In today's business world, information is power. But what does it mean to be a data-driven enterprise? According to Tiger Tyagarajan, CEO of Genpact, it means being the kind of business that uses real-time data sets and predictive analytics to make decisions that move the company ahead of its competition.

As Tiger explains, start by asking yourself what kind of outcome your business is trying to deliver. That could mean helping a client achieve a specific goal or gaining market share against your competitors.

The next factors to consider are how to get started quickly, using the most current data available, while minimizing costly risks. Create a digital transformation strategy that uses predictive modeling and machine learning algorithms to help you determine exactly what data you need, gather the data in real-time, and provide a combination of data science and machine learning to create an ideal data management platform and predictive analytic tools for your business.

But this kind of digital business transformation can meet with resistance among your employees. After all, the changes you want to implement may introduce changes to your company’s working culture and processes.

For this reason, be sure to have a clear vision of an overarching, data-driven business strategy and set winnable goals for the different teams in your company. Gaining early wins around digital strategy can help employees adapt to change and be open to workforce retraining.

Information is power, and data is key. But how does a company shift forward to becoming a data-driven organization?

First, ask yourself what outcome you want. Then figure out what data you need to make that happen.

Understand that changing the way your company does business means starting at the top, with leadership building a clear vision. Then, start with winnable goals that build momentum for your team.

Also understand that retraining and re-skilling workers may be necessary to help them adjust to big changes, such as AI making the decisions they used to be tasked to make. Building excitement in your workforce can help them forge ahead instead of resisting change.

To become a data-driven organization, create a broad vision but start small and gain quick wins. Tiger’s insightful comments are based on experience across many Genpact clients, so be sure to watch the video!

Transcript

This transcript has been edited for length and clarity.

Michael Krigsman: What does it mean to be a data-driven organization and how do we get there? We're live in Brooklyn, New York, at the site of the Formula E Race with Tiger Tyagarajan, who is the CEO of Genpact. Hey, Tiger. How are you?

Tiger Tyagarajan: Michael, great to see you. Thank you.

What is a data-driven organization?

Tiger Tyagarajan: What is the outcome you're trying to deliver? I think the Envision Virgin Racing outcome is, we want to win the race. You work backward from the outcome. Then you say, in order to deliver that outcome, what are the various decisions that are being taken every second?

Let me take that to the business world. If you are lending to a business and I'm a bank, then I want to make sure that I lend to the best customers at the best prices and I win continuously versus competitors and I delight the customer. Which means, I want to do it fast. I want to do it quick. I want to take the least risk. I want to reduce losses. I want to get the maximum price. You keep working backward. That's built on a lot of decisions that are being taken, which is a combination of prediction and judgment.

More and more, machines are getting very good at prediction, which means you have to have the data platform aligned. Where is it all going to come from? They all come in different formats. How do you bring that together? Some of it is real-time.

I'll give you one example in the commercial lending space. Typically, when people lend, they look at FICO scores, credit scores of all of us. Once you finish the lending and, a year later, you are still a customer and you repay whatever you borrowed, and you enter a 30-day default. You have not paid your bill.

You get a call from someone in the bank. They look at the profile of the customer and they say, "Oh, the person's credit rating is X." Guess what that data is. It's a year-old data, which is completely useless. In reality, businesses should be finding a way to get credit data every day about a person and then decide, in the call with me, what to do with me based on my credit in today's situation, not yesterday's situation.

Which business areas should leaders tackle first?

Tiger Tyagarajan: The real problem is identifying the big problems to solve, so the three or four situations that they end up in. One, a lot of them want to boil the ocean. I want to solve everything. That doesn't work.

The second is, I want to solve the easiest thing. It doesn't generate much value, in which case, what's the point? Companies lose attention. People lose traction.

The best opportunity is to really look at a set of real problems that are worth solving, identify the cost versus the benefit, and then pick very few—the lesser the better to begin with—because then you can drive change. You can bring the data together. You can get traction at the organization. When you declare victory, you actually have delivered enough impact for everyone to feel excited and get to the next journey.

