Saurabh Gupta — HFS Research[00:05]
Hi, everyone. Thanks for joining us today on this HFS Unfiltered videocast, where we really want to talk about how automation and AI can make managed services and business services sexy, right? I think 2023 is going to be a pivotal year for businesses and for enterprises. Look at the macroeconomic headwinds — there’s inflation, there’s supply chain disruption, there’s a recession. Europe is already in a recession, and we think the US will be in a recession next year, whether it’s small or long, we’ll see. And then there’s this talent crunch. I don’t think this is going anywhere unless we somehow magically find a hidden continent under the ocean. The talent crunch has got to be here with us. So enterprises need support for talent resourcing, technology modernization, and process transformation. And while we keep talking about this great resignation, I think in reality there’s this great hurry to get things done. And that’s the power of automation and AI that can change the narrative of business services. To discuss that, I’m delighted to have Shivani, who’s the VP for Automation and AI at Xerox, as well as Bharath, who’s the VP for Strategic Initiatives at Workato. So thanks, Shivani and Bharath, for joining us.
Bharath Yadla — VP, Strategic Initiatives, Workato[01:34]
Thanks for having us here as well.
Saurabh Gupta — HFS Research[01:38]
All right. So let me ask both of you — you’ve been in this industry for a long time now, in this business services, managed services, shared services space. What are the one or two biggest challenges today that enterprises are facing from a managed services context? Maybe, Bharath, we’ll start with you.
Bharath Yadla — VP, Strategic Initiatives, Workato[01:53]
Sure. Fundamentally, like you talked about, it’s a great hurry — that’s the right term to use. With all these things that have happened in the last couple of years and how they came together. But if we go back a little bit historically and see how it evolved from a managed services and shared services industry perspective, the first thing that happened was that a large part of the initiatives were driven based on cost arbitrage, or labor arbitrage, and that was the main thing it was used for. However, going forward, a large part of that got converted into using it with a focus on business operations efficiency, and then using that as a kind of spear tip to drive the strength of the services provided by these managed business providers. But then when the last two years hit, there was a lot more hurry — whether you call it digital transformation, technology transformation, or the way you look at automation, AI, and so on — everything came together, and what I call transformation is something people had to hurry into. When that happened, the biggest thing going on was to focus on the low-hanging fruit, the spots you can start looking at to tie things together, whether it’s simply automating and adjusting to the supply chain disruptions that were happening, and that’s one of the things you’ve seen. But the interesting thing — in fact, 2.5 years ago I wrote a white paper about how things would come together, and I was doing it as a kind of pipe dream, mostly sitting and visioning and dreaming about this — is what I call circles of impact, three different circles of impact. Think of these as concentric circles of impact that I was imagining. When the managed service providers initially started, they looked mostly at the first package, which I call very specific — a specific process, function, or area, like a hiring process or an ordering process. That was the main focus, to get specialized on that, because that was the strength they were bringing in. There’s a second circle of impact I was looking at: how could that be extended enterprise-wide, which means bringing all the functions and processes together so that a shared services, managed services, or business services provider can help these large enterprises take a quantum leap into enterprise-wide impact. The third thing I was looking at is that it’s not just the specific business or function area, or the enterprise, but also the ecosystem impact you have to bring in. That ecosystem impact is where a large part of the new revenue areas are. Take, for example, a claims processing provider who has been working with 150 customers and has an immense amount of knowledge about what will work and what doesn’t. Now, getting that entire knowledge in packaged terms and providing it to a large enterprise like an AIG or a Prudential — whoever the end customer is — and bringing back a new source of revenue, is the area that I think is top of mind for these service providers. That’s coming up, and it’s something I’ve very much taken into conversations with customers across the board.
Saurabh Gupta — HFS Research[06:15]
So, automation and AI for growth, not just for cost.
