An interview: Phil Fersht, HfS Research, talks to Sanjay Srivastava, CDO at Genpact

  • Facebook
  • Twitter
  • Linkedin
  • Email

Genpact has made some major investments over the last year to propel itself to the forefront of progressive digital business process management solutions that encompass elements of Design Thinking (TandemSeven), Artificial Intelligence (Rage Frameworks), its successful foray into CFO consulting, and its impressive launch of its AI platform Genpact Cora.

It's almost history repeating itself from a decade ago, when the (then privately held) Genpact turned the traditional Business Process Outsourcing (BPO) model on its head with its disruptive virtual captive proposition that significantly challenged the pricing models and ability to integrate offshore capabilities into the old BPO model. Now, the firm is breaking the mold, yet again, by making real inroads into infusing AI into business processes and introducing these concepts to its huge global community of finance and supply chain leaders.

One of the key brains behind these developments is Sanjay Srivastava, aptly names the Chief Digital Officer for Genpact, who set aside some time to talk with HfS CEO and Chief Analyst, Phil Fersht, to elaborate more on how a unique business process management provider can raise the bar when it comes to Digital Operations:

Phil Fersht, CEO and Chief Analyst, HfS Research: Good morning, Sanjay. It’s great to spend some time with you again. I’ve known you for quite some time now, but not since your post Genpact days. You went from Network24, Corosoft, Aceva, Akritiv, and really made a career, building tech firms, before selling them to established global businesses. Clearly you are an entrepreneur. Is Genpact your latest start-up or is it something a little bit different for you?

Sanjay Srivastava, Chief Digital Officer, Genpact: That's a great observation, Phil, I really love tech start-ups. My first one was in edge networking - getting media and apps to the edge of the Internet in its early days. The second company was in data center automation - helping clouds scale just before it took off. Our third company was in predictive analytics, at a time when analytics were still mainly descriptive and diagnostic. My fourth start-up was the one that brought me to Genpact, a net native cloud based ERP bolt on for F&A in the early days of SaaS evolution. Across all four of those start-ups, three things stood out to me, which are common to my experience at Genpact.

Number one was the desire to be highly disruptive. At Genpact we embody the notion that innovation requires disrupting our own business. We even have a formal DYR - "destroy your revenue" program.  We embrace it holistically and I think our culture is very innovative and highly entrepreneurial.

Secondly, a start-up culture has a certain intellectual honesty - best ideas win no matter where they come from. At Genpact we have this - rich diverse teams and a non-hierarchical culture that draws upon and extracts the best from it. 

Finally, we have curiosity and velocity. Start-ups excel at this. Large companies get stuck in their past success. Learning, experimenting, agile-developing, and fast-failing are critical traits difficult to incubate in a large company – we have it here. I look at Genpact and for me it is a large start-up. Of course, I don’t want to minimize anything we do here by calling it a start-up, but the reality is we are an extremely innovative, fast transforming company that is super focused on bringing very disruptive solutions to the industry.

Phil:  You've recently launched Genpact Digital - why are you jumping on this digital bandwagon? 

Sanjay: We are a company that is laser focused on digital transformation and we have really pivoted our business since our early days. There is no surprise here, right?  We have a large base of enterprise clients whom we served over many years now, and for them Digital is now a board-level agenda.

In the end, true digital transformation success comes only from end-to-end process coverage. You can setup a checking account on your phone in five minutes, but spend weeks getting through KYC and other middle office processes before you get to use it.  It is the same for submitting insurance claims, or resolving an issue around procurement, or tracking something in your supply chain, or getting a status from a customer on invoice payment.  The experience is very digital in the front office but drops as the process makes its way through the entire value chain.

Clients struggle because transformational success depends on the end-to-end client journey. And that journey, which you talk about Phil at HFS, starts with the OneOffice concept. But to do that, industry acumen and process knowledge is the real key. We happen to be deep in the operations of many of those business processes. I think we see ourselves in a really interesting situation - first there is significant demand acceleration for digital transformation in our markets. Second comes success in digital transformation locked in deep process knowledge and domain understanding. Thirdly we happen to be operators that can bridge new advanced technologies with that deep operational understanding. It’s the perfect setup. So yes. Digital is really strategic for us and we’ve pivoted the entire company to digital transformation.

Phil: Can Genpact create a much cleaner vision of digital than we are seeing from many of these failing digital transformation initiatives from traditional service providers? It seems like a lot of these approaches are packaged up to sell to large companies with deep pockets. From our experience, few have been able to turn the oil tanker no matter how riveting the PowerPoint plan. What makes your vision here clearer and more effective than what we have seen in the past that hasn’t worked?

Sanjay: I think that our approach is actually very different. Because it isn't just about implementing and deploying technology. It is actually about the end-to-end client experience and the complete journey mapping that comes on the back of that. It's different because the reality is most digital transformations struggle not because of lack of technology but because of lack of design and governance. For instance, the success rate of Robotic Process Automation (RPA) is well below where it should be in the industry today. Way below. So why is that? It isn't for lack of technology options. There are many providers now, it's difficult to keep up. RPA is super simple code base and it isn't that difficult to build out.

The challenge of robotics is actually getting the design and governance right. Design - how do you think about the end-to-end processes? How do you take a large footprint of work and make proactive decisions on where to apply, how to apply, what to change and fix before you apply, and where not to? That design of the operating model and the change management around it is really key. Clients that started by just rushing and implementing some robotics are coming back and saying “Wait, how does this all come together? Where is the deep design?”

