Part 2 of the two-part interview series with Genpact’s CEO and Chief Strategy Officer
The traditional offshoring-driven labor-intensive outsourcing model is finding its bottom (and fast) as enterprises aspire to go “straight to digital.” One service provider that’s kept pace with the changing industry narrative is Genpact. The firm built on its LEAN Six Sigma process excellence culture (emanating from its GE heritage) with organic and inorganic investments in emerging technologies as typified by the AI-powered Genpact CORA assets. Company leadership talks about using the “Instinctive Enterprise” to get closer to the customer, which is aligned with our Digital OneOffice approach.
We recently caught up with Genpact’s CEO Tiger Tyagarajan and Chief Strategy Officer Katie Stein in a wide-ranging interview talking about the market developments and Genpact’s past, present, and future. Tiger will also be a part of our leadership panel at the upcoming HFS Summit in NYC on October 1-2 in case you’ve still not registered for the event!
Tiger and Katie talked about their best-ever quarter in six years and their recent strategic acquisitions in Part 1 of the interview. In this edition, Part 2, we discuss the challenges with reskilling, data, mindset, and what sort of company Genpact will become in the next two years!
Phil: You talked about the increase of transformational elements within your engagements—you’re having to think a lot more about staffing-up with people who can service that. How challenging is that, and is there a specific profile you’re looking for? How are you building more consultative people who can think outside the box with the clients?
Katie: One factor is the broader reset we’ve just done, which is more forward-looking. We call it Project Orion, but it’s really looking at transformation services for us three and four years from now. We have a roadmap for which specific capabilities we want for the next four years. For example, in the supply chain, what specifically do we want to be able to deliver? And there’s obviously a people, scaling, and talent plan that’s being developed along with that.
In the meantime, we also have “Genome.” Genome is a mass reskilling program that we’re deploying at enterprise scale across Genpact. As the first step forward, we consider a group of “I”s. Think of an “I”—it’s deep and narrow, i.e., you have many subject matter experts. That’s really our history, people who knew a specific process. Back to your point about how customers want to buy and who we’ve hired, we have historically focused on “I”s who are deep in a specific technology or specific digital tools. But where I think we’ve realized there is a weakness, in our firm as well as in our clients’, is with the “T”s—the integrators who can stitch across and truly become the program managers, bringing various capabilities to stitch and deliver the solution.
We’ve been building up the program management office, and the Genome program is targeting all 90,000 people with a set of in-the-moment, in-the-skillset training. For my subject matter experts in F&A, it could be storytelling training. It could be the softer skills of consultative delivery. I won’t call it consulting yet, but consultative delivery. I’m taking machine learning, for example. I never need to code in Python; however, when I am facing a client, I need to understand how it works and what it would mean in their context.
Tiger: I can have great RPA people, but if they don’t know where to apply it and how to apply it, and if they don’t understand the industry and the process…
We’re also still big believers in the science of process, like Lean/Six Sigma, which our old parent company [GE] is bringing, which is great. I believe that’s a big journey that the new leadership team is going on, so we’re teaching that again. And then many people are learning AI, machine learning, and all the new technologies, and how to have enough empathetic conversations with CIOs and the buyers to explain, “AI can work here,” or “Don’t bring AI here, it’s a waste of money.” For a lot of our digital conversations, maybe half the time, we’re telling the customer, “There’s no point,” or, “It’s better to apply it there than to apply it here.”
The learnings on Genome are bite-sized, virtual, and in-the-moment, with quick feedback on a platform. The teacher is often someone inside the company who is designated as a guru on a topic, and the learner is any of our 90,000 people.
There’s no point learning something that you don’t want to practice. We want you to learn only things that you can practice tomorrow morning with a client, in a solution. It’s almost caught fire inside the company because everyone is hungry.
