AIIA.Net Benefactor Interview: PR Krishnan, TCS
TCS thought leader shares his thoughts and views from inside the AI industry
PR Krishnan is a transformation stalwart. He's been on the leading edge of technology for decades. Currently based in Channai, India PRK and I connected to discuss how his storied career is affecting TCS' current AI and Intelligent Automation strategy.
Early in the conversation, PRK notes that while Tata Consultancy Services (TCS) has been providing automation solutions for quite some time, Enterprise Intelligent Automation was formally organized recently, "we initiated this practice early part of last year. It was not actually that we were not doing it. We just decided that this could be an area which could probably transform our enterprises going forward.
Because if you look at the digital roadmap that we want to have for our customers, it is essentially going to be built upon four key themes. It is going to be agile, it's going to be cloud, it's going to be intelligence, and automation."
Agile, Cloud, Intelligence Automation
PRK doesn't just conceive that these are central themes to simply digitalizing what you do- he finds that these central themes should change what you do, "we felt that we have an opportunity to transform every aspect of what an enterprise IT is doing today."
With legacy players in the Global 500 dropping out day-by-day, the definition of an enterprise has already been redifined, "there are new business models coming in. There are new productivity teams coming in. There are new ways of interacting with your customers coming in."
No matter what you do now - you must relate to your entire ecosystem, "be it interacting with their customers or be it interacting with their employees. Or they have to interact with their partners and their suppliers. All aspects of that are changing because digital is making it possible."
"If there are business processes being run by the partners of customer, we will see how we can replace it with robotic process automation to improve time efficiencies..."
And so, to recognise those realities, PRK and TCS created three practices to speak to enterprise:
Intelligent process automation looks at all the business processes of an enterprise to decide what can be done to transform the way the operational processes are run today.
It asks the question, "can anything be improved by adopting a robotic process automation?"
The goal is to speak to "new efficiencies and speed of operations allowing customers to go to market with their plans faster."
And the key is that it's not just one-and-done automation, leaving you with a PoC and initial implementation and no way to scale, "it doesn't stop at RPA, it extends itself to the cognitive intelligence and artificial intelligence and machine learning and to all of them. To process some of the inputs which come into an enterprise."
Which leads to the fact that intelligent automation is only in part about autmating repetitve tasks. There is of course value in doing just that, but if you only execute that aspect - you're left with an enterprise that's only slightly more efficient than the one you have today. You have not put a dent in your strategy and you have not set a roadmap for where your organization is going. You've simply made the firm look good for then next few quarters.
PRK is talking about it truly taking bold steps to ensure that you're setting a path that actually connects to tomorrow. And winning tomorrow is all about being able to get value out of today's enterprise dark data, "it could be any type of unstructured information or data which comes into an enterprise."
With that end goal in sight, the conversation is around end-to-end intelligent automation that incluces cognitive capabilities from the start, "in very simple terms, if there are business processes being run by the partners of customer, we will see how we can replace it with robotic process automation, apply artificial intelligence techniques to improve the number of transactions that I can process per unit of time."
Application Services Automation
So much of today's automation focus is on process- whether that aspect has been optimized or not. PRK and his organization saw a further opportunity to take a next step into automating application services.
"This is something like what we felt was a white space," says PRK.
"Not because nothing is being done in this space, but because nobody's focusing this as an opportunity where we can bring in efficiencies in the way we develop applications within an organization to meet the customer's business needs."
Speak to anyone in the industry and it's clear that we've moved from a waterfall to agile approach to ensure that a more rapid cadence in going to market. But if your dev ops teams are running app deveelopment on legacy systems, "there could be a drag in the enterprise."
"We are looking at opportunities in the whole life cycle management for automation opportunities."
Here's where enterprise transformation can fall flat on its face. If one hand doesn't know the other's methodology.
"There is an opportunity for TCS to be able to transform the whole application development landscape moving from maybe a waterfall model to an agile dev ops model," he says.
"Or automate the whole testing part of the whole software development life cycle. Or moving away from traditional infrastructure to a agile cloud infrastructure where the development can happen. There are different opportunities that allow us to make it a much leaner and faster application development environment. So we are looking at opportunities in the whole life cycle management for automation opportunities."
And enterprises who are employing change management effectively to ensure the entire enviroment is on the same page, are "literally bringing up their productivity multi-fold."
Intelligent IT Operations
With a significant focus coming on automating business process, as noted above - and an opportunity to focus on application development to absolutely ensure the entire enterprise is on the same page - PRK seeks moving enterprises to intelligent IT operations as a complete step-change "this opportunity is definitely huge. This opportunity has the good probability to change the way enterprises are run tomorrow."
And this is precisely where the relationship between business and IT must improve and PRK understands that it's TCS' duty to play an important role for the industry: "I think this is where TCS has to play the role of being the custodian of customers business as well as custodian of customers assets."
Enterprises are in the process of learning the lesson that automating business process is only part of the solution. If IT isn't on-board, or included in what the business is doing- eventually you reach a point of diminishing returns. One of the central keys to automation success and true business transformation is, of course, ensuring that IT is on the transformation journey.
"We have the ability to bring the IT into confidence and insure that what architecture we are implementing kind of dovetails IT into it. So that when IT changes, then the automation component will also mature into what it is to be in a newer model or operation."
"Positioning automation as a special initiative is often the culprit in having to start over after initial automation success which simply can't be scaled."
Speed to market might be the issue. In other words, the business is just interested in getting things automated as quickly as possible to ensure that the enterprise keeps pace with competition.
