Join us weekly to discuss the latest and greatest in low-code and digital process automation with executives and experts. Real conversations, no marketing BS. Hosted by Rob Koplowitz, John Rymer, and Ryan Duguid. Visit analysis.tech to get in touch about your personal low-code journey and learn about ways we can help.
Dave Marcus (00:09)
Well, hello everyone. This is Dave Marcus, along with my partner in crime, Francis Carden for the latest in our series on Lowdown on Lowcode from the Analysis .Tech team. For those of you who don't know me, I've been in the automation and lowcode industry for a long time through engagements and with Microsoft, I spent a bunch of time at K2, now part of Nintex, and then I was at UiPath for five years.
Prince Kohli (00:18)
you
Dave Marcus (00:35)
before leaving about a year and a half ago. Recently joined the analysis .tech team. I'm going to hand over to Francis and then we'll introduce our guests for today. We're very, really, really excited to have Automation Anywhere on the podcast with us today. So Francis, over to you. Great to see you again, as always.
Francis Carden (00:51)
Yeah, likewise, Dave and Hi Prince, welcome aboard to our podcast. Just a quick intro. I've been in, Dave and I talk about how long we've been in automation, this great word automation. I started in as a software developer in the low -code space. Let's just call it the free GL world because it didn't exist then.
Back when I was 17, roll forward 40 years in between that I've had a software company that was doing all the nations and mainframe screens roll forward. then had an RPA company before it was called RPA back in 2005. We started open span and then we sold that to Pega systems, one of the low code vendors in 2016. And then I retired from Pega back last year in October. And so really I've got the whole gambit from.
Prince Kohli (01:34)
Thank
Francis Carden (01:37)
automating really old legacy stuff to really being able to reimagine software and software development and so on everything in between. And so we competed with most of the RPA vendors for many years. It's a lot of fun seeing that ride up through that, through our exit and seeing what you guys have been doing from the beginning. And I know you want to...
you know, from an IPO perspective, you didn't get to IPO and I know that, you know, that may actually be a blessing, right? So, but it's been exciting times. And so we're really, really interested to hear your story here. You know, give us a little intro to yourself and what your role is and what led you to this world.
Prince Kohli (02:14)
Thank
Thanks Francis, thanks David. It's nice to be invited, nice to be here. I'll try to do a very quick kind of journey recap for lack of a better phrase. So I started my career now a very long time ago, 25, 27 years ago, building the first DVRs, that's silicon graphics. Some of us were slightly older. I will remember SDI.
And then I was building the original DVRs before it became an industry. then from there, I moved on to the, during the dot com boom and burst, I moved on to my own company. I was a co -founder of an application firewall company. We were building now they're part of everything, every stack in the cloud, every website. At that point, it was new things like, you know, cross -site scripting, SQL injection, buffer workflows. Those were new terms at that point.
So we were building essentially devices and software to do inline real time monitoring and more than monitoring actual blocks and so on. So very exciting then went to Citrix for many years. We were building cloud at that point for what became AWS and then Google Cloud and so on. So being kind of early for started kind of, would say many industry defining plans early, always exciting to do that, always stressful to do that, but always exciting to do that. Then,
As part of that, what I became convinced of, know, virtualization was very big, but what I felt very strongly and I continue to feel that today is data is going to be key. know, when data is platinum, gold, whatever your favorite metal is, precious metal is. So I went to Ericsson and led global R &D for them all over the world. And the reason for that is, excuse me, sorry.
Someone like Ericsson, they have a customer in literally every country of the world except maybe North Korea and India. And because of that, the amount of data that goes through their own network, as well as they manage it, they are both the endpoint and the network and they do application data at the networking layer. It gives them access to a lot of insights. It's not just operational stuff, right? How do you make sure things don't break? It's a full network.
And at the end, things must be very resilient. So you have to use actually very early AI at that point as well. But then also, how do you then understand scale? What do you have to do about scale? What insights can you derive? Then Ericsson, then after that, I came to AI directly, ThoughtSpot, and then Automation Anywhere. I've been here for six years. I'm the CTO of the company. I'm responsible for taking the company towards both AI into AI machine learning and now GenAI for the last two years. And also,
take the company into cloud. Very, very customer friendly from that perspective, easy to use, being in the browser, and highly accessible. So yeah.
