Remarkable fact: 7,500 Wipro employees are actually robots. Working in tandem with Wipro's AI platform, HOLMES, the company is in the vanguard of deploying artificial intelligence and automation in the workplace.
Going forward, most successful workplaces will combine humans and machine learning systems, which will create radically new models – and require new mindsets – to stay competitive with other businesses.
In what ways will AI and automation shape the workplace of the future?
Even more important: how can you business stay ahead of the curve?
Does AI reduce or increase employment? Do human staffers have good reason to be worried about the rise of robots?(3:11)
"This is a question that we ponder about, and we get asked a lot. Do you really think it's a zero-sum game, or you think it is not a zero-sum game? Look at the history, look at what happened when automation was brought in, in the Industrial Revolution. Population has grown, number of jobs have grown, unemployment has come down. The US today has one of the lowest unemployment rates.
"I tend to be optimistic. I believe that there will be more jobs. But the nature of jobs will change. And there will be jobs in newer areas that probably are small today or growing, and there will be significant learning and unlearning that a large mass of population, especially those in operations and those maybe in more sort of manual roles, will have to.
"Absolutely, and not just in cognitive intelligence, obviously that's a hot area, but you're already seeing other surround areas which five years back did not have enough opportunity. For example, design. Because when you automate, you need to still consider human experience. So there are jobs in design, there are jobs in cybersecurity, there are jobs in compliance that five years back were small or insignificant. So there is a growing space all because of automation and cognitive intelligence, which is further spinning out a host of other opportunities in the space."
How is Wipro using AI and automation? How is it that Wipro has 7,500 employees that are actually robots? (8:22)
"So we use it in two ways, one is what we do for our clients, second is what we do to ourselves. For our clients, we have leveraged AI in these three different buckets. Number one is to provide significant automation of operations. Second is to improve customer experience, or client experience. Third is to identify new areas of revenue for our clients. I'll give you an example of each.
"In automation, let's talk about KYC. When banks onboard new clients, especially large corporates, they need to make sure that these are entities who have no connects with other parties that government should not be funding, under FATCA and other compliance areas. This requires a lot of work to go through annual reports of these entities, collect all the material that is available on the web, and create what you call "entity structures". And then obviously compliance officers look at it and validate whether yes, this is good to go or not.
"Most of that work is now automated through our bots, right? So that's one area in which we have done, under the operation space, but there are many other examples.
"Second, in terms of customer experience, I think all of us are consumers. I think this we ourselves see it. For example, a cognitive bot, a chatbot where you can actually talk to a bot. There are advanced version of it, and there are obviously not so advanced versions of it where sometimes it can get very painful.
"It is [also in the] call center, but now that you've taken it right to the edge where I'm interacting with the application, I'm getting my questions or queries answered right there at the application, the call actually doesn't come to the call center, so we call it "digital deflection" which is basically, reduce the cycle time, reduce the pain, and answer my query right there and then in a very cognitive and intuitive way.
"Third, I talked about is revenue generation. So we do a lot of analytics in customer supply chain space to reduce stockouts, for example. Or an example that Amazon was the first one to really bring to fore is next best offer. Right? I'm buying something, what is the next thing I should buy, and you should propose. So a lot of cognitive intelligence to come up with those answers. Again, revenue generation. Buying recommendations or next best offer, what should I be offered next, which have the highest propensity for me to spend my money on.
"How is Wipro using it internally? What we call "Wipro on Wipro." A lot of the work that we do for our clients where we deliver fixed price work for them, we have automated. We report this publicly. So 18% of the work that we do today is delivered by bots. And I think that's the number that you sort of picked up. So it's 18% of the work we delivered is delivered by bots. But internal shared services of Wipro, a lot of the work we do in our finance, a lot of the work we do in our HR shared service, is all automated. To give you an example, Wipro is a publicly-traded company. Every quarter we declare our next quarter's guidance. That guidance today comes through a bot or several bots that we have developed to come up with that guidance.
We started doing that a year back. It used to take 60 people from our finance team to burn midnight oil and then apply their judgment to come up with this. Now that team is eight people.
"And a lot of the work is done by the bot. Of course, a human being still decides whether this is the right guidance to give or not, but it comes through the system, and it has proven that that's more accurate than what humanly we were able to come up with ourselves.
