News & Events
May 27, 2022

Extracting Maximum Value from Unstructured Data with Ian Barkin

Share on TwitterShare on Twitter
Share on FacebookShare on Facebook
Share on LinkedinShare on Linkedin
Aliquam auctor diam.
Privacy Policy
Watch Now
Sign up
Watch Now

A recent poll from Ian Barkin, a seasoned leader and innovator in the automation space, found:

  • 49% of respondents say unstructured data is abundant, but their company can’t leverage it.
  • Another 11% say they can’t find OR extract value from their unstructured data.

In short, for most businesses, unstructured data is a mess.

Join Brad Cordova, CEO at super.AI; Manish Rai, VP of Marketing at super.AI; and Ian Barkin, Intelligent Automation Expert, in a discussion about how to get the most out of unstructured data. Topics include:

  • The importance of unstructured data to the future of work.
  • Strategies for identifying high-value uses for unstructured information.
  • What to look for in unstructured data processing solutions.
  • How to successfully find, analyze, and leverage unstructured data.

Full transcript:

Ian Barkin [00:02:49] And there we go. Hello, everybody. Sorry, I was I was I was slow. The video in 30 seconds is very short, apparently. Welcome, everybody who has joined the feed. If you have just joined us, please do share your location in the chat. We would love to see who's joining us. Looks like we've already got folks from Phillip, Idaho. Good morning. Paul Griffith is with us. Magdalena is from Poland, is joined us. Hello, Magda Partha a LinkedIn user, has joined us from Seattle. That's that's very mysterious. And and Rupert has joined us as well. Good day, everybody who has been able to join in. Please do share, like I said, your locations and your questions as they occur to you, because we are very excited to to take your questions and it'll make this discussion more dynamic. So let's get on with the show. We are here to discuss data, but not just any data of the title of the show suggests. We're talking about extracting maximum value from unstructured data. Unstructured data, maximum value. I am the one who's talking at the moment. My name is Ian Barkan. I am an entrepreneur. I've put courses out in books and other stuff about intelligent automation and the future of work. And I'm thrilled to be joined by two absolute experts on data and general unstructured data, specifically Brad Cordova, who is the CEO, founder of Super AI, and Maneesh Ry, who's the VP of Marketing of Super AI. And before we go on any further, I'll ask you guys to give a brief introduction of yourselves. So so Maneesh, let's start with you. Who are you, sir? And welcome to the welcome to the broadcast.  

Manish Rai [00:04:37] In good to be on this show with you again. I remember a few years ago when I was that automation anywhere we were took a plane ride together I believe to one of the events. Was it Orlando somewhere?  

Ian Barkin [00:04:50] It's always Orlando.  

Manish Rai [00:04:52] For each other over there. And it's been quite a journey. So myself, I'm Munisteri, I'm VP of Marketing at Superdry. I've been in Silicon Valley for close to 20 years and squarely in the automation arena last five years. Living the unstructured data journey at automation anywhere among I was the VP of Product Marketing and got to launch the IDP Product IQ bot and work with my colleagues to help create that category. I spent a year at RPA and I'm super excited and no pun intended to be here at Superdry partnering with Brad to solve not just documents, but the challenge of processing any kind of unstructured data, which ends up being, from all measures, more than 80% of the data in any given enterprise.  

Ian Barkin [00:05:46] Outstanding. Thank you. And best for last, Brad. Welcome. And. And who are you, sir?  

Brad Cordova [00:05:54] Everyone. It's nice to be here. So I am Brad, the CEO, founder of Super High. My quick background is I was doing my Ph.D. at MIT trying to merge symbolic AI and machine learning together. During my Ph.D., I founded a company called True Motion. We built the AI back end for a lot of the biggest mobility insurance companies around the world. As a part of that. We built one of the first at scale AI infrastructures, which could collect data, clean data, label data, structure data. And a lot of people asked us, Hey, can we could use this A.I. platform? Unfortunately, the answer was no at the time, but that essentially was the genesis of Super Easy to solve a broad range of unstructured data problems using. I'm excited to be here and talking about unstructured data processing.  

