Andreas Welsch: 0:21

Today we'll talk about the trends that will shape intelligent automation and AI in 2023. And who better to talk to about it than three experts in this space? Shail Khiyara, Frank Casale, and Ian Barkin. Hey everyone. Thanks for joining.

Shail Khiyara: 0:36

Hey Andreas, thanks for having us here.

Frank Casale: 0:38

Happy to be here.

Andreas Welsch: 0:40

That's awesome. Hey, I'm sure most people in the audience might already know who you are. But why don't you introduce yourself real quick.

Shail Khiyara: 0:47

Sure. My name is Shail Khiyara. I've been in the automation space for the last decade, have had the opportunity to help shape this market and work for three leading automation providers: Automation Anywhere, Blue Prism, and UIPath. I now run an independent customer think tank called VOCAL which, stands for Voice of Customer in the Automation Landscape. And I've also joined and served on the board of Turbotic and joined Turbotic as a President and COO. Over to you, Frank.

Frank Casale: 1:18

Super. Great to be here. Appreciate the invite Frank Casale here, coming to you from New York. I guess most notably founder of Institute for Robotic Process Automation and ai. That's why I know most of the people in this space prior to that founder of the Outsourcing Institute. Had a good run, maybe close to two decades on in the labor arbitrage space. In the last decade, more in the kind of automated labor and related. So looking forward to today's discussion.

Andreas Welsch: 1:46

Fantastic. Great to have you with us and over to Ian.

Ian Barkin: 1:50

Thanks for having us. Yep. Ian Barkin. I have spent a decade in the RPA intelligent automation space. Learned about it actually first and foremost from a conference that Frank threw way back in like 2013. Co-founded Symphony Ventures, which is an RPA consultancy and advisory, which I then sold a few years ago and spend my time now creating LinkedIn Learning courses, generally trying to develop useful and educational content and investing and advising startups and disruptors in the world of work. Thanks for having us here.

Andreas Welsch: 2:24

Fantastic. So I'm really looking forward to our conversation today and if you're just joining the stream, drop a comment in the chat where you are joining us from. I'm really curious to see how global our audience is because we have more than 930 people that said: Hey I'm interested in this show. I want to watch it. Awesome. Two more quick things to share with you before we really get started. Don't forget to sign up for my newsletter at intelligencebriefing.com and stay up to date how you can run AI and automation projects successfully. Now, we have plenty of time for your questions as well today, so please don't be shy and put them in the chat as well. So let's finally jump right in. First question for you, Frank. I, know a lot of organizations have implemented RPA over the last couple of years and many automation leaders are now wondering what's actually next? So how can they build on their existing program or initiative in 2023? And what would you say are the major trends they should look for?

Frank Casale: 3:24

I would say if you've turned on your laptop at all in the past 30 days, we all know what's next. It's ChatGPT, but we'll build up to that. We'll build up to that crescendo. From, where I sit, Andreas the majority of the people that have invested in RPA over the past decade, are typically looking for what I would say ROI amplifiers. And it typically takes, I would say about 12 months before an enterprise user realizes that they need some adjacent technology. In the bulk of what I'm seeing in our space are things like IDP and conversational AI as two kind of big plays that are great bolt-ons. It's not a rip and replace, either one of them are overlays. And the majority of RPA users that I see that are getting real return on investment from the RPA investment is not directly through RPAs, a standalone. It's really more through these adjacent technologies that suddenly begin to have much greater impact. So I'll stop there. Not to take up too much time. But that's what we're seeing. A lot of activity in the IDP space. Conversational AI, conversational business apps. I think that's where the action is right now.

Andreas Welsch: 4:42

Fantastic. Thanks. Anybody else wanna build on that?

Ian Barkin: 4:46

I would add the process stuff, process mining, process discovery, task mining, all of those the gateways into helping enterprises identify where to apply all the rest of the stuff. Hopefully with a healthy focus on fixing and simplifying processes before just throwing tech at it.

Andreas Welsch: 5:05

How, much do you see leaders and companies a actually do this? We love to think about it as this ideal typical stage of, Hey, first you analyze, then you figure out what you can automate, and then you build something top. How do you see this happen today? Is this really how you see it can happen?

