DigitalNOW Podcast | The future of automation – Is terminator to blame for our fear?

Automation… what is it? Is it a robot? Is it software? What exactly does it look like?

If businesses are to stay competitive in the new year, speed their time to market, and reduce costs, automation is a critical part of any strategy. In this episode, we’re joined by Logic20/20 manager, and automation expert, Amit Unadkat. He discusses various AI technologies, the current disruption to the market, and the top automation trends for 2021.

Transcript

Matt Trouville: You’re listening to DigitalNOW, an original business and technology podcast by Logic20/20. I’m your host, Matt Trouville. Each episode. I’ll be interviewing a new expert to learn more about industry trends, fascinating new tech, shifting customer expectations, and the steps every business can take to stay ahead.

Welcome everybody to the future of work and automation with Amit Unadkat who is one of our managers of the digital transformation practice at Logic20/20. He’s passionate about automation and product management. Amit, welcome!

Amit Unadkat: Thanks, Matt. Happy to be here! Thanks for having me.

MT: No worries, mate. First of all, how are you? How’s life these days? I know it’s crazy right now – I just want to make sure you’re safe and healthy and all the good things.

AU: Yeah. You know, all things considered, Matt, [I’m] doing pretty well. It’s been about a year now since I’ve been in Seattle. What I’ve been doing is I’ve been exploring, just being outdoors. I’ve got some great photos, I’ve really reconnected with nature, and I’m trying to make the best of it. How are you?

MT: Oh, great. That is the perfect sort of hobby to go through when we’re in a pandemic. A good idea! I wish I was doing the same thing. I’m good: I’ve got a nine month old, I’m working from home—the whole mixed bag, right? It’s a learning curve for everybody.

Anyway, we’re here to talk about automation. How’s that sound?

AU: You know what, I can talk about automation all week! It’s something that you’re seeing everywhere and sometimes automation is happening and you don’t even know that it’s happening. You’re just taking it for granted. What do you want to know? Let’s talk automation.

MT: Well, I want to know everything—and I know you’re passionate about it. Thank you so much for giving us your time and being here today. I think it’s gonna be very valuable to those listening because you’re right. It is everywhere. You can’t turn around without hearing the word[s] “automation” or “RPA”—all these different use cases for automation are popping up. So, I’m glad we can dive into it.

My first question is: what is it? Is it a robot? Is it software? What exactly does it look like? I know there’s many faces, so can you just dive into that for us and let us know the different types of automation that we’re seeing?

AU: Absolutely, yeah—you already mentioned a couple of them.

First, I’ll start off by saying: I always think of automation the same way [that] I think of our bodily reflexes: it’s something we don’t have to think about, but it just happens, right? I think it’s the easiest analogy you can draw to automation in any sense of the word, whether you’re looking at a business process that needs to happen by itself, or whether you’re looking at a car that drives itself, or if you’re looking at manufacturing processes where products are created, but they’re being done without humans. The whole “without humans” aspect is referring to the fact that no one has to be there, think about what’s happening, and intervene. As long as that’s happening, you’ve got automation.

You already mentioned RPA, which stands for robotic process automation. That’s one example of a software, and RPA is at its most basic level a software, that goes into a system and performs the tasks that a human would, without the human being present. Again, it comes back to this idea: as long as the human is not present, there isn’t any effort from a person to think about what needs to happen, what needs to happen next, then you’ve got automation happening.

MT: That’s why it’s everywhere, right? Because of the convenience of it, what it can do for a business, [etc.] Very interesting stuff.

My next question is: where do AI technologies—like machine learning, for computer vision, natural language processing—where do they fit in?

AU: Oh, yeah. This is the latest hot topic. I would say about five years ago, a lot of the emphasis was on very simple automation tasks. It would involve anything that was deterministic, which is another fancy way of saying rules-based. Now where AI fits into this is: it starts to tackle that thing that I alluded to in the first question, which was that thought process. Anywhere where you have had traditionally automation, you can’t let it run without human intervention.

That’s where AI is now starting to fit in, because you’ve got computer vision, you’ve got natural language processing, you’ve got optical character recognition, all of these technologies within the AI and machine learning world that are augmenting what automation can do. So traditionally, let’s say you were able to tell a software that I need you to do X if Y, and Z.