Michael Krigsman: There's an organizational maturity dimension to this as well.

Tiger Tyagarajan: Absolutely. Yes.

Michael Krigsman: A company needs to gain experience with these types of problems.

Tiger Tyagarajan: Yes. Yes, and that experience also helps in one other thing, which becomes a big obstacle in all of these, which is, let's go back to the decision-making framework that I laid out. We humans love making decisions. I'm the expert, so I'm the guy who makes the decisions. You're going to take that away and take parts of it, not all of it, parts of it to a machine? No.

I'm going to resist that change until my last day at work. How you drive me to change my behavior to actually allow me to give up that machine-driven prediction in order for my job to actually become better because now I can use all my time and judgment.

Let's go back to the risks. The engineering team is feeding, "Hey, driver, you should do this. Sam, do this." Ultimately, the decision is taken by Sam. Sam decides what he wants to do on the racetrack. That never goes away from the driver.

The same here. That shouldn't go away from the human.

What advice do you have for overcoming resistance to change?

Tiger Tyagarajan: Yeah, I'll start by saying a belief that change has to come from the top. A believe that we have to use data to drive prediction in order to get better at serving our customers, serving our employees, and serving our shareholders, stakeholders, and investors. Once that gets driven down, then identifying the initial waves of, call them, projects and initiatives. All that is top-down.

The second thing is, how do you get the troops and excitement in the troops because everyone has to be engaged. Let them percolate ideas.

The third is, they all need to get engaged with retraining and reskilling. Understand the domain and context. Understand the data. They must be able to speak the same language, which is a continuous reskilling program.

What challenges do companies face when driving data-related transformation?

Tiger Tyagarajan: I'll start by saying that you have the traditional way businesses have run, and they run very successfully. That is the single biggest challenge that you start with because the moment you say, "We're going to use data to take decisions," you are going against what has been successful so far. That's one.

The second is, what's going to happen to me? I am the person who used to take these decisions. I am the expert. People come to me.

You're saying it's some data that's going to do it. What is that? If I have power, then I prevent that from happening. I challenge that. I prevent change from happening. That's the biggest thing or biggest challenge.

Michael Krigsman: Okay. When companies interfere with what I think is actually the natural lifecycle of technology adoption, which is, technology moves forward, the implications are, they will be disrupted.

Tiger Tyagarajan: Yes.

Michael Krigsman: What is the antidote to that? What should companies do to not fall into that trap?

Tiger Tyagarajan: It starts with leadership that builds a vision that creates the excitement as well as the—I want to call it—fear of being left behind but the excitement of going ahead, of leveraging data, building, using it for decisions and science and so on, and cascading that across a core group in the company so that people start.

The second thing is making it very positive. If the discussion is, it's all negative, that it's all going to be about your job, my job, then it becomes a negative. It has to be a positive saying, "I'm going to capture more market. I'm going to become bigger. You're going to become bigger. There are going to be more of you," then that creates a positive momentum.

Michael Krigsman: You need to establish some set of positive goals but, at the same time, identify a core group of people that you can take on board with you.

Tiger Tyagarajan: Right. Right, as any change initiative in large enterprises, boiling the ocean doesn't get you anywhere. Maybe set the vision at the broad level, create the excitement at the broad level, and then pick the one or two places where you start driving that. The way you pick those are maximum impact. Leaders are brought in, leaders of our champions, or leaders who are influencers of other people.

Oh, John got into this and John is driving this? I must do something because I know John is kind of the favorite in the company. He's a rock star in the company.

Michael Krigsman: What final advice do you have on making all this happen?

Tiger Tyagarajan: I'll start by saying that it has to ultimately start with, what are the outcomes they're trying to deliver? I guess the outcome has to be around an outcome to the shareholders, an outcome to clients, an outcome to employees, an outcome to society, and making sure that you start from there rather than starting from, where can we use this technology? The hammer searching for the nail is a bad way to go about it. Start from the outcome. Then figure out the best place to really create a big impact that actually other people can look at and say, "I want to follow that. I want to be as big," et cetera.

