In this Digital CxO Leadership Insights series video, Mike Vizard talks with Ritu Dubey, global head of new business sales and market development for Digitate, an arm of Tata Consulting Services, about the forces shaping the rise of the autonomous enterprise.
Mike Vizard: Hello and welcome to the latest edition of the Digital CxO Leadership Insights series. I’m your host, Mike Vizard. Today we’re with Ritu Dubey, who is head of market and business development for Digitate, and we’re talking about the rise of the autonomous enterprise. Ritu, welcome to the show.
Ritu Dubey: Thank you, Mike. Thanks for having me. Pleasure to be here.
Mike Vizard: I think we’ve been pursuing this dream of autonomous enterprises for some years now. I mean, it’s always been a topic of conversation somewhere along the line. It tends to come and it tends to fade, but there’s been a lot of progress on AI and all kinds of stuff lately. So what’s your sense of how achievable is the idea of an autonomous enterprise and to what degree is it even desirable?
Ritu Dubey: Yeah, so I think I’ll, first Mike, a little bit set the context of what we mean by autonomous enterprise. And as you know, automation has been there for a very long time. And when we talk about autonomous enterprise, essentially it’s about having hyper automation and hyper instrumentation within an enterprise so that the enterprise can be self-governing. And if I have to a little bit break it down, there are three parts to this, right? One is that you build a platform and a capability in an enterprise that is capable of doing three things. It’s capable of learning, which means adapting to the changes because IT has become so dynamic in most of the companies. So it’s able to adapt to your changing technology environments, changing policies changing workloads. So this whole adapt and learning capability, which AI and ML kind of a solution brings to brings into place.
The second, the thing is that it has to have the capability of not only to reason, but it also has the capability to also take actions. So sophisticated decision-making, but the ability to take action. So, all of these three things from our perspective closes the loop for an autonomous autonomous enterprise to achieve a high scale of hyper automation. And all of this underpinning, all of this has to be AI and ML because the kind of volume of activities and the kind of volume of data that is being generated today in IT enterprises, it is almost humanly impossible to maneuver through that data and the set of activities and therefore it becomes a big play for AI in enterprises and therefore that is sort of the vision that we are talking about for enterprises, self-governing and has all of these capabilities.
Now, in terms of the question you asked, how achievable it is. It is absolutely achievable, but if you ask me, has everyone reached that utopia, that stage, I would say no. I think everyone is on that journey towards achieving that autonomous enterprise. And there are examples which I can talk about where companies have actually embarked on that journey and actually have seen some superior benefits and results in that space.
Mike Vizard: I think a lot of organizations are challenged with the whole idea of – it seems like they have more exceptions than they have rules, and that makes it hard to figure out what to do with if this, then that. There’s a lot of dependencies and complexities. Is that something I need to go rationalize before I do this? Or can the platform flexibly address all those different use cases that tend to pop up?
Ritu Dubey: Yeah, I think on an average, if you see for most enterprises, I would say 50% actually of the operations work in an enterprise is pretty structured. It is pretty repeatable in that sense and it is quite stable. So I think the starting point is that low-hanging fruit, and even if it is structured and it is stable and repeatable, but there is still a lot of volume of work that needs to be done. I mean there are enterprises that you would see that how you have thousands and thousands of issues that come on a daily basis, they may be repeatable in nature. So I feel that first of all, you have these low hanging fruits where you can directly start applying this kind of capability already because there is a lot of automation in terms of processes or deriving insights that you can do in that space.
So that’s one part of the story. The second part of the story where you say that it may not be rule-based, right? And there is definitely inherent complexity in IT today. So actually there’s a lot of activities that organizations do, which is around planning. They do, around how do you scale your infrastructure to the future, how was the performance? So there is a lot of need to look through a huge amount of data to find that pattern in massive data sets. And again, in this specific case also, that makes a great case for AI because as I mentioned that AI’s capability to learn means that everything does not need to be codified in your environment. You’ll not codify every scenario that exists in your enterprise because AI has that capability to learn. The more it is in your enterprise, the more it learns about your context, the more it is even able to deal with situations which are new for the software.