Frank Casale: 5:24

I guess it's tough to go broad brush strokes and I think my colleagues would agree. There's a very small and active portion, I would say, of the community that continues to invest and push the envelope that's looking for, I would say leapfrog innovation. And those are very much screening in opportunities and with good, top-down support and investment. I would say both in dollars and I would say in mindset if you will. But on the other end of the spectrum we're seeing a sizable percentage of people that are a bit disenamored with with RPA. So they're just sitting there wondering what and maybe what did they do wrong? Who explained what incorrectly or over-hyped what on the buy or sell or advisory side, and everybody's pointing to the other person. And so those guys are going nowhere. But I'm seeing kind of two extremes.

Shail Khiyara: 6:21

I tend to agree with with you, Frank. I think the"just do it" automation era has to a large extent slowed down and we're starting to see a lot more emphasis on value. Frank mentioned ROI earlier. It's ROI and it's what kind of customer experience are you driving? What kind of revenue enhancement are you driving through automation? So automation value is becoming more and more prominent in the decision making for automation technologies. And there are several of them.

Ian Barkin: 6:54

Yeah. I would concur. I know Andreas, you want some more drama, but I I would concur with what we just said. I think my biggest disappointment, the RPA space specifically, has meant a lot to me. I built a team there, succeeded in that space. I think globally we have absolutely failed at achieving the full potential and impact of that one technology. And I think we're on the path to fail at applying all of the other augmenting technologies that Frank just mentioned because enterprises are going about it in the wrong way, right? They're fascinated by the tech. They are not putting in the foundation stones of solid enterprise change with a vision and a mission and communications and buy-in and all that other stuff. And so it's frankly a little disappointing. I think the potential is huge. The actual achieved outcomes thus far are frankly pathetic. And it's up to folks like us on this call to try to educate everybody into a pragmatic, methodical, mature set of activities around understanding and adapting and applying all of the tech available today.

Andreas Welsch: 8:09

Thank you for sharing that and for building on each other's answer. So I'm having a quick look at the chat here. Boy, that's really global. For those of you who have joined from. Lebanon to Morocco, to the U.K., Austria, Italy, the U.S. So that's really awesome. Thank you for, joining. Let's, take one question from the chat. Maya is asking what are the major misconceptions companies had about RPA and where was the disconnect?

Ian Barkin: 8:42

This is so easy a business analyst can do it" was the narrative that was dumped into the market for years. And when you tell people it's easy, they prepare for easy voyages which means that they under prepare for the real voyage ahead. And therefore they didn't educate broadly enough. They didn't allocate enough funds, they didn't put together the right SLAs and objective measurements to decide what good is early stage, mid and later stage in their maturity and scalability. And so I think you end up with a bunch of abandoned POCs and pilots that prove the point, but never really embedded in a habit.

Frank Casale: 9:24

Yeah, if I could pile on. As I saw it, and I've been involved almost since the beginning, I think there were things that people should have known and I would also say there were things that none of us knew, but we learned over time. What we should have known is, at least, I felt was fairly basic, is when somebody sets out to quote unquote increased productivity. To this day I hear, and my colleagues may agree or disagree, but if, I'm pulled into a discussion and the mission is increased productivity and I can't get three people in that organization agree on what that means, I know that the business case is dead. So, the more specific and the more definable, the more measurable an outcome is, then you could look and say, okay, fine, does this technology get us there? And I would say in many cases RPA may not. Cuz at the end of the day it is non-intelligent automation of a task in many cases a subtask. So how far does that get if we're able to free up 10% of Ian's time each day by some nifty tool, cuz we're automating his travel expenses reporting. And how do we know? How does he know? How do we know if that increases his productivity? We'll never be able to connect the dots as I see it. I don't think I'm being a pessimist. I think I'm just being a pragmatist. What most of us didn't know going into this is even if you ran the numbers and it looked like it worked, most of us didn't anticipate the cost on the back end to manage and repair these bots. I don't know, maybe Ian and Shail, maybe you guys sensed that early on. That certain was a surprise to me and many folks that I knew that you say. Wow. And, I would argue maybe not too dissimilar from those of us that came from the outsourcing space where we thought that in the early days in the nineties where we were doing outsourcing deals and I was involved with a handful of really big ones where the assumption was, okay, we ran the math, ink the deal. Let's go, have a steak dinner tonight and we're looking good. And suddenly we say, wow, we're gonna need a team of 3, 4, 6 people in maybe two, three different cities to manage this thing. And then suddenly that kind of the numbers got funky from that point forward. Yeah.