Ian Barkin [00:06:49] Outstanding. It is very cool. Hold on. Let's pull. Let's bring us all back. Hello, everybody. First of all, I love the I've got to love the MIT reference and I love the idea that you effectively you'd solve the problem for yourself. Everyone else saw was valuable and wanted it. And I mean, what better market validation? And then, man, when you said living the data life, I was pretty certain you were going to say living la vida loca. I wish you do it. And that would have been funny. So okay. So yes, in a minute you would share that. We did share flights. We shared an Uber as well. Yeah. So we just really confirmed for everybody how close you and I are. But what we then did share more recently, which is where I got the chance to meet Brad, was we were all in New Orleans at the Intelligence Automation Winter Congress, which sounds sort of dark and ominous for some reason. But it was winter in New Orleans, so it was actually nice out. And I went there because I was just listening. I just wanted to see where the state of the market was, what the topics that people were discussing were, whether it had evolved from RPA to broader intelligent automation, whether folks were focusing on citizen development more. And one thing that was clear, absolutely clear was data was on everyone's lips. Everyone was trying to to figure out how best to leverage the assets that they have or tap into assets that they sort of knew were in there somewhere. But they just had struggled to get to. So so let's let's let's start with data here, which is I mean, data obviously is somewhat clear to folks, but this concept of unstructured data was really the turning point of of where everyone was trying to figure out how to get value from that unstructured data. It's the topic of this discussion, perhaps, Brad, could you define for for us and our listeners what you think of when when you say unstructured data?  

Brad Cordova [00:08:53] Yeah, definitely so. I'd like to start by saying unstructured data is not so well, unstructured data is not as structured data, and I think we all know what structured data is. I think the simplest way to think about it is structured data is data that you can fit into your Excel spreadsheet or your Google spreadsheet has rows and columns. Machines can understand this. You can use RPA on it. Macros software on this. So what is unstructured data? So unstructured data is data that's unstructured, obviously. And so these are things like images, audio, video, unstructured document, text, satellite data. And one of the hallmark. Properties of unstructured data is that it's useless. It's unusable we call unusable data until you provide structure for it. So if you have an image, you take it for granted because you're your your brain. When the image hits your retina, it goes through a lot of processing before it gets structures in your brain. Or when you hear a sound, it goes to your cochlea and essentially becomes structured to you. Even in even in our bodies, we we have to structure unstructured data. And so one of the most important things is that if you want to use unstructured data, it's so valuable, but you need to structure it first. And this is the whole idea behind unstructured data processing. And then I think the second most interesting thing before I passed the miniatures, if you look at the data, the amount of unstructured data looks like a classic exponential curve because everyone's taking pictures on their phone, videos on their phone, so many Iot devices and the amount of structured data looks like a flat line when you plot it on the same graph. And so I think this was something interesting that I didn't maybe intuitively realize, but when I saw the data, it really blew my mind.  

Ian Barkin [00:10:52] Outstanding. Water. Water everywhere. But not a drop to drink. Enterprises have it all over the place but can't tap it. So it's it's as good as salt water in that sense. I mean, it's in that analogy.  

Manish Rai [00:11:03] That if people need more desalination plants to make the water usable, you know.  

Ian Barkin [00:11:10] There we go. That's that's an entire marketing angle. You should experiment with many of the desalinate in the world today to Maneesh and Brad mentioned everything to to the cochlea and brain function. Do you have anything to top that?  

Manish Rai [00:11:27] No, I was just amazed with the analogies and examples he brings brings up in relating to that. I mean, for those of you who don't know, Brad has almost like a photographic memory and a good example is like his wife is German and he took it upon himself to learn German and it took him only six months to start speaking fluent German.  

Ian Barkin [00:11:54] And I don't know if it helps any of us to just be super jealous of one of the speakers on this call. And I don't even know.  

Manish Rai [00:12:00] How to search for unstructured data. It's like, little more advanced than mine.  