Mike Vizard: I think a lot of organizations have struggled with everything from robotic process automation to chatbots and all these other things that they’ve put in. What mistakes do you see people making as they go down this path? Or is it simply just a matter of trial and error and getting experience and it’s the cost of learning?
Ritu Dubey: Yeah, I think robotic process automation… I think one of the challenges with robotic process automation has been that because it’s just a matter of what we say, mimicking human action versus mimicking human behavior. So AI is about mimicking behaviors of how humans think about solving problems. And robotic process automation was all about let’s just, we do a manual process just digitize that manual process. And what happened with that is we ended up creating a lot of bots and whenever there was a change in the rule, then you had to change the bot. So essentially it created more work than it helped. It helped initially of course, but it ended up creating more work. But when you shift towards intelligent automation, that’s where again, going back to the capability to adapt to the changes that happen in your environment and to adapt to the changes in the context without codifying everything. That sort of ensures that as you grow your enterprise, as you add new technologies continuously, AI is able to adapt to that kind of solution.
I think the starting point is definitely… I mean you have to start in an area, you have to start with the problem statement, which is the most painful for you. And I think one of the areas where we actually see this kind of solution applicable is sort of what we call the business assurance side. So business assurance for us is can you ensure that the revenue generating systems of your organizations are always up and running? Because I mean that’s sort of generates a revenue for the company and that has become a great use case from our point of view where we are seeing the applicability of solutions. And I’ll give you a few examples. So for one of the utility companies that we had, they had a lot of what they called unbilled accounts that were not paid by the customers and they were not paid by the customers because there was a lot of manual intensive work that the energy company or the utility company had to take care of before the bill was sent to the end client.
And there was a huge backlog of those bills because on a daily basis this company generated almost a hundred thousand bills per day and there was a lot of errors, manual errors. Now just applying a technology like this to that single process, one single use case helped them save almost $4.2 million on an annual basis. It’s just one use case. So I think it’s a matter of finding where is the maximum pain point for you and starting at that place that, can it help you solve that problem. The other thing that we see, which is important because return on investment is always going to be important for such a technology because you’re not just going to implement it for one use case. So as we start to implement the starting point, we also see that customers try to create the whole business case. Where else can it be applicable?
And I think if you start by doing these two set of things, we have found that the customers are able to achieve their ROI. In fact, our data point shows that it’s able to achieve, we are able to achieve almost 185% ROI in a span of three years of course on an average. And some customers have even been able to achieve 800% ROI, but they are outliers. But on an average, a great ROI is possible when you start looking at from the perspective of what’s your pain point, where do you apply it to begin with so that you can of course propagate that message internally because everyone has to buy into it as well for the success of the program and for sustainable automation.
Mike Vizard: Do organizations sometimes attempt to maybe boil the ocean and they try to go too far too soon? And as that seems to be a lot of the dangers that I see with a lot of organizations, then they have a reaction that’s understandable. But is it really about just picking the right task at the right time to get the ball rolling?
Ritu Dubey: Yeah, I think you’re right. Some of the companies that we have worked with who have really gone all the way, I think they have set it out as a program. They actually call it AI first culture. This is our culture, this is the culture moving forward. And when you set that kind of a context in your enterprise, then everyone starts to start playing with it. And the other thing that I see is also can AI become for citizen development? So of course you are starting in a certain area, but can you create that capability or give that capability to the enterprise to start playing with it and start seeing that if it can help them in their job? So I think if you have that kind of a perspective and build that narrative in your organization, success will come. And we have seen that success in many customer organizations, but the ones that tend to fail is when they don’t set that context right, because you see there is always a fear.
There is a certain amount of fear in terms of AI and jobs. And all we can say is in all the enterprises that we have done, it’s about augmenting the human intelligence. So it’s actually helped people to do something different in their organization, which is much more value adding. So that’s the narrative I think organizations need to build. And to just give you a data point, one of the retailers, Canadian actually retailers that we work with, they said their ambition was to have almost 80% of the workload in their enterprise to be AI first. And with that mandate, they have gone ahead and every possible entity of the enterprise, whether it is their retail stores, whether there’s payment systems, ERP, whatever they can find, they have been able to drive that culture and drive that very quickly. They have achieved 60% already. So in a span of four years, they went to 60%. And of course they’re doing the journey. So it’s possible, but it has to come from the right level and it has to bring people along on that journey.