Andreas Welsch: 11:41

Perfect. Maybe let's move on to the next question. Thank you both for sharing your, perspectives. So if I put my old IT hat back on for a second and it of goes in the direction also what Ian and Frank have said. I think there are a lot of landscapes that have evolved over time, right? Maybe you've integrated an acquisition or onboarded another division that's on different platform and you inevitably get to a multi-vendor operation. From an IT point of view I know that's never really ideal, because you're trying to standardize much as you can. So I'm wondering is this multi-vendor automation landscape really the new reality? And how can automation leaders optimize their footprint?

Shail Khiyara: 12:25

No, that's a good question. As you heard Frank and Ian talk about the fact that there are multiple technologies in automation out there, right? There's IDP, process mining, machine learning capabilities, cognitive X, Y, Z, and analytics in addition to the RPA technology that we're talking about. So what I'm hearing from customers is that this multi-vendor landscape and the desire to have multiple technologies to do various different things under the umbrella of enterprise automation is inevitable. There are about 40 to 45,000 customers using RPA across the globe. Today, 40 to 45% of them are using multi vendors. And that is continuing to grow as we see commoditization, particularly of the attended automation capabilities with some of the new large entrants that have entered into this space. What's really interesting is that over the last, I would say three years or so, about 11 billion of investment has gone into RPA. A lot of it has gone into the three major players and then the long, tail of 65 other players that exist in the market. However, if you look at the acquisitions over the last two years, 14 acquisitions totalling about one and a half billion dollars. So 10 billion gone into the core RPA vendors, one and a half billion dollars gone into acquiring 14 companies for large entrants like Microsoft, IBM, SAP, and various others to enter into this market, starting to commoditize attended automation. So with that kind of a dynamic, there are a collection of technologies customers are using. They're struggling with how to stitch these technologies together because the core fundamental focus of RPA vendors has always been to sell bots. The orchestrators are lacking innovation, in my opinion. And because of that gap, customers are struggling with, if I have multiple vendors, how do I manage these vendors? Where's the single pane of glass to be able to manage the entire lifecycle of automation? Andreas, if you walk into a CFO's office today, she cannot tell you that she can't give you financial information because she doesn't have a system for it. She does. If you walk into a CRO's office, he can't tell you that he won't give you sales forecast. He has a system for it. It's a CRM system. If you walk into a CoE today, not all of them, but a few of them, you do find Excel spreadsheets, SharePoints PowerPoints, SMS messages to manage the entire lifecycle of automation to manage all these multitude of technologies that exist today. And that is a significant challenge. And we're starting to see the emergence like we saw in the network world, like we saw in the cloud world. Where we saw network optimization tools, cloud optimization tools emerge. We're starting to see the emergence of these single pane of glass automation optimization tools. There's still a long way to go in this market, but I think coming back to your question. Yes, I think the multi-vendor landscape is inevitable. Ian, Frank what are your thoughts.

Frank Casale: 15:41

Yeah, I'm happy to weigh in on this. My feeling is if you really want to de-risk it and have one neck to choke, then outsource it. It's pretty straightforward as I see it. However, realize there's no risk free path, so you outsource it. You have predictability there, but you're giving a margin and you're giving up potentially best in class solution. But that's the way to go. On the other hand, if you want best in class and you have a comfort level that your IT team has some experience cuz you really want to get beyond them learning this on your dime, if you will. If you're the internal client and, I would say the majority of global 2000 have some experience here, then you go best in class and maybe bring in ideally the third party just for success insurance and build something awesome. But even then now you have integration, creates uncertainty. And I would assure you, whatever you like this year, you will not like next year. It is a bit of a dilemma. And then again, looking where we are now, and I promise I'm not getting a commission, that each time I say ChatGPT, somebody sends me$50. If you look at where we are now, we're on the precipice of the whole game changing so, we just need to be aware of it. It's still young and it's not enterprise ready. It's very cool. It's very interesting. I believe it's likely to change a whole bunch of things. That's my understatement, but nonetheless I don't see an enterprise ready. One last thing I'd like to say. When we think about the buyers, obviously there's a lot of different ways to define large, medium and, small, et cetera. But what I've seen, and I think many of you have seen is, a significant difference in traditional companies versus digital native organizations of how they do this. So I would caution against the big kind of macro view of the market because the way I would approach a strategy at JP Morgan and the way that would be or not be embraced would be totally different the way I would approach it at Spotify, a place that was born in the cloud, digitally native. In the latter scenario, which is small, digital, native, non-regulated industry, that's gonna be a shorter cycle, less landmines. We're just gonna be moving and grooving. It's gonna be a lot faster. You get involved with a big bank or an insurance company, the ultimate impact could be significant, but it will be painful. I know I have battle scars doing those deals, and I know a couple my colleagues do.

