Copyright 2019, Forbesindia.com
As software bots get more complex, only a multi-talented workforce will thrive: Wipro COO
By Harichandan Arakali| Mar 4, 2019
Wipro's President and COO Bhanumurthy B.M. talks about how software bots are transforming the services industry, Wipro's third-gen delivery of services, and the right talent to survive the next digital era
Bhanumurthy B.M., President and COO, WiproIndian IT companies see automation and software robots at the heart of the next generation of IT delivery. Their clients expect no less. From finance to manufacturing to retail, software robots are touching every vertical and human jobs are either getting eliminated or transformed, and new ones being created in their place.
At Wipro, President and COO Bhanumurthy B.M. expects that software robots will independently take on increasingly complex processes in the future. In such a scenario, employees who can learn new skills on the go and adapt to a world where they will be routinely expected to work alongside bots will thrive, he said, in a recent interview with Forbes India. Edited excerpts:
Q How is the world of digital services changing?
We have started to look at our next-gen delivery. This is Wipro’s ‘generation three’ delivery (of services). In general, the first two generations of capabilities that India’s largest software services providers acquired were about demonstrating to their customers their ability for delivering high standards of quality from centres in India.
The third generation that is now being demonstrated to customers is about showing a deeper understanding of business domains. Therefore, Indian IT companies are using automation and within that, software robots-based automation, to rewire business processes. And increasingly on the cloud.
Q How will Wipro deliver the next generation of services?
Wipro relies on three pillars to deliver these next generation services. The first big pillar is about making application engineering a lot more digital. The work methods are different, the tool sets are different. The company has built its own (software) rigs to make software development much more agile, with use of techniques such as CICD (continuous integration, continuous delivery).
The second pillar is about an automation-first approach. This is about Wipro’s cognitive automation platform called Wipro Holmes. This is about how to make human staffers work along with bots to change the customer experience.
Third, how does the company ensure that any given project gets the best and the most relevant people working on it, even if some of them aren’t formally on the team that’s helming the project, but are available somewhere in the company. This may also involve some crowdsourcing, which Wipro internally calls ‘Top Gear’.
Q How has automation changed?
Automation has evolved from eliminating manual work on particular tasks — like say modifying an employee’s password — to a bunch of tasks that form a particular process, like ensuring that the right staffer has the right level of access to various software systems, related data and so on. Previously this all used to be fairly rule-based and involved “hand-coding”.
Now, there are sophisticated building blocks and general platforms available off-the-shelf from companies such as Automation Anywhere. And companies like Wipro have enthusiastically adopted them and started to build their own tools.
Automation has evolved from eliminating manual work on particular tasks — like say modifying an employee’s password — to a bunch of tasks that form a particular process, like ensuring that the right staffer has the right level of access to various software systems, related data and so on. Previously this all used to be fairly rule-based and involved “hand-coding”.
Now, there are sophisticated building blocks and general platforms available off-the-shelf from companies such as Automation Anywhere. And companies like Wipro have enthusiastically adopted them and started to build their own tools.
Another aspect of the evolution of automation is about getting the software to mimic human understanding. For example, understand the language in a contract, and perform various tasks or execute various processes based on that understanding. At that level the software bot or bots becomes ‘cognitive,’ and is able to learn, for example, the intent behind the contract.
Consequently, over the course of the maturity of the cognitive bot, it will become increasingly accurate and behave like a human being. Currently, Wipro does a lot of work in both, rule-based scenarios as well as on the cognitive platform Wipro Holmes. About 250 customers are already using Wipro Holmes.
Q Give us examples of how the bots are handling work of increasing complexity.
Now the question is how intelligent the robots can become and how complex a task can they handle. For instance, some of the bots that Wipro has deployed can already read through a set of financial documents and establish relationships between various entities in the documents and cull out information on the risks various parties would face.
Other bots the company has built for customers work on scheduling supply-chain routes and yet others can forecast the failure of industrial equipment and engines, and so on, based on the conditions in which the engines are used.
Q What are your customers telling you about automation?
Automation is on the agenda of every major corporation and every solution offered by the Indian IT companies must have an automation component in it. Wipro’s own analyses show that nearly a third of its customers explicitly see investments in software robots as part of their budgets for the next 12 months. Close to 80 percent of the customers broadly see automation as ‘strategic’ to their businesses.
Wherever there is a large amount of data and several variables influencing a business, one will see an interest in automation and software bots. Banking is clearly an area where there is much interest in new technologies. Traditionally too, financial clients have tended to lead in adopting new technologies. Retail, insurance, health care — where the problems are fairly complex — are among the verticals where there is much interest.
In these verticals, there are multitudes of end users and many variables that need to be correlated in attempts to find patterns and build templates. Failure forecasting in manufacturing is another area.
Consequently, over the course of the maturity of the cognitive bot, it will become increasingly accurate and behave like a human being. Currently, Wipro does a lot of work in both, rule-based scenarios as well as on the cognitive platform Wipro Holmes. About 250 customers are already using Wipro Holmes.
Q Give us examples of how the bots are handling work of increasing complexity.
Now the question is how intelligent the robots can become and how complex a task can they handle. For instance, some of the bots that Wipro has deployed can already read through a set of financial documents and establish relationships between various entities in the documents and cull out information on the risks various parties would face.