Ian Barkin [00:12:05] Okay, well, now I resent Brad. Thanks so much for that. So. Okay, damn it. I have to write it down in a little piece of paper in my pocket and I forget where they are. So, so. Okay, so data, everyone is talking about data. Unstructured data is this massive resource that in some cases is very hard to tap. So what we thought we would do was ask our collective communities in through the use of LinkedIn polls because everyone loves a LinkedIn poll. And we thought we'd ask folks what they sort of what state they thought enterprises were in, in using or accessing or even just being aware of this data. And so what we did and what I bring, I bring graphics, but I'm just going to slap on top of our faces. But so what we did in a LinkedIn polls, we asked of our companies, do people think and we sort of specifically chose our companies rather than is your company? Because we sort of wanted people to speak more broadly to the world that they were privy to. But our company was able to extract data from or value sorry, value from their unstructured data. And before we show you the results, I'll show you the options that we gave folks. And then I want to chat among ourselves about where where we think, where our gut says the answer should be. So the options we gave were sort of the classic. Yes, sort of sort of know of yes, the oil is flowing. We made a data is oil analogy. So, yes, the oil is flowing. Enterprises can tap into this unstructured stuff and and glean value from it. Then there was the sort of it's somewhat tapped, but it's at this point unusable. There's the more definitive, no, we can't find this stuff or we can't use it. And then there's the huh option of I'm not quite clear what you're saying and, and I could benefit from some some clarity, so I'll leave that up for a second. Maneesh, where did you think the votes were going to come in? What was going to be the winner in your head?  

Manish Rai [00:14:03] In my mind, sort of was going to be a winner, but I thought yes, would have been less than 20%, frankly, more around 15% or so.  

Ian Barkin [00:14:15] Gotcha, Brad.  

Brad Cordova [00:14:18] So if I'm honest, I had no idea just because this is it's asking for kind of a subjective value assessment. If it was like, hey, do you have air models that given this 30% ROI, then I do have a very clear answer. But if you asked, are you fast, maybe some, some three or my son would say he's the fastest person in the. What is it like on an objective scale? Usain Bolt is much faster. So I honestly didn't know.  

Ian Barkin [00:14:46] That is a very data oriented answer. So no. And I guess my gut was. And I'm sort of still some of the some of the punch line here. But our community is is one that's in this fishbowl and hears about this stuff a lot. So generally they'll they'll vote more optimistically than I think the state of of the world actually is. So but this was like. So I think I think more people think they've got their act together. And in fact, most enterprises don't have a clue or don't or don't have a hope of tapping into this stuff quite yet. So drum roll, please. The the responses came in and the answer was a sort of a resounding sort of. 50% of folks thought that they'd sort of gotten their handle on this unstructured data they had. In some cases, they they tapped it, but they were having trouble using it. And then they they frankly, shockingly large person 38 said that, yeah, this stuff's flowing like we got our we got our game together. We're tapping into the unstructured and we're getting value from it. Thoughts.  

Manish Rai [00:15:52] I was really shocked because, you know, a survey from Esso and from the community we talked to came out. And when you look at the data, like how many companies have more than 20 boards, what vintage of respondents, and it was less than 19%.  

Ian Barkin [00:16:11] Right.  

Manish Rai [00:16:11] And similarly, about 18% responded that they've moved beyond and they're looking at digitizing data and looking at document processing. So about 9% each, around 18%. So I was hoping to see a number in that range. So so I was really shocked to see such a high number.  

Ian Barkin [00:16:34] Well, and I guess and nobody's going to weigh in and say, I have no idea what I'm talking about or what you're talking about. So the two or 3% of of unsure have the ha I guess wasn't necessarily surprising to me. But again, I do think it was it's optimistic. Yeah. Now that said.  

Manish Rai [00:16:52] We came back from Cisco and we think the annual gathering of all the automakers in the world. And frankly, customer prospect after prospect, they were talking to us about documents because they're just very early on in processing documents are struggling. So that also prompted us to think that, you know, they didn't look like customers who had the oil or the data flowing be.  

Ian Barkin [00:17:23] Good to I think the technical term is had their shit together but yeah so so they were not that so okay. So let's let's pivot. We were we're halfway through our time. I want to start picking your your incredible brands as to what good looks like and how one gets too good. So let's, let's just assume for the sake of discussion that 38% really do, in fact, have their act together and that they really are tapped into this. What are the hallmarks of that sort of enterprise? What what is that enterprise doing right to to really be tapping into this unstructured data, converting it into value and using it? And Maneesh, I'll start with you.  