How should business leaders choose the right data science business projects?

Tiger Tyagarajan: Michael, we are learning as we have done this with a large number of customers. There is a new learning that we are beginning to develop. That learning is going after the big game-changing move. While it's ultimately the most interesting and right thing to do, it may not be the fight first thing to do.

If you want quick wins, and you want people to say, "Wow, this is pretty cool and not that difficult," then it may be better to grab smaller things that drive. Small improvements that actually don't take away my job and start that. Then, layer on top of that.

The good news is that all of these technologies, the second time is a shot to drive improvements. You can watch the improvement deliver results pretty quickly. The cost involved is not that much.

By the way, you'll have failures. If it fails, pull the plug and move on to the next one. You have a portfolio of these. Start with the easy ones, small ones. Create wins and go to the big ones.

Michael Krigsman: What's the purpose of starting with small ones?

Tiger Tyagarajan: Quick wins. Show success fast. People lose patience and fatigue. The faster you show success, people say, "Yeah, this works." The bets are smaller, to begin with. You're not trying to boil the ocean.

There's another big one. AI, machine learning, et cetera are still new use cases and new sciences. There is a lot of experimentation. When you experiment, by definition, there's going to be failures. Big visions that then lead to big failures, unfortunately, will lead to, "I'm not going to try this anymore," which is tragic.

Starting small, having some wins that then ensures future success is, I think, a better way to go.

Michael Krigsman: Eventually, we all are reading about the hype that AI is going to change our company.

Tiger Tyagarajan: Yes.

How much of AI is hype?

Tiger Tyagarajan: It's not hype at all. It is not hype at all. We've got enough examples in the pharma industry with some of the patient safety areas, some of the areas around safety in engines, in preventative maintenance, in lending and the speed of lending and so on. Enough examples of complete game-changing moves.

All I would say, Michael, is to create the vision for the game-changing move and then say, okay, we all bought into the vision but we will start here and prove the case. Let's put all our energies on this, which is smaller, more contained, and can deliver results faster. Then layer on top of that.

Layering and what I call version one, version two, version three, version four. Ten versions in a year is not a bad idea.

Michael Krigsman: You need to have the clarity of having the vision of where we're going and what's possible.

Tiger Tyagarajan: Yes. The north star is important. Why are we doing this, in the long run, is important. The size of the prize being so big is important. The fact that this is game-changing is important.

All I'm saying is, having agreed on that, I think jumping wholeheartedly into the big bang may not be the right one. CDs or small banks may be better.

What lessons for the enterprise have emerged from your partnership with Envision Virgin Racing?

Tiger Tyagarajan: I would say three things that I would point out. Number one, getting talent that is diverse in the way they think, in the skills they bring, in the language they initially use, which is very different. In our instance, often it'll be domain, process, and data versus algorithms, software, and so on. Then, finding a way to make those teams work well together, which means it's leadership. That's one.

The second would be, how do you create an ecosystem which is not just the team itself but the team connects to other teams outside—some inside the company, some outside the company—and driving that ecosystem?

The third is, how do you get the vision to be big enough to change the game and really go after changing the game? But do it by first saying, I'm going to play faster, better at the current game and show that it works.

Michael Krigsman: You're operating on both these levels. On the one hand, you need this very broad, expansive vision. But at the same time, you need a game plan that is incremental here, incremental there, small changes.

Tiger Tyagarajan: Yeah. Yeah, and high-frequency change.

Michael Krigsman: The high-frequency change is a key.

Tiger Tyagarajan: Yeah. These are not large ERP, SAP projects that takes two years and, after two years, you realize, "Oh, how do we make this work." These are, okay, 90-day sprint, the next 60-day sprint, another 90-day sprint….

How should companies think about culture change and experimentation?