And so intelligent automation is often set up as a special initiative instead of ensuring that business and IT are on the same page from the start. And positioning automation as a special initiative is often the culprit in having to start over after initial automation success which simply can't be scaled.
"You have to understand whether you're talking about deploying bots in the case of the RPA. Or deploying any of the intelligent automation in the IT operations space. You will realize that at the end of the day, the operations guy on the floor is going to be worried that if something goes wrong, the bot will not be questioned - I will be questioned."
Change management being mentioned, I asked PRK how much of his time is actually focused in change management versus automating anything? How much of his time is spent with management levels? Or with employees in general? Or simply engaging in change management?
His response: "Change management happens, like you said, in two or three levels. The first level is that when we initially start engaging at the executive level, I think at the executive level, they kind of get it in the sense they know that it's got to be done. They just have to be careful about organizational messaging when they start implementing these initiatives."
Having ensured success through demonstrating what is necessary in terms of top-line communication, PRK and his team can then dive in to OCM, "working with our customers to insure they send the right messages the employees- the right messages to their business, to their partners. So there is work which is done ahead to insure that our initiators are well received when you go to the floor. Because at the end of the day, if that doesn't happen it could actually thwart the whole effort to bring in transformation into the organization. Which is natural in anything we do new."
When I asked him how similar this was to the ERP implementations or initial offshoring of yore, and he said it's essentially the same, "it is part of part and parcel of what we do as a consulting initiator to do the initial part of any and all this transformation initiators who work with our customers. There are three things which happen during the whole up front processes, what we call as consulting. One is about organizational change management, what does it do in terms of impact to the employees? Or to the partners? Or to the customer's business? That is a communication which has to be prepared and the organization has to be ready to accept this transformation."
And once he and the team are within the organization, it's on to process optimization which leads to the actual change in tools and technology.
And so if transformation is continuous, what about the technology at hand?
With a bird's eye view, while integrating machine learning into the enterprise is absolutely the current or next step for many, it's a difficult step to take.
"If you look at the automation roadmap or the transformation roadmap we want to drive for our customers, you will find that artificial intelligence and machine learning algorithms would essentially be adopted in one way they're at the highest level of abstraction. We will try RPA, then RPA will take you some distance. Then you start putting heuristic patterns, pattern matching, behavior patterns, and make that RPA much more effective. Then you start getting into machine learning algorithms because there you start the tool which you are deploying or the platform you have deployed to start learning from either history or what it can predict for the future. So it is kind of getting complicated as we go down that maturity part. I would argue that our enterprises have done that to a large extent. Most of them will have come to if I have to call them at level one, level two, level three. AI level is the highest level of maturity. Most of them will be struggling between level two and level three."
Learning from history to predict the future is all about the inputs- text, images, graphics, video.
"Most of that, there is an element of AI ML in it. That is where most of the noise comes from or which is the most visible part of what AI ML can do to you- that we can get this unstructured, semi-structured data and convert it into structured data."
Sense, Interpret, Integrate
Most enterprises have take the step to sense their data as noted above, they're up to interpreting that data through machine learning. And once that data is interpreted, the enterprise can integrate those learnings back into the organization to better serve customers.
"That particular maturity in terms of being able to integrate what comes from the front end and integrate it to the back end where the maximum knowledge about enterprise- it's at a very immature level. We have some distance to go."
Cloud Enabled Transformation
And so, we're truly at the beginning of realizing the potential of AI for enterprise.
And understanding that PRK references lessons learned along the way, "transformation of enterprises does not happen in isolation in automation. It's a combination of different things. When we say combination, we talk about automation as one piece. The intelligence we bring about enterprise, we are able to encapsulate in a learning platform. The third part is how can we make the enterprise agile? The fourth, which is something which you have not talked about in this conversation, is about putting them on a cloud. The cloud gives them the flexibility of elasticity. They kind of being elastic about their resources they want to deploy in their enterprise. So these are the four big things, which actually make up what we call is the transformation roadmap for a customer."
Transformation is incremental. Work on the front-end today. Work on the back-end tomorrow. Simply focus on data for the next two quarters. No matter what your focus, ensure you've got a solid strategy. Ensure your strategy isn't scattershot.
"What TCS is doing is we are able to take it in chunks of different parts of enterprise and be able to do it in a very incremental transformation model. So that we don't disrupt their business too much that they become really uncomfortable."
Machine-First Delivery Model
You've got your solid transformation strategy in place you're optimizing business process, you're automating application services and you're engaging intelligent IT Operations.
"The most important theme of the machine first delivery model is to say that our delivery model will empower our people to do more and make them much more efficient. That's the bottom line. Anytime you talk about automation or you talk about machine, they're essentially telling if you augment the capability of your workforce and make them do and empower them to do better than what they're doing yesterday. That is the fundamental theme on which this has been built. It is not about machine versus man, but it is about machine with the human."
The inference is that tomorrow's most successful enterprises will only find the top of the pyramid if they allow automation to make your carbon-based workforce as efficient as possible through automation. It's not 1 minus 1. It's 1 plus 1.
"The way we communicate to our people is that the transformation is not about technology, the transformation is also about the people. They will have to learn something different from what they have been doing until that day. Their roles will change, their responsibilities will change, and their competencies will have to be enhanced."
The key to machine learning is human learning. The human enterprise workforce will only succeed by improving one life-form at a time. Transformation is about showcasing to the current brightest talent, that they are needed, but only if they themselves transform along with the enterprise.
Stay up-to-date with AiiA.Net for more case-study led insight from TCS.