Manish Rai [00:18:00] Yeah. In my mind, I think if I look at the automation journey of most enterprises, they're starting with structured data. So they have to have a fairly mature RPA in place and they have sufficient number of processes automated so that they're ready to move on to the next phase and start tackling the unstructured part. But 80% of the remaining. So that's one hallmark. They have somewhat of a RPA maturity and two, they are driven by auto AIS. So they're looking at it. What is what are the types of unstructured data that are causing bottlenecks or they are heavy burden to the enterprise? And in my experience, the number one thing that comes up for a lot of enterprises are business documents. And then moving beyond that, people start looking at email centric workflow email. People are also looking at conversational AI, chat bots and understanding that for customer facing websites. And once you move beyond that, you start looking at some of the processes that involve images. So we are seeing customers who have visual inspection type use cases in the manufacturing plants, in the supply chains where they could be processing those images automatically to do that. And and finally, they're moving to other types of data like audio and video beyond that. So that's that seems to be the logical progression in my life.  

Ian Barkin [00:19:44] Outstanding. Okay. And so, Brad, it's from your experience of having having used tools that solve this problem for yourself. So being a company that was the archetype of good that everyone else wanted to tap into the special sauce. And now having a company that that does help people. When you walk into an early prospect and you're describing your sort of your philosophy in the world, what what's the telltale sign that they get it like that that organization really does have their act together, that they are making the right investments or steps. What what is a sign of good to you.  

Brad Cordova [00:20:19] Yeah. So there's there's three things concretely that I see are the are the biggest variables of of the winners and separate the losers in the space are the soon to be winners but not quite yet. And the first one Maneesh touched on this is the companies that treat AI or automation more as like a science experiment where you create a cool demo and then you read on your website where an AI company just doesn't work in. Today's world any more. The the businesses who are successful automating with AI, they don't just think about quality, but they think about the tradeoff between quality, cost and speed. And they take these models and make A.I. business applications. And today's successful companies expect an hour away from AI. It's out of the world of this is a cool science experiment. This is called demo and and really giving this enterprise quality guarantee, cost guarantee and speed guarantee. That's really important. The second thing is, is using a centralized, unified A.I. platform where this platform can harness any worker. So I think the dirty little secret of A.I. is that a lot of what makes A.I. successful is having humans in the loop, whether it is humans or labeling the data to teach the A.I., whether the humans are quality, checking the AI. And so a platform that can, let's say, use the best AI, let's say if it's Google or your internal model or or Amazon, but also having in that same loop humans and treating it as a feature, not a bug and integrating that also with some kind of software, let's say RPA and you're always using the best and you're always routing to the appropriate worker. That's important. And in the second part of this unified A.I. platform is that you have all your applications in one place. So we all have iPhones now or Android phones, and all your applications are on one phone before you'd have to buy this thing to do X that thing to do Y, this to do. But now all your applications are in one place. It's really important to take this iPhone App Store idea seriously in A.I. And then the final is what I call automatic automation. So you may have some workers be able to automate in a certain degree, have a certain quality costs and speed, but you really want this to automatically improve over time. And so this isn't just improving the quality, but improving the quality, cost and speed automatically while you sleep as you process more data. And then the final piece is not adopting a new platform and throwing everything away that you've invested in previously, like your RPA solutions, your your BPO solutions. You want this to all integrate together. So when you're cleaning data labeling, structuring, training models, monitoring, you don't you don't want to throw away what you have or make a totally new investment. You want all these things to work together. And so the companies I see that are successful do these things outstanding.  

