sustained performance for a period of time that you trust the other half and say, "Wow. They've got it right. I've got to do my bit and, if both of us do our bits, we'll end up together much better." I think that's one piece that's really, really important.
We've also learned other things. Going back to experience for a minute, experience is a really important part. For me to come up with a recommendation to do this or that to a race driver at the heat of the moment at the speed he or she is going at with all of the forces around them and get them to absorb that in the instant and act on it, I have to think about the experience. I have to think about, how do I actually expose that information to them? How do you visualize? How do you communicate? These things become really important.
Iterative, agile, small wins to address the trust part. Visualization, articulation, communication to address the experience part. Then actually having a common set of goals. It's, of course, easy in racing because it's obvious. It's black and white.
In most environments, we'll end up in a scenario where we have supply chain planners and then we have machine learning technicians. We have to bring the same magic together. It always comes down to those three things. How do you experience? How do you get a commonly shared set of goals that are visible? Then how do you build trust through an iterative, agile, small win, big outcome kind of approach?
Michael Krigsman: Any final thoughts or advice to businesspeople who are listening and saying, "Yeah, I want to inject greater collaboration or have my organization evolve towards incorporating a greater degree of collaboration"?
Sanjay Srivastava: Well, look. I would say our learnings are, you have to get the culture right. The culture of continuous learning, of continuous innovation, of rethinking, of thinking outside the box, it's an important thing to get right. That's not an easy thing to change, so continuously working on that and driving cultural change is an important part. I think you need a backbone and an infrastructure, an ability to actually have this cohesive sort of connective fabric that allows us to sort of have the backbone upon which to innovate, that becomes really important.
Then I cannot overemphasize the word that we use a lot, "bilingual." You have to bring a convergence of disciplines. You have to bring people that can combine different things and come up with a converged innovation that comes from different disciplines.
If you focus on those three things, they're probably our biggest learnings and probably my number one advice.
Michael Krigsman: Sanjay Srivastava, thank you so much.
Sanjay Srivastava: Indeed, it's been my pleasure. Thanks for having me.
Published Date: Nov 12, 2019
Author: Michael Krigsman
Episode ID: 633