Tiger Tyagarajan: Oh, no, Michael. You've just nailed it with that question itself and the word "culture." In the end, if you were to force me to says, what's the one thing that is the ultimate secret sauce, it's culture. How do you get the culture to embrace this? How do you get the culture to actually become multilingual, bilingual? How do you get the teams to work together? How do you get the AI expert to be listened to by the risk expert rather than the risk expert saying, "But you don't know underwriting. I've done it for 20 years. Don't tell me how to do this," and vice versa?

How do you get the AI expert to say, "Move out. My algorithm is going to work." No, your algorithm is not going to work in my business. Understand my business.

How do you create that culture? How do you create the culture of, failure is okay; let's learn and move on? How do you create the culture of, the technology is not the big thing here? It's actually, what outcome are we trying to deliver? Sometimes, cool technology guys are searching for the next place to use their next new toy, so it's all about culture so, therefore, it all starts from the top.

This transcript has been edited for length and clarity.

Michael Krigsman: What does it mean to be a data-driven organization and how do we get there? We're live in Brooklyn, New York, at the site of the Formula E Race with Tiger Tyagarajan, who is the CEO of Genpact. Hey, Tiger. How are you?

Tiger Tyagarajan: Michael, great to see you. Thank you.

What is a data-driven organization?

Tiger Tyagarajan: What is the outcome you're trying to deliver? I think the Envision Virgin Racing outcome is, we want to win the race. You work backward from the outcome. Then you say, in order to deliver that outcome, what are the various decisions that are being taken every second?

Let me take that to the business world. If you are lending to a business and I'm a bank, then I want to make sure that I lend to the best customers at the best prices and I win continuously versus competitors and I delight the customer. Which means, I want to do it fast. I want to do it quick. I want to take the least risk. I want to reduce losses. I want to get the maximum price. You keep working backward. That's built on a lot of decisions that are being taken, which is a combination of prediction and judgment.

More and more, machines are getting very good at prediction, which means you have to have the data platform aligned. Where is it all going to come from? They all come in different formats. How do you bring that together? Some of it is real-time.

I'll give you one example in the commercial lending space. Typically, when people lend, they look at FICO scores, credit scores of all of us. Once you finish the lending and, a year later, you are still a customer and you repay whatever you borrowed, and you enter a 30-day default. You have not paid your bill.

You get a call from someone in the bank. They look at the profile of the customer and they say, "Oh, the person's credit rating is X." Guess what that data is. It's a year-old data, which is completely useless. In reality, businesses should be finding a way to get credit data every day about a person and then decide, in the call with me, what to do with me based on my credit in today's situation, not yesterday's situation.

Which business areas should leaders tackle first?

Tiger Tyagarajan: The real problem is identifying the big problems to solve, so the three or four situations that they end up in. One, a lot of them want to boil the ocean. I want to solve everything. That doesn't work.

The second is, I want to solve the easiest thing. It doesn't generate much value, in which case, what's the point? Companies lose attention. People lose traction.

The best opportunity is to really look at a set of real problems that are worth solving, identify the cost versus the benefit, and then pick very few—the lesser the better to begin with—because then you can drive change. You can bring the data together. You can get traction at the organization. When you declare victory, you actually have delivered enough impact for everyone to feel excited and get to the next journey.

Michael Krigsman: There's an organizational maturity dimension to this as well.

Tiger Tyagarajan: Absolutely. Yes.

Michael Krigsman: A company needs to gain experience with these types of problems.

Tiger Tyagarajan: Yes. Yes, and that experience also helps in one other thing, which becomes a big obstacle in all of these, which is, let's go back to the decision-making framework that I laid out. We humans love making decisions. I'm the expert, so I'm the guy who makes the decisions. You're going to take that away and take parts of it, not all of it, parts of it to a machine? No.

I'm going to resist that change until my last day at work. How you drive me to change my behavior to actually allow me to give up that machine-driven prediction in order for my job to actually become better because now I can use all my time and judgment.