Ian Barkin [00:23:33] I have to say, from, from my, my in my day, I'm going to use old as as a as a as a useful thing here. But having been in the BPO industry as long as I have and been in the RPA industry, that's a really insightful and respectful comment you just made as far as don't throw away that what you have and build upon it and and add value by integrating it so that it was. I appreciate hearing that. So okay. So so those are hallmarks of an enterprise that does things right there. I think we all believe that most enterprises are still struggling in many cases because they've just been working their way through this sort of digital just sort of maturity curve where they've they've played with many of the tools you mentioned with the RPA and other intelligent automation things. They're starting to address documents as a source of unstructured data. Let's let's spend the last sort of 10 minutes doing two things. One, if anybody has any questions, put them in the chat and we will absolutely answer them. Please do that. And I'll I'll I'll say your name again if you want to hear it online and we'll ask the question. But but more so, I'd like to hear Maneesh and Brad, your perspective on how you help enterprises that that aren't the hallmark of of mature and good yet how does super I work with with your clients to to to give them that that insight to help them understand that cost quality and speed triptych and how they how they invest appropriately, how they educate internally and how they how they structure their strategy going forward.  

Manish Rai [00:25:12] If I can take a stab at it like we have a lot of companies coming to us who have been disillusioned with the the ADP or intelligent documents processing solution they have deployed. So if you look at it, Ian, you and I have been living this space for five years and know what so-called modern IDP solution started emerging five years ago. But what we are learning is most of those solutions have put put a new interface on the same old engine, which is the OCR, and made it more accessible and usable for the business users. But they haven't fundamentally rebuilt the engine underlying to process documents with greater accuracy. And so what they love about us is that we have an open platform that can pick the best model because ultimately we don't care about building our own models, but picking what's the best to deliver right away to the customer. Second thing is a lot of those solutions don't have all the capabilities for managing the human part of processing documents very well. They leave it up to the enterprise and what they love is they're built to hold. We have built a whole resource management engine which treats EAI and human workers very similarly and routs the job to the best worker out there. And we have tools, but humans need motivation. They need, you know, some oversight. They need gamification, they need vacations and they live on breaks. So we built all these capabilities very randomize the output so we can find people who are just clicking on the first option all, all the time. We ask them if they're still there, we gamify to keep them engaged, and we can also crowdsource and measure the quality of the input of the users. And then that allows us to root out the best human or AI to actually deliver out away because ultimately the customer doesn't care what you're using underneath. All they care about is this is the metric, this is the outcome we want. And can you deliver to that in in the best cost possible for it. So for a given task, maybe humans outperform EHR and we leverage them at play in other tasks far outperforms humans, we leverage it. So that's something very unique. And finally, the point Brad mentioned is, is guaranteeing quality and we haven't seen anybody else step up and see that, hey, we stand behind and we have all the mechanisms where we combine the output of AI input and the humans and measured the quality of each worker to actually guarantee that output quality to you.  

Ian Barkin [00:28:18] Outstanding. Brad, what do you have to add?  

Brad Cordova [00:28:23] So I spent a lot of my time trying to make I unsexy. And what do I mean by that? There's so much hype and so much false marketing in the air space. It kind of sickens me. And you have these two personas. You have one persona that has been burned by solutions. They've tried and there's it doesn't work, or you have someone who believes it. And it's like, I can do everything. I can read my mind, it can solve cancer. And, and so I spent a lot of time just trying to educate what's what's real and not. And at the end of the day, all I care about and all I want is for our customers to be successful. And it's rarely that this super sexy, brand new trillion parameter model is the way to go. It's oftentimes the linear model or just a workhorse model coupled with the humans, but giving guarantees. And then there's a lot of great technology in our case behind the scenes where you have a router and a combiner and use reinforcement learning. By the end of the day. I don't really care about any of that. I just care about delivering results. And so we spend a lot of time making sure people really understand what I can and can't do, choosing the right project and making sure they get an ROI in terms of quality, cost and speed, because in the end, that's all that matters. And and that's what that's what I that's where I becomes ubiquitous and changes the world when we make it practical. It's not the the cool flashy demo. It's it's on the, let's say, unsexy problem that humans are working on. And humans shouldn't work on a lot of these problems. We want humans to be more human and automate a lot of this stuff. So a lot of time is spent in the practical nature with our clients either trying to like we like to, we like to show rather than tell. So we just say just use our product and and you make your own decision. I don't like all this marketing and it's it's not good for anybody long term.  