"I'll give you another example which is even more complicated. This, we delivered for a client of ours. This is a public listed company. They have products. They are consumer products, clothing. They have number of products and number of channels to sell that product. And today, they leverage our bot, which we have jointly developed with them to come up with a prediction model. And again, the prediction model is better than what they used to do internally.
"And now you look at this, weather can change, and it can impact the sales of their products. Their sales, competitor sales price, all of that can impact the actual revenue that the company now gets. So this is very, very complex. Yes, so bots have come a long way."
Wipro is using "new methods, new models, and new mindsets to navigate the fourth industrial revolution." What does this mean? (3:43)
“This is a good question because a lot of people think, and some of our clients also think that for them to leverage cognitive intelligence, it is about technology. I don't believe it is at all only about technology. It is much beyond technology. So we have come up with something that we propagate. We call it a "4M model."
“So it is Method, Model, Machine, and Mindset. And all the clients and enterprises need to work on all four elements. I'll give you an example. So when we say "method", method is about them creating an entire value stream. It is not good enough for them to automate small parts of the value stream and become automated only in the small parts. They need to really become automated in the end-to-end value stream, for them and the clients to get significant benefit out of it. Right? "Model" is obviously about data model, who owns the data, where does data come from?
“Then obviously machine. 'Machine' is of a lot of technology and the data models behind that. And mindset, I think is very important, because for clients to leverage data, they have to change how they have traditionally made decisions, right? They have to trust the data. They have to also be ready to fail. Because initially when you start using these algorithms, it is quite possible that initially your results will be worse off than what you were used to. These machines learn very quickly, and then you need to be willing to spend the time to make them learn, and then eventually they'll be far better than what a pure human could.
“Sometimes you may leverage machine learning in a place where probably it's not fit, but you may take certain learning out of that experiment and then apply it to a different place. So you need to have that fail fast mindset even... Even here. So it is a lot about mindset. It's a shift of how clients have run their enterprises. So that's why we don't believe it's just about technology. They need to work on all the forums.
“Because of the advancement that is happening, it is quite possible you may fail on something today but six months later, that may become a great place for you to invest in because now you have new sources of data that have come in which were not available to you. So you need to really learn from every failure and be adaptive, to then apply this learning to new space."
What advice would you give to businesses that want to get on-board with AI and automation? Many companies are struggling with AI adoption. (2:35)
“That's a great point. A lot of enterprises are using AI, but maybe 4% are using it effectively. I think the difference really is these four or five things. Number one, I truly believe that they have to have a vision for AI. It can't just be a small experiment that they're doing on the side without having a north star or a vision. It is okay to start small as long as you still have that long-term vision of where AI needs to go, how do you become data-led enterprise or agile enterprise.
“Second, because it requires change management, it truly has to be top-down. We have seen, and in many cases the finance example that I gave you, the CEO of the company actually came down to where we were working. He actually not only learnt the work we were doing, in his enthusiasm, the CEO actually learnt how to do coding.
“Because he wanted to set a personal example of how this is, one, easy to do, and second, personal involvement in the success of this. So it has to be top-down. Third, because it is not just about technology, it is about change, there has to be a lot of investment that clients need to do in training, communication, change management. And the training has to be not only for the staff that is going to use it. Obviously, that's important because they need to learn how to use it, how to now unlearn what they were doing, but also for the staff that is going to get impacted because of this automation. Going back to your earlier question about how does it impact.
“So there has to be something done for those people as well. So I think between those who are successfully doing it and those who are not doing it so successfully, these three are sort of the core differences."
What does the future hold for the next 2-3 years, for AI and automation? (2:08)
“This is a very exciting field. So, I would maybe take two examples. One, I truly believe there would be end-to-end cognitive scale applications in enterprise. Today enterprises are automating parts of the value chain, as I said. But in the very near future, you would have end-to-end cognitive application. A good example, just because you talked about customer experience and call center earlier. There would be an automated customer call center application which will be self-sensing, self-responding, self-learning and self-adapting. That will happen in very near future and there are no issues with that.
“Second, which is very exciting space, I believe because of the culmination or growth in 5G, IoT, cloud, and hopefully very soon quantum computing, I think that will lead to tremendous growth opportunities. Very exciting space...
“Yes, and very exciting space because all of us have seen Minority Report, the movie, would be prevention of crime. I mean, once you have this... That is very much a reality. Obviously, there needs to be some work done by the authorities, but technically it's all possible in the very, very near future."