Bharath Yadla — VP, Strategic Initiatives, Workato[06:15]
Correct. Absolutely, without a doubt. So not just for cost, not just for efficiencies, but how can you spear-tip the growth and the new sources of revenue that make a top-line impact for enterprises? That’s the most important thing, and that’s where there is a lot more to be unlocked.
Saurabh Gupta — HFS Research[06:37]
Yeah. So, Shivani, let’s ask you. Bharath has been painting these concentric circles in his head — and they’re very aligned, Bharath, by the way, to our three horizons at HFS. But you know, Shivani, automation has been there for like a decade now. RPA was invented 10 years ago, or perhaps even earlier — we just started calling it RPA 10 years back. But if you look at our research, half or more than half of enterprises are not really satisfied with their automation initiatives. On one hand, we’re talking about driving new sources of growth, which is completely fair and should be the aspiration; on the other hand, we see this level of dissatisfaction. How do you marry these two things together, Shivani?
Shivani Agarwal — VP, Automation and AI, Xerox[07:39]
Yeah. So I think, for getting the real value from automation, a process has to be well understood in depth. As Bharath was saying, initially it was a cost arbitrage play, and then it was a process play, where you bring a particular process enterprise-wide, you consolidate that process, you build expertise in it, and then you automate. If that approach is taken, then the success of automation is likely to be much higher, because then you can anticipate, as the organization is automating the process, what different directions the process can take and what different business scenarios you need to cater to. And how do you improve that? In Xerox we call it the impact of the digital workforce. So you can put a bot on a process, and it may be that what’s in process is only 10% of your total transactions, but there is a journey you have to undertake to go from that 10% to 80%. You also have to realize that there will be some transactions where it will not make any sense to automate, because they are rare scenarios for which you may not get the returns. So really, anticipating the different directions, based on different business scenarios, that the process can diversify into, and factoring all of that into your automation plan, is what will give the business the satisfaction. Because for them, if the bulk of their transactions are automated, then there is real meaning, because that is when they start to see the benefits of getting things done faster, the benefits of accuracy, less rework, and getting information on time to make business decisions. So that has to be the vision. And also, it has to be centrally driven. I have seen many organizations where it is very fragmented. So, let’s say, Europe may be using one tool, the US may be using one tool, the Middle East may be doing something else. Then it becomes extremely difficult to manage and maintain these different islands, and it becomes expensive. So the cost play that you were going after, over time, starts to see diminishing returns. And also, tomorrow, when as an organization you consolidate your processes into, let’s say, one ERP or one CRM tool, then you are not able to easily update your automations to reflect those changes.
Saurabh Gupta — HFS Research[11:03]
No, that’s fantastic. So it’s process transformation along with automation — not just throwing technology at it and assuming everything is going to be hunky-dory, right? So, Bharath, how is Workato different than others? Or is it, because these are common challenges? What are you doing to address some of these things?