Andreas Welsch: 18:29

Awesome. Anybody else wanna add Ian?

Ian Barkin: 18:32

Yeah. This dates us all, I think. But, that conference I mentioned that Frank held, that introduced me to RPA was an outsourcing institute conference, not an RPA conference. And RPA came out of the outsourcing industry because we were using whatever capabilities, tools, resources we had available to us to make an outcome for our client. In that case it was transactional processing at a more affordable rate. We happened to use arbitrage, so we went offshore and used people in places like India. Then RPA came along and it it matched that to some degree cuz it enables us to continue to do that process transaction activity. But with a different agent. It wasn't the Indian, it was a robot. So we said it would happen that way, but then we got too fascinated by the tech itself and enterprises started to be told that they could do it, so that you don't need a third party partner who happens to be an expert in this. You can just fire up your HR team and ask them to spend their evenings and weekends digitizing the tasks they worked on that did not work. So to Frank's point working with an expert I think will more likely be the path to success here, but it rests on all of us to figure out a little bit more modern, interesting collaborative business models and collaborative structures so that yeah, you're gonna pay somebody. So there will be some margin, but I'll tell you that margin will be a function of all of the pain, technical debt and waste that's happened in enterprises in the last 10 years, trying to figure out and deploy the RPA, IDP, process mining, conversational AI, et cetera, that they're tinkering with right now. So I think the future is some sort of hybrid where you're working with experts, still employing the capabilities in the subject matter experts in. Maybe to scout and find good areas to deploy this stuff. But that's how we're gonna really make the most of the tools available to us. And the need, the big enterprise is not the digital native ones that can just start from scratch and be, more be more structured and, digital from the beginning. That's, the way we all need to go.

Andreas Welsch: 20:42

Fantastic. Thanks. I think well rounded answer and good to see the different perspectives here from you. So if I take a look at the chat, there are some questions about AI. For example, Michael is asking: Will ChatGPT and its equal, but opposite generative AI detection, become a new market category, if not in 23, then in 25? Or Christine is asking: Do you believe that more attention towards data monetization in individual organizations will be the trend that will shape AI and automation in 2023? I think personally I have an opinion on the generative AI topic and I believe at least for the time being that we're, still as an industry, as leaders, figuring out where can we actually use this and where can we use it in a way that it makes sense. If, we on one hand have proprietary or models. On the other hand, we have open source models. If we use the ones that are proprietary like CHatGPT, we don't really know what data has gone into it. And you see some discussion and concerns, for example, over at Amazon. Hey, this thing spits out code that looks suspiciously familiar to our own IP. Or questions around information security more broadly. Please don't copy confidential information into this chat prompt, or are the answers that it generates really reliable, or do they just sound plausible? So I think there's still some things that the industry needs to figure out. And, similar I think also to, RPA and, other AI scenarios. We might want to have the human in the loop for some time to review it. It's good, at least at the moment, right? In the scenarios we see, and at the moment it's good as a kickstart for ideas and to get creativity flowing. Maybe to reword or rephrase a few things. And, I'm sure you know this. We'll, see a thousand more ways to use this. I don't think its own category. I think much AI it'll help a new kind of application, be powered and, deliver more. And I think also on the other side, organizations looking into monetization, I think there are a few that they're doing this or will be doing this and will be doing it well. But also depends on the underlying architecture. Do you have the data strategy, the processes, things in place to be able to do that. And maybe part of that goes back to digital natives and established companies as well who might be able to do it easier if you're starting new in the cloud, relatively new compared to having legacy and history. But I'm curious what do you think? What are you seeing there? Maybe just briefly. We have a few more minutes before we wrap up and wanna get to Ian. And, the question I have for you as well.