Phil: So it feels to me like you’re very much focused on AI, in process, as the key differentiator taking this forward. So where do you see that going? Where do you think clients can genuinely get value today based on the skillsets that they have, that they’re trying to get, and that you can deliver? And how do you see this shifting in about two years’ time?
Katie: In F&A, we’re stepping back and looking at AI. Many clients fall victim to constraining their imaginations and starting with, “Where do I have data?,” or constraining themselves to some data science group, thinking, “Where can we apply models?” Genpact hasn’t been doing that. We’ve been stepping back and saying, “Okay. I process invoices, what do I do, over and over again?” We might have thousands of people who match an invoice... so then we say, “What is the prediction here (that is the AI) that can enable human judgment or, where accuracy is high, supplant the need for a human?”
That is a classic example of where a little bit of RPA and OCR, plus a little bit of machine learning, can come in and actually fundamentally transform speed and accuracy.
It has immediate productivity equations for our clients, but, more importantly, a number of these examples allow us to say, “Ah-ha! We used to be able to say to you, ‘We can lower your costs by 20%, 30%,’ ‘We can eliminate work.’ Now what we’re saying to you is, ‘We can bring intelligence back into the system that allows you to go track leakage, that before, you couldn’t go after.’ Because $10 on every invoice wasn’t worth going after before. The whole cost equation wasn’t set up that way. But with AI dropping the level of cost, we can go after that leakage for you. Genpact now can guarantee the cost, and we can come back to you and guarantee to collect the leakage that was in your system.”
Tiger: Also, there are industry-specific processes; for example, in most insurance companies, both personal and commercial lines are challenged with having to decide when a deal comes in, “How much time do I spend on it? Where do I spend time first?” And that’s a combination of “What’s the risk? What’s the price I’m going to get for that risk? And what’s the probability that I do get a good price and that I’m going to win the deal?”
So that’s the ability to use past data and then quickly look at the data that’s coming in on the applications, and say, “This particular deal, even though it came in just this morning and there are hundreds of other deals waiting, I want the best underwriter to look at it now because it’s fabulous risk, it’s a great pricing I’m going to get, and if I give a good price, I’m going to win it now.”
Phil: How do you come to your clients with this proposition? There’s a lot of talk around platforms, and whether this is the right approach to take, or is it more about frameworks and engagement? You’ve got your own CORA platform or framework. You’ve got one, Accenture has one, KPMG has one, and DXC has one. Do you think that’s the future? Or do you think it’s more mindset and engagement models?
Katie: As we’re deploying machine learning or other techniques, I think there is inherent value for invoicing, trade promotions, or other areas, in the first 70% of assembling the engines and having them talk to each other. In the trade promotions example, first, I must ingest data. Second, I must extract pieces of information from the data. There is a basic level of process for the technology itself that has to happen. On top of that, my belief system is, for the client, specificity, “Oh, okay, you’re Retailer X. Well, your invoices may be different from someone else who’s in the pharmacy distribution chain,” so there’s necessary customization on top. I don’t think that a platform solves a problem ubiquitously.
Tiger: It’s not a customized piece from scratch every time. There’s no value in that. Neither is it a standard platform that everyone just buys into, like an AWS. When you do it for one customer, and you take it to the next customer, in the same industry, that has the same problem, the chances are 30% of that is going to be similar.
So, you start 30% faster. You take it to the fifth customer, it’s 70% the same. As long as it’s the same industry. And that goes back to the other question you asked, are clients willing to change? It requires getting people’s hearts and minds. Leaders are getting there; the middle of the organization is where it’s not easy. Where we get traction is by brewing competitive juices. If two customers in an industry move, five other customers look at it and think, “Oh my, what am I going to do?”
When Walmart moves, the whole industry moves. If Walmart moves, all retailers move, Amazon moves, all CPG companies move—everyone moves. When Bridgewater moves, the whole financial services industry stands up and says, “What is that? I want a look at that.” When iconic brands move, and we’re associated with that [as Genpact], we can take those frameworks to others.