Bharath Yadla — VP, Strategic Initiatives, Workato[11:26]
So thanks for bringing that up, it’s an interesting point. I want to extend what Shivani was talking about, and also the previous question you asked about people having adopted intelligent automation, which has been there for more than 10 years, but, like I said, why has it not given the anticipated or expected results? So let me peel this apart again, pointing back to the same paper I wrote. What I was looking at 2.5 years back is two important things that were happening. One is, what does enterprise-wide automation look like? In my mind, this is what I’ve been talking to customers about, in a concise fashion: it’s a spectrum. You could look at it from simply taking the human-in-the-loop tasks that are there and start automating them, toward the end-to-end process that Shivani was talking about — entire process transformation. And then extend it into how you make it completely like a Tesla autopilot. That’s again a dream — having the entire enterprise automation, the whole operation, business operations, being automated just like Tesla on autopilot. So I painted a picture of what I call about six stages of automation. It goes through about six different phases of automation from a maturity standpoint, if you think about it. The first one, which I call the task-level automation, is basically where, like you said, RPA was invented about 10 years back, and even prior to that, but then it came into existence, and that’s where the focus was largely. That’s what I call the first mile of automation. The next 100 miles of automation is a big journey you have to bring together to drive that out. That involves bringing the systems together, like consolidation of the systems, modernization of the systems, which involves bringing the data together, and this is being done in a separate silo — basically all the middleware and integration that brings the data and systems together. That is being done in a different silo by IT, where the task-level automation is being done at the business level. So that’s one part of it. Then the third level, thanks to Salesforce, is what I call the SaaSification. Basically, 30 years back there were only 11 modules of SAP R/2 that people were happy about, but now, on average, in any enterprise you take, there are close to about 800 different apps — apps for this and apps for that. The moment you talk about different apps, there’s an orchestration between the apps that needs to come together, which led to the API-based approach. A large part of the fight when RPA emerged was that APIs are difficult to deal with, but now the market forces have converged many players to go with the API thing, and they’re talking about API automation being part of automation as well. That’s level 3. Level 4 is what Shivani talked about, which is end-to-end business process stitching, completely getting the transformation together, which is what the heavy-duty BPM players were doing. But now, with all this effort, you need lightweight workflow systems that come together. So that’s another part of it. Then level 5, which I call automated decision making: once you stitch the entire process together, the decisions are happening, the human interventions need to happen, and that gets popped up into their experiences — like, for example, a simplistic order approval system, PO approvals, and so on. You look at those and intervene, bringing those experiences in from a business decision-making standpoint, which relies on AI/ML, or just simple data, rules-based, or it could be deep-learning-based now with GPT-3 and so on. You could get to that level. And level 6 is what I call Tesla on autopilot, where the business is a little bit self-learning, kind of understanding itself. So these six levels are the biggest thing I see — people just started with that, which I call the first mile. It’s not that it hasn’t taken off; it has taken off, but that’s only the first mile of the operation. The next 100 miles is where the actual journey is, to get the entire thing together. Now, the second part of the question you asked about — sorry, you had a question, please carry on. The second part of the question you talked about was how is Workato different? So when we founded Workato, the idea was work automation — that was the main thing, and that’s how the Workato name came into play. The idea was that for doing this work automation, you have to have the data, process, and experiences together in union. You have to stitch those things together, in whatever way and fashion, and that’s exactly what Workato as a platform brings: data, process, and experiences together in one unit. So, to say it shortly, in all six levels we talked about, level 2 to level 5 is where Workato plays extensively well. So Workato becomes the vehicle for the next 100 miles of that automation journey, and that’s exactly how we look at ourselves.
Saurabh Gupta — HFS Research[16:47]
Fantastic. No, that was very professorial, but that’s great. So, Shivani, how far are you from that Tesla autopilot stage, and what role is Workato playing in your next 100 miles of automation?
Shivani Agarwal — VP, Automation and AI, Xerox[17:04]
Yes, so when you say autopilot, that goal is a little far, but we have a path to it now. We did start with RPA. We started automating a lot of our basic transactional activities that were time-consuming, high-volume, where we just had to swivel data between applications. So we took all that mundane activity away from our people. We have automated a whole lot of reporting activities and data entry activities, but then there is a lot of exception management, automated decisioning, and bringing the human into the loop of the automation, which is what we are working through right now. We are injecting a lot of document AI into our automation right now. This whole journey, the industry calls it hyper-automation. So we are moving from plain-vanilla RPA to hyper-automation very rapidly. And if we have to get to an autopilot stage, it’s not going to be one technology versus the other, but stitching together all these different technologies. So, for example, for a document-based process, you will have document AI, you will have OCR, you will have RPA, and you can have Workato. Workato is coming into the picture for us to perform that last-mile activity of pushing the extracted, validated data into the business applications or ERPs. So we are also using Workato in our commercial offering, where we are automating processes for our customers, and we are realizing that a lot of ERP systems our customers use have ready-to-use Workato connectors. We obviously don’t want to be building APIs or connectors for each and everything and maintaining them. So, since Workato offers a library of pre-built connectors, and it also offers an easy way of building recipes where we can transform the data we have extracted from a document or received from an external source like a customer portal or an incident ticket, we can then transform that data to meet the requirements of another business application and automate the transaction using the back-end connectors. So it speeds up the automation — that is one thing — because you’re not working at a transactional level on the UI interface; rather, you’re doing an API call to instantly pull the data into the business application. So one is speed — you are able to do business now at the speed of software. And the other thing is maintenance. Whenever you do UI-based automation, we know there will be changes to the UI or the application we are working on, whether it is Oracle or SAP; everybody does these quarterly upgrades to their platforms, and those quarterly upgrades have some impact on our user interface. So it may impact 1 out of 100 bots we have running on that particular application, but we have to test all 100 every quarter. So making our automations more robust, and taking away the brittle nature of the UI-based automations wherever we can, is what led us to working with Workato.