Let's go back to the risks. The engineering team is feeding, "Hey, driver, you should do this. Sam, do this." Ultimately, the decision is taken by Sam. Sam decides what he wants to do on the racetrack. That never goes away from the driver.

The same here. That shouldn't go away from the human.

What advice do you have for overcoming resistance to change?

Tiger Tyagarajan: Yeah, I'll start by saying a belief that change has to come from the top. A believe that we have to use data to drive prediction in order to get better at serving our customers, serving our employees, and serving our shareholders, stakeholders, and investors. Once that gets driven down, then identifying the initial waves of, call them, projects and initiatives. All that is top-down.

The second thing is, how do you get the troops and excitement in the troops because everyone has to be engaged. Let them percolate ideas.

The third is, they all need to get engaged with retraining and reskilling. Understand the domain and context. Understand the data. They must be able to speak the same language, which is a continuous reskilling program.

What challenges do companies face when driving data-related transformation?

Tiger Tyagarajan: I'll start by saying that you have the traditional way businesses have run, and they run very successfully. That is the single biggest challenge that you start with because the moment you say, "We're going to use data to take decisions," you are going against what has been successful so far. That's one.

The second is, what's going to happen to me? I am the person who used to take these decisions. I am the expert. People come to me.

You're saying it's some data that's going to do it. What is that? If I have power, then I prevent that from happening. I challenge that. I prevent change from happening. That's the biggest thing or biggest challenge.

Michael Krigsman: Okay. When companies interfere with what I think is actually the natural lifecycle of technology adoption, which is, technology moves forward, the implications are, they will be disrupted.

Tiger Tyagarajan: Yes.

Michael Krigsman: What is the antidote to that? What should companies do to not fall into that trap?

Tiger Tyagarajan: It starts with leadership that builds a vision that creates the excitement as well as the—I want to call it—fear of being left behind but the excitement of going ahead, of leveraging data, building, using it for decisions and science and so on, and cascading that across a core group in the company so that people start.

The second thing is making it very positive. If the discussion is, it's all negative, that it's all going to be about your job, my job, then it becomes a negative. It has to be a positive saying, "I'm going to capture more market. I'm going to become bigger. You're going to become bigger. There are going to be more of you," then that creates a positive momentum.

Michael Krigsman: You need to establish some set of positive goals but, at the same time, identify a core group of people that you can take on board with you.

Tiger Tyagarajan: Right. Right, as any change initiative in large enterprises, boiling the ocean doesn't get you anywhere. Maybe set the vision at the broad level, create the excitement at the broad level, and then pick the one or two places where you start driving that. The way you pick those are maximum impact. Leaders are brought in, leaders of our champions, or leaders who are influencers of other people.

Oh, John got into this and John is driving this? I must do something because I know John is kind of the favorite in the company. He's a rock star in the company.

Michael Krigsman: What final advice do you have on making all this happen?

Tiger Tyagarajan: I'll start by saying that it has to ultimately start with, what are the outcomes they're trying to deliver? I guess the outcome has to be around an outcome to the shareholders, an outcome to clients, an outcome to employees, an outcome to society, and making sure that you start from there rather than starting from, where can we use this technology? The hammer searching for the nail is a bad way to go about it. Start from the outcome. Then figure out the best place to really create a big impact that actually other people can look at and say, "I want to follow that. I want to be as big," et cetera.

How should business leaders choose the right data science business projects?

Tiger Tyagarajan: Michael, we are learning as we have done this with a large number of customers. There is a new learning that we are beginning to develop. That learning is going after the big game-changing move. While it's ultimately the most interesting and right thing to do, it may not be the fight first thing to do.

If you want quick wins, and you want people to say, "Wow, this is pretty cool and not that difficult," then it may be better to grab smaller things that drive. Small improvements that actually don't take away my job and start that. Then, layer on top of that.

The good news is that all of these technologies, the second time is a shot to drive improvements. You can watch the improvement deliver results pretty quickly. The cost involved is not that much.