Francis Carden (05:07)
Yeah, that's a great story. a lot of our guests have great stories, a lot of history there. You mentioned the word injection. I think you have to be a real science engineer to understand in the tech world what things like SQL memory injection are all about. That's something we need to discuss over a beer print. That's deep.
Prince Kohli (05:15)
you
Yes, sir.
Francis Carden (05:26)
And I also have discussion around data being key and I have some love and hate relationship with data per se, so we can also discuss that over a second beer. listen, historically RPA, right? Most RPA initiatives have been centered around task automation. That's how it started. know, screen scraping was a term, I think, and then RPA became a term. And I think that's still what many people think about the RPA space. And I think that's also what automation anywhere is very well known for.
Do you see with your role as well, AI evolving to be more of a provider of that reimagining processes, end to end process orchestration, and even dare I say it, that low code term that often gets misconstrued with visual IDEs, et cetera. Or, you know, and you're moving towards that higher value around the enterprise business apps, you know, that future domination of what they're going to look like. Or is AI staying predominantly with the bread and butter and it's great business model that it has and been very successful at?
Prince Kohli (06:24)
Okay, so lots of very interesting topics in there. And let me see if I can do first a broad brush and then talk about a few of them. So RPA, what was traditional RPA, if there's such a term, was really the very beginnings of what we call Gen 1 automation, true automation, not just kind of macros and so on. mean, that's okay, I wouldn't, I would say that's probably very low level automation.
So RPA was the first actual progenitor of real automation in enterprises. You could go across apps, you could take data, you could do controls, you could do actions. And we call that Gen 1 automation. And then what we call Gen 2 is when machine learning came in into RPA for the first time. And that allowed you to start understanding the semi -structured or structured documents or messages. Also, you could do a much better job of understanding screens or...
text on the screen or applications itself. And then two years ago, give or take, with Gen AI came what we call Gen 3 automation. And the difference in them, if you look at it from a customer perspective, with Gen 1, could affect whether it's bottom line or top line, you could affect millions of dollars. With Gen 2, you could do tens of millions. With Gen 3,
we actually have customers who are already starting to see either top line or bottom line impact of hundreds of millions of dollars. And they have gone in public. They have gone public about doing that. So the change that has occurred has been exponential for the last many years. And we gently the ability to create automations via natural language interfaces without anyone really understanding how the software gets written because it just seems automatic to them.
And the ability for the software to handle the environment around them across applications, across processes, you you said that earlier process orchestration. And that has become, I would say it's, it is beyond anyone's imagination or best hopes a few years ago.
Francis Carden (08:22)
So we do, we do, we do concur with that. think that's that's a very strong statement. What I was trying to lead to with my questions is that RPA generally is automation of as is processes, wherever that's as is documents, as is screens, as in usability by users. The end to end processes, we reimagine them in low code says, okay, what's the next generation of processes and entire applications going to look like? Are are you moving in that direction as an AI? Do you see moving in that direction?
Prince Kohli (08:48)
Yep. You know, that's, actually, so we actually have done that. And let me, the best way to say that is not just me saying it. Let me give you a customer example, a real customer example. So one of our larger customers in South America, they're a large oil and gas company. What they have done, and I'm going to kind of jump directly into AI and Gen. AI with this, is they're using Gen. 3 AI to build AI agents on our platform.
AI agents are that, and these are basically they build a tax agent. They're able to use those tax agents now to not just understand the data they have in the enterprise, to not just drive automations, but to actually start understanding their tax regime. Are they underpaying, overpaying? What are they historically done? What should they be doing? And across different countries, right? And based on that, they can actually make great decisions about their own tax filings.
Then there is the customer, very large bank, who has gone from many, many days of tax loan prep into hours using a loan agent that they built on a platform. So what has happened is what used to be, going back to your previous question, what used to be RPA, the industry for us in our use cases are expanded to go across end -to -end large complex processes that may run over days or weeks or forever, may never end.
allowing you both orchestration, resiliency, as well as insights into how the company is performing. So I think that has kind of been a massive sea change in how people perceive automation. It's not no longer task automation really, kind of any cutting edge platform today, state of the art platform today, should give you process orchestration and process visibility.