Frank Casale: 23:32

Okay. May I, in a friendly way, disagree with you with opponents?

Andreas Welsch: 23:37

Sure. I'm more optimistic than you seem to be, Andreas with regard to ChatGPT. And I think it'll be significant, I think from a standpoint of category. I recommend that we all think about less about what it is and we focus more about what it does, which you alluded to. I think it's all about use case. And I think there are people on this line now that have ideas for use cases, and I would say the right use case with the right commercial strategy is your fast track to independence. I think there'll be billions of dollars made. Some people will make millions, some people will make billions for anyone. And the barrier to entry is very low. Which is it? Okay. Double-edged sword. But still the, opportunity for someone now with an idea, getting a whole of technology that is cheaper than cheap. Why? Because it's free it is significant. So I've been watching this very closely. I've been fortunate enough, maybe the older guy on the panel here, I've been through a couple of cycles. This is significant. However yes, there will be a ton of lawsuits. I think IP becomes debated. I think it becomes very, very fuzzy whether you're in the insurance game or a college student doing a paper. I think that becomes very fuzzy. However literally this weekend you could launch a business and go. There are tremendous opportunities. One quick thing, if I may, on data monetization. From what I've seen, the discussions have been amazing. I've seen very amazing stories. I think for the average company, if you're a bank or an insurance company or in telecommunications or you're a media company, the odds of you investing the time, by the way, you may have a billion dollars worth of IP data and yes, data is the new oil, blah, blah, blah. I haven't seen it. I haven't seen it yet. I think it's likely gonna be an independent organization.

Shail Khiyara: 25:39

Look I'm very bullish on what you just said, Frank. Very bullish on generative AI as a whole. I wouldn't say just focusing on ChatGPT, but if you look at the marginal cost of energy, it's approaching zero. That's going to happen to the intelligence as well. Marginal cost of intelligence approaches zero, and it's happening at an extremely fast pace. If you look at Transformers, when you look at BERT back in 2017 to ChatGPT now in less than slightly more than five years. The next five years, you're going to see transformers even evolve even further. Where prompts or questions are not going to be the key focus areas anymore. It's going to be the conversation that you and I are having right now that is going to be prevalent in terms of how you interface with these systems. So I do think the opportunity is immense. I do think that there is significantly fast-paced technological innovation that is going on. To Frank's earlier point, what you may like today, you may not like tomorrow. Same applies to ChatGPT as well. You may like it today, but there's something new coming tomorrow very quickly. That's one aspect of it. The other aspect of it is I think it's going to change the structure of interfaces in society as we know it in many ways. Frank talked about lawsuits. There's already discussions about how to use this or not to use this in schools. There are already discussions about how, who owns digital art as an example. So these are new questions that are coming forth that we have not dealt with before as a society. So that's going to be a very interesting change across the world. And by the way, ChatGPT, right now, the interface is English. Wait till that changes.

Andreas Welsch: 27:21

Excellent point. Maybe Ian I'll go over to you for the next question that we have, so we can finish somewhat in time. But I really appreciate the good conversation and discussion we are having here. Also the questions in the chat go in that direction if there are these huge shifts and this new momentum for automation, for AI. I know you often post about this and call it the workers of the future in the future of work. So I'm, wondering what does it mean? What do you mean by that and what does it take to be a future worker?