By the way, you'll have failures. If it fails, pull the plug and move on to the next one. You have a portfolio of these. Start with the easy ones, small ones. Create wins and go to the big ones.

Michael Krigsman: What's the purpose of starting with small ones?

Tiger Tyagarajan: Quick wins. Show success fast. People lose patience and fatigue. The faster you show success, people say, "Yeah, this works." The bets are smaller, to begin with. You're not trying to boil the ocean.

There's another big one. AI, machine learning, et cetera are still new use cases and new sciences. There is a lot of experimentation. When you experiment, by definition, there's going to be failures. Big visions that then lead to big failures, unfortunately, will lead to, "I'm not going to try this anymore," which is tragic.

Starting small, having some wins that then ensures future success is, I think, a better way to go.

Michael Krigsman: Eventually, we all are reading about the hype that AI is going to change our company.

Tiger Tyagarajan: Yes.

How much of AI is hype?

Tiger Tyagarajan: It's not hype at all. It is not hype at all. We've got enough examples in the pharma industry with some of the patient safety areas, some of the areas around safety in engines, in preventative maintenance, in lending and the speed of lending and so on. Enough examples of complete game-changing moves.

All I would say, Michael, is to create the vision for the game-changing move and then say, okay, we all bought into the vision but we will start here and prove the case. Let's put all our energies on this, which is smaller, more contained, and can deliver results faster. Then layer on top of that.

Layering and what I call version one, version two, version three, version four. Ten versions in a year is not a bad idea.

Michael Krigsman: You need to have the clarity of having the vision of where we're going and what's possible.

Tiger Tyagarajan: Yes. The north star is important. Why are we doing this, in the long run, is important. The size of the prize being so big is important. The fact that this is game-changing is important.

All I'm saying is, having agreed on that, I think jumping wholeheartedly into the big bang may not be the right one. CDs or small banks may be better.

What lessons for the enterprise have emerged from your partnership with Envision Virgin Racing?

Tiger Tyagarajan: I would say three things that I would point out. Number one, getting talent that is diverse in the way they think, in the skills they bring, in the language they initially use, which is very different. In our instance, often it'll be domain, process, and data versus algorithms, software, and so on. Then, finding a way to make those teams work well together, which means it's leadership. That's one.

The second would be, how do you create an ecosystem which is not just the team itself but the team connects to other teams outside—some inside the company, some outside the company—and driving that ecosystem?

The third is, how do you get the vision to be big enough to change the game and really go after changing the game? But do it by first saying, I'm going to play faster, better at the current game and show that it works.

Michael Krigsman: You're operating on both these levels. On the one hand, you need this very broad, expansive vision. But at the same time, you need a game plan that is incremental here, incremental there, small changes.

Tiger Tyagarajan: Yeah. Yeah, and high-frequency change.

Michael Krigsman: The high-frequency change is a key.

Tiger Tyagarajan: Yeah. These are not large ERP, SAP projects that takes two years and, after two years, you realize, "Oh, how do we make this work." These are, okay, 90-day sprint, the next 60-day sprint, another 90-day sprint….

How should companies think about culture change and experimentation?

Tiger Tyagarajan: Oh, no, Michael. You've just nailed it with that question itself and the word "culture." In the end, if you were to force me to says, what's the one thing that is the ultimate secret sauce, it's culture. How do you get the culture to embrace this? How do you get the culture to actually become multilingual, bilingual? How do you get the teams to work together? How do you get the AI expert to be listened to by the risk expert rather than the risk expert saying, "But you don't know underwriting. I've done it for 20 years. Don't tell me how to do this," and vice versa?

How do you get the AI expert to say, "Move out. My algorithm is going to work." No, your algorithm is not going to work in my business. Understand my business.

How do you create that culture? How do you create the culture of, failure is okay; let's learn and move on? How do you create the culture of, the technology is not the big thing here? It's actually, what outcome are we trying to deliver? Sometimes, cool technology guys are searching for the next place to use their next new toy, so it's all about culture so, therefore, it all starts from the top.