Governance is the other challenge. Many of our clients have done some 10 robots here and 15 there. Lots of excitement, some worked, many failed, but great learnings. Now we are saying forget 10 or 15, let's deploy at scale so how do you get 5,000 bots up and running? No one wants to bet on one small software vendor, so you probably settle on two different RPA vendors. Then there is invariably a couple of different versions, one in the US and one in Europe, this is just how these things go. With a variety of security patches here and there. On top of that, your processes are changing weekly, apps are updating monthly, data formats are constantly evolving. So now you've ended up with 5,000 robots on mission critical deployments with a thousand moving pieces, it becomes super critical to have the right governance. Because of our process centricity this notion of governance, command and control, is an area where we put a lot of focus on.

Phil: I really feel that we are in the calm after the storm within our industry now. In your experience, everybody is becoming an AI expert, right? For traditional businesses, the focus is shifting much more, I think, to renovation as much as it is innovation where there are many techniques like RPA, machine learning, cognitive to get more from existing infrastructure as opposed to the rip and replace of investing in brand new systems. Your new Cora offering appears to be an intent to pull together digital analytics and AI into one cohesive platform. Are customers truly ready for this? Are you looking to innovate or renovate as you roll this out?  What's the real plan?

Sanjay: I think it's a great question, Phil.  Look, our clients tell us that they are struggling on three fronts with digital transformation. First the noise level is unmanageable - every day there is yet another new technology, another AI science, coming to market. There are two examples of what your competitor did. Here are two examples of what the start-up down the road is doing. And then your cousin's son is starting off from college and building a company in this space. Every CXO is almost inundated with this kind of stuff coming through.  One of the first things we did with Cora is curated a comprehensive set of technologies that you can build a framework out of and drive transformation in a managed fashion.

Modularity is also important because clients will say, "Look Sanjay, we are going to implement 5,000 robots this year, but in the next 18 months, we want to apply machine learning to all this new data and turn them into cognitive agents for NBA (next best action) so we can respond to queries – and this all needs to be on one road map”.  That's the natural progression of digital transformation. As you digitize, the next set of opportunities open up and the ability to have a curated set of technologies which are available in a modular fashion is invaluable. The investment you are making today is protected for the future.

Second, is the ability to make digital work in the context of a large legacy enterprise IT Infrastructure. This goes to your point about renovation vs. innovation. Fortune 1000 companies all have an existing group of businesses, a large set of clients, and a legacy IT infrastructure in place. So doing something really cool and fascinating in some standalone online channel may be easy to do, but making it work in your existing business is much harder. With Cora we are building a mature API that is extensible and that allows for many of the ecosystem partners to participate. This allows us to ensure, as we implement a machine learning algorithm here and feature engineering in text classification there and a robotics implementation somewhere else, that in fact all of them can leverage each other. And more importantly, can leverage the existing IT foundation of our clients.

The third challenge we hear from clients is governance. Here is an example. Suppose in my office I have 100 people working on this floor, if 50 of them don't show up tomorrow morning, I'd know it within ten minutes. But if 100 robots can no longer log in, because someone changed the password policy, I may not even know for days. So, in this transformation from the human manual workforce to a digital automated workforce, how do you manage errant bots? Machine learning is great but who is making sure that that data it’s learning from is not biased. Chatbots are fantastic but how do you make sure they don’t pick up and use politically incorrect language. This notion of governance is really key. And Cora does that with a command and control center approach that allows you to manage, not in an IT log file level, but at a process level. You understand where your control points are, visualize the choke points in your process and can systematically address that. To go back to your point about the calm after the storm, I do think there is a lot coming, and getting organized with the right framework and curated technologies and command and control mechanisms, is the need of the hour.

Phil: I look forward to seeing how this develops with time, Sanjay. I've got one final question. What is truly the next big thing Sanjay? You know you've always seemed to find yourself working a demand curve to perfection. So where is this one going to take us by 2020?

Sanjay: Phil, our biggest belief is that artificial intelligence is going to redefine the meaning of the word 'work'. Much of what we call work is actually the manual exception management on the processes that can’t be straight through processed because the technologies available then couldn’t solve for the last mile. This will change with AI.  It won’t happen overnight.  But as a company we are convinced it will happen. AI, and more specifically deep reasoning and convolutional neural net capabilities will change that paradigm. The discussion will need to change from the future of work to the work of the future.  And that work of the future will be built on contextualization, on understanding the domain, on the process nuances. We think of the work of the future as the area under the curve of many thin vertical spikes on a thick horizontal bar.  The horizontal bar represents infrastructure, core AI algorithms - scale technologies that allow you to solve AI capacity and infrastructure - companies like Google Microsoft AWS and others play here. The vertical spikes built above represent the work that is done to apply AI in the context of a process and an industry, really a specific problem, and require deep knowledge of the domain, we play there.

We think the world of AI is going to look more and more like that. We will partner deeply with those companies that provide core machine intelligence technologies at scale and we will bring in the practical ability to apply AI in the context of a specific business problem, connect it to the rest of the enterprise and bring in the governance to evolve to this new digital workforce. We are very very passionate about AI.

Phil: This is great stuff Sanjay. Thank you for your time today.

This interview was first published at HfS Research.