Ian Barkin: 27:55

Sure. Ultimately the divide of IT and business has held back true evolution just cuz business knows what they need and IT knows what they can build and what they have. And so the last 10 years we've aspired to bring those groups together or at least empower business to run a gray rogue IT operation where they can get done what they think they need to get done in the future. Because every company is a software company. We all need to evolve our skills, and enterprises need to realize that. Cuz as we've just discussed, there'll always be a new tool. There always will be. Ultimately we need to build a flexibility and a foundational digital literacy and digital quotient within enterprises so that we can assess incorporate and apply whatever comes next. If it's ChatGPT or if it's RPA or whatever else. And, then we need to be able to adapt the organization design really good incentive structures and initiatives to make the most of it. Because right now, in RPA, we are hoping business could do it. It's a little hard. We got it doing it, but we don't have enough IT folks. We don't have enough developers in the world to build all of the apps and the configurations that businesses believe they need. And so you're leaning on your citizenry to do that. And yet the citizens aren't necessarily. Skilled up yet to be able to do that. So most of the citizen development initiatives that are happening in the world are pilots. They're small in scale and impact. And so over time, foundation of digital literacy, we need to create a better organizational structure with incentives because the smart people, the ones who actually know how to do this, are using it to digitize their jobs and then just go play video games for half a week. So those are the ones who have the moonlighting and the gigs and the side hustles. So they don't have any incentive to, to share what they've, what their ingenuity and their ability have allowed them to develop for their enterprise, cuz what's in it for them. There are a lot of pieces that need to be worked out to truly harness the ingenuity of everyone within an organization. Make them capable of adopting the tools that come out as they come out and then to have these meaningful discussions about how we're gonna experiment and improve out their value and apply them in a broader scale.

Andreas Welsch: 30:20

Perfect. Thank you. That's awesome. I think that's a really good way to also end today's session on. But before we do, I was wondering if you can each summarize one key takeaway for our audience today. Starting with Frank and to Shail and to Ian again.

Frank Casale: 30:39

Wow, that's tough. I believe I coined a term about a decade ago digital labor and I for years kept thinking that we're almost there. RPA? Nope. Intelligent automation. Close. Maybe full on AI. Not sure. I believe we're approaching that tipping point where intelligent technology will shift from assisting knowledge workers to emulating knowledge workers. Now that's exciting and also fairly scary. So there will be people that would be able to find opportunity here. But regardless of who you are, buyer, seller, advisor, entrepreneur, wanna be entrepreneur I would recommend that you think about the following and I jotted them down because I think it's important for each of us as people, workers, professionals, and you have a life, you have a family, I would say. To power through the change. Most relationships win. Biggest network helps a mindset toward continuous learning. You can't, if you get stuck with the old dog and the new tricks, you're stuck. And last, by not least, financial runway could be a good amount of turbulence depending on who you are. I think at that point you're, you have the best odds of powering.

Andreas Welsch: 32:01

Fantastic. Thanks. Going over to Shail.

Shail Khiyara: 32:04

Yeah, look, there are two things that make up organization. It's people and everything else, right? So I would say that it the same applies to automation as well. Automation, it's not the bots driving your automation. It's actually the people involved in it and the ecosystem around it. So my key takeaway is that focus on the software aspects of automation, which is the culture. How do you create a borderless organization? How do you create a no fear zone? How do you actually empower people to drive automation inside your organizations? The last thing I wanna say is that look I just want to express my gratitude sitting here listening to industry experts who. In many ways helped educate the market, Frank through IRPAA and other means, Ian, through LinkedIn Learning and, other means as well. And Andreas, what you're doing as well. So I'd encourage the audience as a takeaways, follow them. There's a lot to be learned here.

Andreas Welsch: 33:02

Perfect. Thank you. And very nicely put. Hard to disagree with that one. So over to Ian.

Ian Barkin: 33:09

And hard to follow those two guys. Man, what the hell am I left with saying? I agree with everything that Frank and Shail just said. It is about people. It's about that commitment to digital literacy. It's a more exciting future now than it was 10 years ago when I started playing with this stuff. It's certainly more confusing but it's just because we've got more tools and more players and more enthusiasm in the space. So, you really owe it to yourself to stay plugged in and educate. As the guys just said before. The group on this call really has committed and dedicated a lot of their energy to doing that. To exploring for their own understanding and then sharing for the understanding and the development and betterment of our communities. Follow Shail, follow Frank, follow Andreas. Those guys are putting out incredible content. And thanks for having us, Andreas. This is a great discussion. Really appreciate you putting this together.

Frank Casale: 34:07

Thanks, Andreas.

Andreas Welsch: 34:08

Yeah, so thanks for joining us. When we started this, we started on RPA. I wasn't quite sure how far into into AI we would get. And here we were talking about generative AI and ChatGPT. So I think, we really covered all the major trends and really appreciate you sharing your expertise with us and the community here as well. Yeah, so I think the only thing left to do then is to close us out. Thanks Shail, Frank and Ian for joining.