Dave Marcus (10:30)
Cool. So Prince, guess a question for me is obviously you guys have a really big customer base today. And as you evolve and as you innovate, I think it's, know, and if you do a good job with the education, there is a relatively straightforward way for those customers to understand what you're delivering that's new, that's interesting and compelling. And I was at your user conference recently. And so it was really interesting to sort of see this, almost this evolution.
if you like, especially in, you know, not just in messaging, but in product as well from, you know, the fact that the term RPA pretty much never got used in the three days of the entire event, which was kind of fascinating, especially bearing in mind that last week, Gartner shipped a new RPA Magic Quadrant in a very strange, arbitrary way, but we won't get into that today. It's, I think for me, one of the things that will be interesting is
Outside of your existing customer base, one of your sellers, for instance, or your channel are talking to a net new prospect who is looking for a set of capabilities, what is it that makes them a sweet spot for Automation Anywhere? And what is it that makes, that showcases to them that Automation Anywhere has a differentiated value proposition in what has become a pretty crowded market?
And I don't mean crowded market around RPA. mean a crowded market around the convergence of everything from traditional BPM, low code, RPA, iPads, and everything else. And so it'd be interesting just to try to get a perspective on sort of what makes you different.
Prince Kohli (12:11)
You know, I understand the clouded market past. What has happened is that marketing is converging. I don't know if you have noticed that. Marketing is converging. People are talking about the same terms. know, obviously everyone, if you don't say LLM or foundation model, you don't exist. AI agents is becoming the new fancy term as well. So, and which is unfortunate because for customers, it becomes their problem in trying to understand what is real, what is not real, and what value do they get.
Then there is, and I'm actually answering your question. You will see how I'm coming at it. Then there are other companies which mix personal productivity with enterprise productivity. Correct? And unfortunately, they use the same terms. That's some, I would say unnamed, but companies that provide your usual office suites. Now they will also say, they will say automation as well. They will say agents, and they will also talk about productivity. But to a large enterprise,
Creating PowerPoint faster and doing better grammar checks in email. That's one kind of productivity. You I look, use it. I like it. But where does it help a company with their bottom line? Can you measure that and say, if I want to improve next year, the CFO says I want two extra points in margin. Will you, can you say that if I give $20 per user and add AI to one of my applications, can I improve that two points of margin?
The answer is almost certainly not. So you really have to think of the processes that the company cares about, whether they're customer -facing processes, internal processes. So going back to then the question on how do we position ourselves and why do customers care about what we do, because that is really the key question. We have two things that, at least two things that customers seem to care about. One is we are able to talk in terms of the problem statements they have.
For example, there is a very large hospital in actually the Northeast of the US who have used us to create agents that help, whether it is helping the doctor, helping the doctor do after -visit summaries and then write up case reports or helping them do parts of diagnosis and using many different images and so on that are out there. So there are things that are existing processes that they can make better.
and maybe add new things as well as new technologies come. So for them who, customers understand AI, but what they lack is a platform that allows them to build high fidelity AI agent that allows them to have governance on the data, have governance on the models, understand what the data provenance is, then do ABC experiments because there are so many models out there, then able to map them to the enterprise data.
And at the same time, say, what I'm getting is low hallucination. And when needed, have human in the loop. So to have that end -to -end view on something that is real and something that is already deployed, not something that is promised three years from now or two years from now, that, I would say, has been the reason that customers like us.
Dave Marcus (15:11)
So the, sorry, Francis, go ahead.
Francis Carden (15:14)
No, no, no, go. That's good.
Dave Marcus (15:16)
So I guess, you know, what came out really clear at your user conference was this whole notion of sort of automation anywhere as a vehicle for creating, facilitating, composing, you know, these AI agents, whether you want to call them agentic or not, you know, in a marketing context doesn't really matter. But I also noticed this very heavy focus around orchestration. And so
You know, we're seeing right now that a lot of vendors are sort of not pivoting is probably the wrong word, but making a really big deal about investments that they're putting into orchestration because ultimately at the end of the day, there's a lot of stuff that needs to be brought together. Stuff that either you are creating and owning and need to coordinate or stuff that other people are creating, owning, but someone needs to coordinate in order to make it such that it becomes part of a bigger
experience and to your point, measurable through business outcomes rather than just measurable through simplistic productivity gains.