Mike Vizard: Speaking of bringing people along. I mean, most people will leave their job because they don’t get along with somebody, but the second-biggest reason is that whatever they’re doing is kind of mind-numbing and they just get tired of it. So can we eliminate a lot of the toil that folks over time come to resent and maybe they might actually like their jobs better?
Ritu Dubey: No, exactly. I think can we take away all this repetitive, mundane kind of tasks that people do and give them an opportunity to learn of course, and do something different and something new or even become what we say trainers of such softwares and solutions and absolutely, I think AI makes the case for it. And there are many scenarios where the existing teams have really adopted it because it helps them do not only their own job better, which they’re mandated to do, but helps them take the opportunity to also learn something better or something new.
Mike Vizard: How do I put the guardrails in place? Because as one wise one said to me: It’s one thing to be wrong, it’s another thing to be wrong at scale. So how do I put something in that’s automated that has all this infusion of AI, but make sure that something bad doesn’t happen?
Ritu Dubey: Yeah, I think the AI that we are talking about, I would say that you can define certain rules of what it is allowed and what is not allowed to do. That’s the first thing right? There is always that kind of a setup that is possible with these kinds of solutions. But the other thing also I would say is from that standpoint is these, when you deploy these solutions, you just don’t go and say just here, go ahead and let AI take care of everything, right? There is always a gated process. You have to be able to trust the technology to do the right set of things. And we have seen in our experience that usually there is what we call the hockey stick effect, right? You go in, you prove, right? And there are people who actually validate what the machine is doing and is it inferring the right set of things, is doing the right set of things, and then that trust is built over time.
Before that, you can say that hockey stick effect takes place. So it’s always going to be about that. And of course you do certain principles, you define certain principles in your enterprise, one which are controlled, centrally controlled, but the other, you have to define that guidance and guardrails that you are allowed to do a certain set of things or you should not be used for certain set of things. And I think that the corporate governance is important around that. But what we have seen is within the such solutions itself, you can define what AI is allowed to do and not allowed to do and gradually build that comfort and trust over time.
Mike Vizard: What is the one thing, or maybe a couple of things that you see organizations doing over and again as they start this journey and just makes you shake your head and go, “Geez, I wish everybody knew the following so that they could avoid this heartache at the beginning.”
Ritu Dubey: Yeah, I think one thing I feel is that, which organizations don’t think about is, it’s not about just mimicking what you do today, right? I think they don’t see the potential of AI as an opportunity to reimagine what they’re doing today. And I think that the great potential of AI is not just around, let’s just do what we are used to doing, but can we do things different and make it much more effective? And I think that’s one of the things, that’s what I see is one of the things that they don’t recognize upfront and it takes some time to work with them, but once they see that, it just opens up their mind because that’s where the revolutionary impact of AI comes into the picture.
The other thing I feel is that this notion of what AI is, I mean there were certain technologies, sort of what we call the general purpose AI thinking that we have achieved a general purpose AI kind of status. We have not. There is AI that is defined today that works well in a certain domain, and I think it’s important to understand what that domain is and how it can work in the domain such as the IT as a domain and how does it work in the context of it. So I think that level setting of what AI really is today and what it can achieve today, I think that is where I feel that sometimes people have different notions and it’s important levels of that.
Mike Vizard: All right folks, you heard it here. If you’re using AI to just automate the same old process, you might be missing A: An opportunity and B: The point. Hey Ritu, thanks for being on the show.
Ritu Dubey: Thank you for having me, Mike.
Mike Vizard: All right. And thank you all for watching the latest episode of the Digital CxO Leadership Insights series. You can find this episode and others on the digitalcxo.com site. We invite you to check them all out. And once again, thanks for watching, and we’ll see you next time.