Ian Barkin [00:30:36] Manish, as as as a former chief marketing officer and as your VP of marketing, what do we say? What how do we feel about that? No, I do absolutely, totally agree. I, I again, as I said earlier, I respect in them impressed by your your pragmatism towards the market. But what used to drive me crazy in the days is when I would go to a prospective enterprise, not a client, but someone we were trying to to win over. And they would pose the question and they many of them did at one point, and it was worded almost exactly the same every time. And they would ask, Why do I need a RPA if I can just use I? And it just I mean, it was exactly what you described. It was just it was a it was an unknown and sort of founded faith in in the concept that it could do magic. And so they didn't understand that technology is, in fact, hard. Some of the more traditional technology is exactly what's the right sort of solution for the problem that they're facing. RPA is one of a number of tools and a toolset. And so it is that belief and magic and and a fascination with shiny things that I think is has constrained true digital transformation for way too long. So I love I love your approach and and your and your emphasis on sort of the right worker for the right task, whether that worker be a person or an algorithm or or a combination. Outstanding. We're at the end of our time. I'm going to I'm going to.  

Manish Rai [00:32:13] Address some comments here, if you could address them.  

Ian Barkin [00:32:18] Yeah, sorry. I was I was blathering and missed which one we want to put up. What do you.  

Brad Cordova [00:32:22] Think? I think the system is similar. Oh, sorry. Go ahead.  

Manish Rai [00:32:27] Go ahead. Brad, you want to take the first part?  

Ian Barkin [00:32:31] We're talking about this one.  

Manish Rai [00:32:32] Yeah.  

Brad Cordova [00:32:33] Yeah. So the the the the first thing they're asking very similar questions, but I'll just read the question. So I think you're saying. Okay, when you talk about the advantages of a centralized platform, what do you look for in a platform? What does it make sense to go with something already integrated with their ERP? So I think the first thing when you're looking at a centralized platform is you want it to be flexible. So you you want this application layer where you can plug in a certain machine learning model, let's say, from Google or Amazon or an open source, or if you have a great machine learning team, plug that in because we view machine learning models as commoditized and there's so much innovation in the space. Whatever machine learning model you choose today, there's going to be a better one next month. And so any approach you choose shouldn't be locked in to a specific vendor or specific machine learning model, or else you're going to get outdated so quickly. And I can't stress that enough. The same thing is true. It should be able to suck in A.I. models. It should be able to suck in any human, whether that's internal to your organization or a BPO or somebody in in South America or India or Asia. And then it's and then it should also integrate to your RPA or ERP solutions. In terms of your second question, does it make sense to go with something already integrated with your ERP? So I would say if it if it has these elements I just mentioned, then it doesn't make sense. If it doesn't, then then it doesn't make sense. And the other thing is you want something that has enough to a to guarantee the quality. So machine learning is powerful and I think this is kind of the point of Thomas Pointer. I don't know if that was a pun, but if it doesn't give guarantees around quality costs and speed and clear business are wise, then I would also consider trying something else. And then and then the last question which is very similar what are the best software tools for AI companies and what are the challenges currently needed to overcome and help the to capture quality of unstructured data? So. So it's hard to answer this in general because I would need to know where you're at in your journey. Let's say you have a team of expert PhDs in machine learning. I would say just directly use Google, Amazon, Microsoft or some open source because your team will figure that out. But for most people who don't have an army of data scientists or machine learning engineers, then I think using Google's tools directly or machine using machine learning directly is a is a really bad approach because machine learning is powerful, but it's it's a really pain in the ass to use. It's like the is like the high interest credit cards of technical debt. And what I would recommend is to use an application layer where then you can plug in these great machine learning models to get the power of them. But then you get the governance, you get the guarantees, you get the management, you get the integrations, everything that you need if you're not an expert in this.  