Katie: If you were to scatter plot the number of areas where we’re targeting AI, at any level of scale, it’s still quite small. Because, a lot of times, we get in there, and it’s not an AI problem, it’s a workflow problem. AI is a set of math, it’s a set of algorithms, it’s a set of models. But it’s ultimately coming back to the fact that there’s a business problem, and understanding which tool in your toolkit to use to solve it… And I think that’s one of the hardest things I’ve observed with clients, is getting them off of, “I need blockchain, I need AI, I need…” and to go back to the problem first, and then let us solve the problem, using what we have available to us. For the industry, that’s still a big challenge, and we’re actually perpetuating it among ourselves.
Tiger: Because otherwise, it remains a proof of concept. “The proof of concept has worked. Okay, now what do we do? Um…I don’t know. I’ve got to bring it into my company and implement it into my workflow,” and sometimes that takes forever.
Katie: A hammer looks like it can be applied to all things, and then you get into the specificity of the client, the context, the extent of the process, etc. It’s this one-size-fits-all hammer mindset that is a real underestimated challenge.
Phil: Yes. Absolutely. I think you’ve really hit the nail on the head…
Katie: [Laughs].
Phil: We won’t let our analysts do any more generic, horizontal pieces on blockchain, or RPA, or AI. It has to be within context.
I think you’ve nailed it: start with the problem, and then work back. I feel like we’re having the same conversation we had 10 years ago where we were talking about SEP and smart enterprise processes.
A lot of the RPA technologies were around in the 80s…
Phil: One final question: having watched the journey for the last 13 years, you’ve become a lot of the things you said you would, and now you’re looking at this process-first approach with a lot of your clients. What do you think is going to happen, generally, in the next couple of years? What sort of company are you guys going to become?
Tiger: Let’s push the horizon out more—five, seven, 10 years—and let’s start with the world. So, if one believes that things like AI and machine learning are going to become more ubiquitous, the things that we use really smart people to do today, like prediction, can actually be done at the base level by algorithms.
The value, therefore, of, “Hey, I have a great algorithm, I have a great prediction mechanism, or I have PhDs who can write algorithms,” is going to become lower and lower. The real value is going to come from things surrounding it that actually use those predictions. And it starts with: Do you have the data? Who has the data? Can you have proxy data that then allows you to have real data over time?
Do you have the domain and the process understanding to build the context of whatever AI you have, and whatever problem you’re trying to solve? And then, can you have the guts to take decisions and to take action based on all the predictions the machine is telling you? And then, can you feed that loop back into the machine?
The core of the value proposition, which I think we will continue on the path of, is going deeper and deeper into industry verticals and services of our choice, driving outcomes for our clients via understanding the client, understanding the data, and bringing data to life across many clients to add value back to industry.
We should be one of the leading AI users and value creators for our clients in the world. We won’t be the big horizontal AI provider, but we should be one of the leading proponents of using AI to create value—we have all the ingredients.
I don’t think any of this is going to happen alone. We think, obviously, acquisitions is one part, but I don’t think restraining ourselves and constraining ourselves—thinking, “Either we do it ourselves, or we acquire it”—is the right path. We think ecosystems of partners is the way problems are going to get solved.
Katie: What we’ve learned off the back of our recent retail win with Walmart, looking at the end-to-end, not just at the end-to-end within Walmart, but the end-to-end of their private label provider, the end-to-end to the other CPG providers, and you think about something as simple as receivables, flowing back and forth, and how one can start to look at that problem, as a connected ecosystem of an entire industry … It’s early, and there are lots of challenges, but opportunities start to bubble up if we can be savvy enough to catch them, which makes it something where a single client really can’t do it on their own. It requires providers, like Genpact.
We thank Katie and Tiger for their candid conversation and look forward to continuing the dialog at our upcoming summit, where Tiger will be speaking at the leadership panel.
Click here to read part one of this two part series