Other bots the company has built for customers work on scheduling supply-chain routes and yet others can forecast the failure of industrial equipment and engines, and so on, based on the conditions in which the engines are used.
Q What are your customers telling you about automation?
Automation is on the agenda of every major corporation and every solution offered by the Indian IT companies must have an automation component in it. Wipro’s own analyses show that nearly a third of its customers explicitly see investments in software robots as part of their budgets for the next 12 months. Close to 80 percent of the customers broadly see automation as ‘strategic’ to their businesses.
Wherever there is a large amount of data and several variables influencing a business, one will see an interest in automation and software bots. Banking is clearly an area where there is much interest in new technologies. Traditionally too, financial clients have tended to lead in adopting new technologies. Retail, insurance, health care — where the problems are fairly complex — are among the verticals where there is much interest.
In these verticals, there are multitudes of end users and many variables that need to be correlated in attempts to find patterns and build templates. Failure forecasting in manufacturing is another area.
Q Is automation killing jobs, but also creating new ones?
There are two ways of looking at automation. One is about building software to take over dull, repetitive tasks and release humans to do more creative work. The second is about tasks that are already of a higher order and of a nature that makes it difficult or even impossible for humans to attempt them.
For instance, a telecom company will have large volumes of data on various plans it offers and might want to know which customers are using what kind of data plans — something that one can throw software and raw computing power at, which on the other hand would be virtually impossible for humans to attempt. Humans can get involved here once the software has thrown up various patterns, and then use their knowledge of the business context to act on those patterns.
This all is changing the skill sets required of humans to take on various roles in the IT services sector. Previously, it was about ‘how can I write a piece of software or how can I use something that has already been developed by the customer’. Today one still needs the programming skills to configure a bot to do part of the work, but also the ability to apply human judgment once to bot has thrown up some information.
Therefore a potential recruit may not necessarily be very good at a particular programming language or another. But needs to have the aptitude for programming in general. And how fast can he or she pick up new skills — something that we call ‘learnability’ in the industry, and this is even more critical.
Q What are you doing about this at Wipro, and what are your biggest challenges?
Within Wipro, employees are all expected to up-skill themselves by training on one or more of the new requirements related to providing digital services to customers. Through 2018, already over 100,000 employees had each picked up a new skill relevant to the digital era.
On challenges, one is, in scenarios where one there are scenarios where one works with the customers to understand what’s needed. If the expectations are very high, that’s one hurdle to cross. Cognitive bots tend to become better with use, as they learn over a period of time. If a customer is expecting a high degree of accuracy in the beginning itself, then it will never get into mainstream. Higher expectation can lead to slower deployment.
The second challenge tends to how comfortable employees are working with bots. They will increasingly need to do some work on their own, but much more with the bots. This is a cultural conversation for most organisations. Third, the skill sets required for building these bots, are a challenge. Recruits who are ‘multi-talented’ are preferred. Some of these skills are completely new.
Two or three years down the line, the complexity of the tasks that bots can handle would have increased manifold. Consequently, the requirement that human beings be part of these processes and make high-level judgments will be even more critical.
One critical area will be about ethics — how bias-free will the bots be, because there is always the risk of the builder of a bot injecting his or her biases into the way the bot functions.
There are two ways of looking at automation. One is about building software to take over dull, repetitive tasks and release humans to do more creative work. The second is about tasks that are already of a higher order and of a nature that makes it difficult or even impossible for humans to attempt them.
For instance, a telecom company will have large volumes of data on various plans it offers and might want to know which customers are using what kind of data plans — something that one can throw software and raw computing power at, which on the other hand would be virtually impossible for humans to attempt. Humans can get involved here once the software has thrown up various patterns, and then use their knowledge of the business context to act on those patterns.
This all is changing the skill sets required of humans to take on various roles in the IT services sector. Previously, it was about ‘how can I write a piece of software or how can I use something that has already been developed by the customer’. Today one still needs the programming skills to configure a bot to do part of the work, but also the ability to apply human judgment once to bot has thrown up some information.
Therefore a potential recruit may not necessarily be very good at a particular programming language or another. But needs to have the aptitude for programming in general. And how fast can he or she pick up new skills — something that we call ‘learnability’ in the industry, and this is even more critical.
Q What are you doing about this at Wipro, and what are your biggest challenges?
Within Wipro, employees are all expected to up-skill themselves by training on one or more of the new requirements related to providing digital services to customers. Through 2018, already over 100,000 employees had each picked up a new skill relevant to the digital era.
On challenges, one is, in scenarios where one there are scenarios where one works with the customers to understand what’s needed. If the expectations are very high, that’s one hurdle to cross. Cognitive bots tend to become better with use, as they learn over a period of time. If a customer is expecting a high degree of accuracy in the beginning itself, then it will never get into mainstream. Higher expectation can lead to slower deployment.
The second challenge tends to how comfortable employees are working with bots. They will increasingly need to do some work on their own, but much more with the bots. This is a cultural conversation for most organisations. Third, the skill sets required for building these bots, are a challenge. Recruits who are ‘multi-talented’ are preferred. Some of these skills are completely new.
Two or three years down the line, the complexity of the tasks that bots can handle would have increased manifold. Consequently, the requirement that human beings be part of these processes and make high-level judgments will be even more critical.
One critical area will be about ethics — how bias-free will the bots be, because there is always the risk of the builder of a bot injecting his or her biases into the way the bot functions.