Prince Kohli (16:18)
So that is correct. I mean, I can only agree with you, right? Now there is in that field is also called observability, right? Orchestration and observability, right? What is going on and what do I do about it? So when you have that as part of the automation platform itself, you're able to go many clicks deeper. For example, if there is a process you're orchestrating and then you want to be able to say where it's a process, where is it taking the most time? And then what do I do about it?
including whether it's a task level or a process level, whatever it may be, to have the insight at all the time, to have insight all the time into how my company is performing, because what is a company? A company is a set of processes, right? Complex processes, processes. What are my people spending time on? Those things become very interesting. And as you said in our user conference as well, the fact that now that there is software in many of these things allows, whether that's a CIO or CFO,
to have a single view on a dashboard about what their teams are doing and how they are doing it, gives them more insight into where the improvement may be in the future, where the investment should be in the future. So those things are becoming very real for most of them.
Francis Carden (17:21)
So, interestingly, I was thinking about a question after my first question and you kind of...
touched on it in your second one, and you were talking about process mining and task mining and all kinds of visibility. you know, and I am passionate in believing that people need to do orchestration regardless of their observability and process intelligence is a fantastic piece of value, but you still need to get the thing orchestrated and you need a workflow engine. You need something to start, because I think that's the future of the way work gets done.
But you touched on this terminology issue. know, it is so noisy out there. Who isn't saying a genetic, who isn't saying generative AI, who isn't saying chat, GVT or models, LLMs, it's like, and not just within the RPA market, but other platforms, bigger platforms, like service and, you know, PEGA and Appian and so on.
Prince Kohli (18:09)
everywhere, everywhere. That's right.
Francis Carden (18:14)
I want to get back to kind of seeing if can just squeeze a little bit more of that answer to you. Where do you, do they compete? Do you compete with ServiceNow? Do you compete with Appian and Pegger and Microsoft Power Automate? Or do you, are you quite happy to compete in this bit given even the analysts typically like to bucket you into an RPA MQ way versus a low code versus a BPM, but it's
Give us your take and please feel free to say what you think you want to be, what AA will be doing next, what's coming to solve this and it's a conundrum for you, I see.
Prince Kohli (18:52)
Thank you for that question. love that question. know, I see there is an interesting view, right? There is an analyst view, then there is what we feel is going on. And then there is a customer view. I, know, customer view is like, you know, the only truth out there, right?
Francis Carden (19:03)
Yes, agreed. Well, remember I was a vendor, I was a vendor for so many years. I've only been an analyst for about a year, so I'm on your side.
Prince Kohli (19:11)
I said this to everyone because analysts are awesome because they talk to so many people and they obviously get insights from many smart people and having been given their own experience, they're able to bring those insights together. But I feel the customer view in the end is really the sole, the only truth that matters. And I think that one thing that the industry, we do a disservice to our customers is that we overuse terms that should not
that are, that provide no value. And you know, I will give you an example of an LLM. My own personal view is an LLM or a foundation model or to say, you know, open AI versus cloud and so on. It's a way to confuse a customer, honestly. Right? You know, you, you keep on saying, I have these 80 LLMs in my kind of in my quiver and you can use whatever arrow you want. That's okay. But you don't want to tell the customer I am better than someone else because they have 60 and I have 80.
I honestly doesn't matter, right? What matters is if you want to do, for example, a document automation as part of your long process, you want to use it in the contact center or you want to use it in healthcare or supply chain. In that, are you able to solve the problem that they have or help them solve the problem that they have using the right tools? Are those tools available is the right way to think about it. Agents are an architectural choice, right? They're being what we call kind of the right LLM, the right rag service.
the right action model because action models are very important. You're able to bring that all together into an entity. And, you know, we call the entity agent, but you could call it whatever you like. What is happening is that, you know, the Microsoft uses, you know, productivity, but really they mean personal productivity. ServiceNow, when they say automation and AI, they're really talking about what, about the perimeter of ServiceNow itself. What works with ServiceNow, right? What gets data in and out of ServiceNow.
And when they say, therefore, we'll automate your environment or other your company, they don't really mean that. What they say is things that were just around us, we will talk to them via whatever automation that we put in. But no real process exists in a company that is one or two apps. Pretty much every server will show you most real processes are across 8, 10, if not 30 apps. I mean, nothing is across two. Everything is across many, many apps.
