Excerpts from an interaction
You have spoken about the need for talent to be ‘bilingual’, especially in companies that serve as tech partners for businesses. Why do you see this as such a critical issue to focus on?
The world does not need so many machine learning experts. They need to know how many of these understand the domain and can apply the tech in it. Technology by itself is just an instrument. You have to be able to use the instrument for business needs. At Genpact, our employees are deep in tech and in domain. We are trying to take those strengths and blend it together into one composite. Each of our colleagues understand both sides of the whole. When you do that, you approach a problem set much more holistically.
The insights become different and richer with kind of holistic approach. And it has a direct impact on business outcomes. At the end, data can show you some insights, but you need to go to judgement from there and for that, you need someone who understands the domain. This is hence a big focus for us.
Even as you say that talent needs to be bilingual and that the domain and the technology both needs to be understood, do you not think that adding more jargons, augmented intelligence, for instance, when people are still trying to grasp artificial intelligence does not really help the case?
I must confess I am among those who have probably contributed to this problem of tech jargons. But too often artificial intelligence (AI) is misunderstood and misplaced. It is a misnomer that if you have AI, it is all automated and hence human is not required. The benchmark becomes that if you cannot replace it, it is not AI. In my mind, that is the farthest from the truth. The reason we speak about augmented intelligence is because it is not replacement, but addition. It is about how you combine technology to augment the intelligence of a human being.
AI is a prediction machine, but you need to apply judgement to that prediction. The combination of the two takes you to action — any business decision is done like that. Augmented intelligence brings this concept of ‘computer in the group, human in the loop’ much better than AI.
We are constantly reading comments on how technology has become a business imperative in Covid times. But the same statements have been made now for nearly a couple of years now. Were corporates not taking it seriously before February 2020?
You are right. All language on digital transformation that we were speaking even a couple of years ago, is still the same. But the pandemic has propelled the need to drive business change. The tech was there but the business impact has changed now. The job now is not just about the technology but making a prediction from that dataset, in an efficient and cost-effective manner. The tech partner now needs to bring added advantage to the business— businesses need it. The digital transformation that we have been speaking about as an industry, has new meaning and relevance in Covid times, and has become the only way to drive change. Digital transformation is coming to the rescue of manufacturing, supply chain, distribution — its role has significantly changed to even become life or death for some businesses.
The pandemic has also required people to revisit their businesses to become more relevant to current needs. Do you see change management truly assisting in adoption to new ways of work?
Covid has made everyone a data analyst and thinking through change management is important. Part of change management is getting the right mindset and focus on issues. And then taking those recommendations and implementing it in real life to drive change. During the pandemic, several businesses came to a grinding halt. Companies had to revisit their overall business models so that they can make the shift and move to a different world.
This will not happen by chance. Look at the amount of programming and information that the Indian government had to do to get the word of the pandemic out. The same needs to happen to enterprises and in deploying technology.
Despite technology becoming such a business imperative and the rise of the likes of AI, there is still a degree of uncertainty when an action is thrown out. Business managers some wariness in accepting it as is. What is your advice on how tech players should tackle this?
In fact, the number problem with AI today is adoption. People are still not sure how and whether it is going to really work. Most enterprises that are looking to deploy AI are also seeing ways to address this. They just do not want to know the analysis and advice or action, but also the reasoning behind it. You cannot take a black-box approach if you want better adoption of the AI. The solution is developing white-box algorithms. This involves bread crumbing so that every element can be broken down and is accessible with just a click. This ability to go to data source that drove the end point decision is similar to how it would be if this was done manually. At Genpact, we are trying to recreate human curiosity in the world of AI with this approach. This is a great way to drive adoption, making a company much more agile and resilient.