Manish Rai [00:35:42] Yeah. And I can take the question directed at me. It's a hard question, but I would like to address the elephant in the room. Thomas, I don't know which part you agree or disagree with my outlook, because what I said was the first generation IDP solutions were not very flexible. And I think that's exactly what you're saying over here, that you found that some of the approaches were not flexible. They had a fixed set of OCR. They're using it as a black box of EAI models inside and you don't have visibility and then you don't have a quality guaranteed associated with that. And I couldn't agree with you more on that point that yes, those solutions were not flexible. And our approach here and thankfully a lot of companies in the market, you know, they were not API first companies, they were in RPA or some other business and they put out that solution and they lacked a lot of deep expertize. Why is the difference with some areas of our foundation, our engineering team comes from people from Google Green team, from the original founders and the Microsoft brand himself as a ML BSD dropout from M.I.T.. And and so we have solid engineers who understand all the challenges and how to overcome right to where you could have over fitted model or you could have a loss by overtraining it and things like that, whatever can go wrong. And we have taken steps to make sure that we maintain the quality and improve the quality over time continuously and do the active learning in the right way possible without relying too much on the outliers and having them kind of affect our underlying models and impact results.  

Brad Cordova [00:37:44] And I I'd like to address this because I agree 100% with Thomas and this is what I've been saying for years and years, that if you only use machine learning model, you're going to have horrible approaches. Unless you're a total expert in this, then you but then even then they don't use just machine learning. They know how to wield this to it to work. But this is exactly why we built superior AI, because if you just try to use the Google model, you're going to have these problems that you have. And I think more people need to speak out about this, like Thomas, and you need to use humans in the loop. I think there's this weird ethos that using humans is is somehow a bad thing. It's one of our human brains and one of the most wonderful things in the universe and there's a lot of fraud going on. Was like I only. And then behind the scenes, we all figured out that they're just using they're just using humans. And it's important to not just use machine learning. And I think that's why people got so many bad results in the past. And so I think that's an important point. And I agree. And that's why we personally and I we built this even to machine. We built this application layer on top of machine learning because machine learning only approaches or machine learning first approaches. If you're not the expert in this, it's not going to work for you. I can promise.  

Ian Barkin [00:39:07] Outstanding. Well, thank you, everybody, for questions, Thomas. For your for your spicy one, too. That was great. Well, we're over time, I want to all wrap up and say thank you, Brad. Thank you, Maneesh. For for for doing this with me. It was fun to to do the poll to hear from so many people what they thought about the relative in general maturity of of the enterprise space as far as its understanding of active sort of strategic initiatives around and success with tapping into unstructured data and in. Converting it into to value. Fascinating. Just because there is such an excitement around all things digital and digital transformation in general. We all touched on our sort of our histories and our heritages that that are informed by various different elements of that from RPA and intelligent automation to machine learning. So it's an exciting space out there. I love Brad's wrap up of of the human in this amazing AI. I've got one. I think it's okay sometimes. And and I do agree that that sort of human in the loop orchestration of work with using the right tools for the right jobs and really unleashing humans to be as creative and and empathetic and and as human as we are is is the winning algorithm, the winning calculation. So, so much.

Manish Rai [00:40:33] In fact, I do a small plug in on me for Brad and I will be talking with my old colleague from Automation anywhere, Rob Hughes, who's with Tech Open on May 4th. And we'll be actually tackling human centric automation as a topic. Very cool. Stay tuned.  

Ian Barkin [00:40:52] Great. And so if you if you were on this broadcast and you think it's useful, do share it. The link is available on LinkedIn, so easy enough to share with your friends and colleagues and other people who I know will benefit from the insights that Brad Mooney shared. Brad, any any last words before we close the broadcast?  

Brad Cordova [00:41:11] No. It's great to have these conversations and appreciate it.  

Ian Barkin [00:41:16] Beautiful. Thanks, everybody. Have a great Thursday.  

Manish Rai [00:41:19] Thanks, as always. A pleasure to be hosting an event with you.  

Ian Barkin [00:41:24] It's fun, right? Yes. I'll see you at the next Uber in Orlando. Oh, right. Okay.  

Brad Cordova [00:41:31] All right.  

Read less
Share on TwitterShare on Twitter
Share on FacebookShare on Facebook
Share on GithubShare on Github
Share on LinkedinShare on Linkedin
Share on LinkedinShare on Medium
Button Text