So to, example, say have ServiceNow say, I can automate your enterprise or to have Salesforce say, I'll automate your enterprise or to have Microsoft say, haha, I'll automate your enterprise. I it's misleading at best and really unfortunate for the customer budget and for the customer expectations.
Francis Carden (21:49)
They're spending a lot of money with those vendors though, aren't they? Growth is amazing. And that whole platform based space, software application development space. It is growing.
Prince Kohli (21:55)
Yes and no, right? Yeah. Sorry to interrupt you, but you know, yes and no. See, because I think sometimes when someone reports AI numbers, they mix developer, you know, what did you use in your GitHub? They'll mix up with AI numbers, right? So you really have to dig into the numbers for saying who is using what. If I sell you a number of licenses, $30 a user, is that the same as consumption?
And then how do you measure the ROI on that? That is a challenge. I actually ran into a customer in Japan, one of the largest companies in Japan. name them, but you know, well -known company. And their CIO asked me this question. He was very honest about it. He said, look, all my vendors, right, many of them that you named, they say each of them wants to charge me X dollars per user. How do I decide? Right? I mean, you are also here selling me this process automation. How do I decide? And my answer was,
Don't take our marketing, right? Don't accept our marketing. Understand where the value is coming from. If you can show whether it's a point of margin or whether it is something running faster or better, if you can measure it and you can put a dollar value on it, you should go with them. I don't care who it is. And that's the right answer. It's always the right answer. If you think you can affect your enterprise by doing parts of ServiceNow and parts of Salesforce, maybe that is the right answer.
In our experience, in our customer experience, the right answer has always been the abstraction, the right abstraction is the process. That's what the company cares about. They don't care about apps. mean, who cares about apps? In cloud computing, example, virtualization, you forget about CPUs and networks and so on. You care about the higher level thing called an entity that can take all of these things together and run your applications. Similarly, in this space, you don't care about one or two apps. You don't care about Excel and so on. You care about getting something done.
That's what they should care about.
Dave Marcus (23:47)
So Prince, with that in mind, besides investing in sort of the, what I'm gonna call a broad range of core platform services, whether they be gen AI driven or not, are you guys also investing sort of further up the business value chain in either, know, formalized business accelerators focusing on particular outcomes, you know, end to end applications, and I'll use that term in their broadest context.
because nothing's just end to end out of the box. But is that an area that you guys are putting investment into as well so that someone basically is just not coming in and saying, you know, I got to start with Automation Anywhere's platform with kind of a blank canvas and can do more, you know, initially with an understanding that they get to do, you know, 60 to 70 % of the work with the assets that you provide. And then the rest is work that they may need to customize for their own purposes. Do you see that being like a, you know,
Big domain for you guys.
Prince Kohli (24:45)
It's huge. I don't have a marketexture slide to show you here, but if we did, at the top of the stack is what we call solution accelerators, really. That's what they are. And they allow you to go to 60 to 80 % of... It's not just a template. mean, let's say you're in service operations, or contact center, or wherever else it may be. You can actually start off by saying, here are all the right agents, here are the LMS I use, here is how most of the work is going to get done.
And then the rest you can fill based on kind of your own custom, you know, the actual applications you use and so on. Because, you know, like you were implying earlier, you want the customer to, you want the user and the customer to get started as quickly as possible so they can see C value as quickly as possible. So we will ship with even pre -trained, you know, good prompts for each LLM, all that, all that stuff. You shouldn't have to start from kind of day one.
because that's a lot of work. It requires a lot of data scientists, it requires a lot of experimentation. So if you're able to ship with most of it and then allow them to tweak, makes it, it takes, you know, it will save them six months, eight months of having to do starting from scratch. We will come with pre -integrated with Rack Services even, right? So you can automatically get low hallucination. We can connect with your enterprise data, data stores, and that value is real to the customer.
Francis Carden (26:09)
So the, now I go back to this terminology statement. think that the talking about value, I think is a really good approach, automation anyway, and focusing on that to the customer. You've got this bit of conflict going on with business and IT because business do want to fix things quickly because in the past it's taken so long to fix it properly.