Saurabh Gupta — HFS Research[20:17]
Fantastic. All right, this was a terrific session, but I won’t let you go, both of you, without answering this question. So, as we’re getting into 2023 — and I’m pretty close to the new year — what would be your one wish, if it could come true, for next year? Shivani, what’s your one wish?
Shivani Agarwal — VP, Automation and AI, Xerox[20:56]
So I wish we could train ML models with less manual effort. That is one wish. And the other is that we could spend less time on maintenance and more time on driving our automation journey forward. So those are, I think, the two things from an automation perspective, where we are always on the lookout for technologies that will allow us to use AutoML to do trainings.
Saurabh Gupta — HFS Research[21:42]
Shivani is in a greater hurry. And Bharath, what’s your one wish for 2023?
Bharath Yadla — VP, Strategic Initiatives, Workato[21:52]
My one wish, which is very apparent for me right now — in fact, you and I talked about this quite heavily — is what I term the fatigue that has kicked in because of the UI-based path to automation. People have used the UI-based path to go on that automation journey, and that is something very prevalent with customers. If there’s one wish I could do, I wish I could just throw a wand, like Harry Potter, and make it go away. But that’s not that easy to do. So that’s exactly where the 2023 mission I’ve taken up on my side is to go in front of customers and say there are places where you do need to use the UI-based automation, which is green-screen automation — go use that. But the key thing is that when there are modern systems, and you’re living in the modern systems, you can drive the automation in a very different way. You’ve got to think about stitching up the data, processes, and experiences together, and that’s exactly how you drive those automations and make your life much easier. So there’s a place for this and a place for that. That’s the core aspect, and I think that’s the mission in 2023 that I wish I could take to customers and say, this is what you should be looking at.
Shivani Agarwal — VP, Automation and AI, Xerox[23:06]
Yeah, as I said, it’s not one technology versus the other, but bringing them all together and using them appropriately based on the need of the use case.
Saurabh Gupta — HFS Research[23:33]
No, this was a fantastic conversation. Thank you, Bharath. Thank you, Shivani. What I’ve realized over the last 15 to 20 minutes that we’ve been talking is the power of “and.” As you mentioned, Shivani, we, as humans, have this tendency to keep comparing one versus the other. We don’t realize the power of “and” — if this and that can join hands, whether it’s people, process, and technology; not people versus technology, man or person versus machine. It’s not just one type of automation, but multiple types of automation. So this conversation has helped me realize that power of “and.” And hopefully some of you who are listening to this will also realize that it’s not really, as Shivani mentioned, one versus the other, but one plus one. That’s going to make our dreams come true in 2023. So thank you, Bharath, thank you, Shivani, for taking time out of your extremely busy schedules, and I wish you happy holidays and a very, very happy New Year.
Bharath Yadla — VP, Strategic Initiatives, Workato[24:35]
Thank you so much, and I wish you a very happy New Year as well.