Right. So that they're quite happy to have automation to get the value because it's either IT has got such a backlog. know they're never going to get it or nobody's told them about some new technology. And I think that's one of the things that, you know, as I go through this semi retirement and I'm retiring on the low code, no code, you know, moving on out of RPA, I see these platforms as so predominant that the, what you've just described some great use cases with AI.
there are some phenomenal use cases of AI across these models because you get the governance and security, you get all that protection, those layers, and then suddenly you come along with AI. It doesn't need to relearn the governance because it's already on the very things it's being applied to. So that's exciting. And so that leads me to this question we've asked on our other podcast. Typically, we ask this question. So you may have heard it anyway. So not going to put you on the spot.
Given specific reference to Microsoft and how much they're investing in low code and automation technologies and the differentiation you see with AA and forgetting all the complexities of all the terminologies will come in emerging and confusing. How do you see this working? You both compete and partner with Microsoft. So are they a threat? Are they more of a partner? And I don't want to put you on the spot with that, but...
where is it going to benefit everybody, what they're doing, and what are you doing in respect to that?
Prince Kohli (28:03)
So, know, said Microsoft is always an interesting company. think I don't know if there has ever been an enterprise software, a serious enterprise software space. By serious, I mean large, where you don't have to compete with Microsoft at some point or the other, because whenever there is enterprise software, I mean, you there's always, they're always there, right? And, and, and sometimes they succeed and sometimes, you know, they fail, but they always try because that is a space they want to always play a part to be a part of.
Francis Carden (28:17)
Always.
Prince Kohli (28:29)
So therefore in that sense, we do compete with Microsoft, but the way that occurs is that Microsoft, they have their bundles and they have their packages and their power automate and so on. But they really focus on personal productivity. So therefore, as long as we can demonstrate to enterprises that it's not about email and Excel micros and PowerPoints. And what you should do therefore is have a real
deployment. You want to automate this long running process, whether it's a supply chain or whether it is manufacturing or whether it is, you know, in healthcare pick any really, you know, back end or front end. Which is why don't you just go and try it, right? We will, we will come in, we'll use our solution accelerators and then a few days to a week or two, we will have it working. That's never going to happen with Microsoft. They just don't have the tools. They don't have the products. They don't have the understanding of that space because they work at a personal
at a person level. So therefore for us, while there is always FUD, because the word use are the same, the terms use are the same, we find that when the rubber meets the road, we are able to show our other customers, show our products, and then any POC in any real deployment, it's easy to show value. In some ways, it is a blessing to compete with Microsoft. You get some free marketing.
But you get the know you get the fad. It's the nature of the nature of the game
Francis Carden (29:54)
like the way you've answered that. think reality is that it's the fox in the hen house. You do have to find out. In my early, early years, and there a of people on this call who don't even know what this is, but I used to compete with Shareware. We had a term in Malaysia, but Shareware was a thing. And of course, Microsoft with all the packages and their bundling as well. But there's some really cool shit going on over there. Let's just say that. And there's a lot of cool shit going on in lots of places. Dave, do you want to follow up on that?
Prince Kohli (30:06)
Yeah, there you go. Yeah, yeah, Yes.
Dave Marcus (30:20)
Yeah, yeah, absolutely. So Princess has been really, really helpful. I want to be conscious of everyone's time as well. It's been super helpful to understand your perspective, you know, how AA is sort of looking at this world. I think for a long time, not a long time, for a while to come, there's still going to continue to be a ton of noise out there and figuring out how to get beyond the noise.
becomes really, really useful. It's partly why we as analysis .tech actually exist in some ways and why we sort of look to establish this, because we wanted to try to help organizations, whether they be vendors or customers, kind of get beyond this noise and get to some level of pragmatic reality. So again, thank you for the time today. For those of you listening into this call,
We'd love to have some feedback. We'd love your comments. If you agree with us, please let us know. Please repost as well. If you don't agree with us, then set up a dialogue, because we are really, really happy to do that and to respond as well. So again, Prince, thank you for the time today. We really appreciate you giving up some of your valuable time. Thanks all, and enjoy the rest of your day.
Francis Carden (31:07)
Definitely.
Prince Kohli (31:09)
Thank
Francis Carden (31:25)
Thanks Prince. Thanks Dave.
Prince Kohli (31:27)
Thank you. Thank you. Bye.