Now what happens if X didn’t equal Y and then didn’t get to Zed, right? You would need someone to come in there and intercept that automation and make sure it’s working correctly. Whereas now what we’re finding is you’ve got self-correcting models. You’ve got bots and automations that are saying “You know what? It’s not always going to be that X goes to Y and then goes to Zed. Sometimes X might go to B and then go to D and then end up at Y before going to Zed. And even if it doesn’t get to Zed, it’s fine.” Those are decision points that machine learning is stepping in to fill, because once you’ve got a sufficient amount of data, you can then train a model to say, this is what I really expect to happen. If it doesn’t happen, then you go and figure out what the next steps are.

I think the simplest way to think about AI with respect to automation, is it really takes away the need for a human being to come in and say, “Is this working the way I originally designed it?” Because the system will start to figure itself out. It’ll start to train itself, it’ll start to recognize patterns, and it’ll just do the job, but not doing it specifically in the way that it’s being coded. It’s very exciting stuff.

MT: Wow – that is very exciting. Are there stats around how accurate that is? I know there’s probably different models and different software to use in different angles. Is there an argument saying that it’s more accurate than a human? Is it still in the testing phase? Where are we in that process? Is it fully adapted? What are we seeing here?

AU: Very relevant question, Matt. Accuracy and the precision that the models can have—it’s always a work in progress.

I don’t have a number specifically for you and that’s because it really depends on each use case. I will tell you, however, it’s predicated on data quality, right? When you have a process that is relying on a machine learning model, the better quality of data that you have, the higher volume of data that you have, the more likely it is that you’re able to then train the model in a very accurate way.

MT: Yeah—and it’s ever evolving, right. As you said and that’s the beauty of it. It’s going to keep learning and it’s going to keep evolving and moving forward. So, that’s very exciting stuff. Robotic process automation or RPA, as we’ve been talking about—is one of the leading ways that business to automate their processes.

Can you talk about sort of some more use cases? Can you explain that to a five-year-old? I think that’s a really crucial point that the people listening to this will want to hear about.

AU: Definitely. The way I would explain RPA to a five-year-old—a keen five-year-old who wants to know—is it’s basically a software.

[It’s] something that sits on a computer that does [the] work that a person does, without the person being there. That’s the very high-level way to think about it. A five-year-old at this point might be a little confused, in all fairness. What I would say in terms of one of the best use cases to think about (and every company can imagine this): bringing on a new employee is a very repetitive process. Once an employee has been offered a job, the process to get them enrolled into the various benefits systems, to get their payments and stuff sorted, and also to get them integrated into the environments that they need to work in.

There’s a lot of stakeholders that are involved in that from HR to finance to IT. This is one of the proven areas of automation that many companies who have started down the RPA path have already figured out. You could really reimagine what a business process should look like when you have automation in play because you’re now no longer relying on someone to be there, to accept the information, to think about where this information needs to live, and then ultimately put that information in a system.

MT: Just so you know, I’m the five-year-old. That’s why I put that into [a] question. I need to be told like a five-year-old.

AU: No, that’s great. You said you have a nine month old. I bet your nine month old is going to be all up in over this by the time they’re using their devices, right? This will be so natural to them that they won’t have to distinguish between what is artificial intelligence versus what is just regular software. They’ll just be like, “Yeah, it’s a system. It’s here to do my work and here’s how I’m gonna use it.”

MT: Yeah. It’ll just be the normal, right. I mean, she’s already trying to use my phone, so she’s already on the path. Working between different systems—that’s gotta be a huge time-saver for any business, which like you said, you can then go and put that time into more value-added tasks. I think that’s crazy to think about. That’s not being adopted as quickly as it potentially should be.

AU: A fair point, and I’ll say this before we move to the next question: the business case for RPA is very appealing, right? If you have a piece of software or a bot that can do the equivalent of a human being’s worth of work, you can have it run 24/7 and the amount of work that you could process with one bot is equivalent to three people’s work, right? Because you’ve got eight hours and you’ve got 24 hours in a day, so you’ve essentially got approximately three employees worth of work, and—get this, the license cost for that one bot is a fraction of what it normally costs a full-time resource, right? So, approximately $15,000 for a one license to operate fully autonomously.

So, you can see just based on those numbers alone, the business case is quite substantial, once you start to save some time.

MT: Yeah, very appealing to any business looking to upgrade their systems. What’s the demand for RPA in the market right now? I was reading something that said that it’s estimated that RPA spend will reach $2.4 billion by 2022. You know, that’s not too far away. So, obviously the market is seeing this and there’s demand. Whereabouts are we at?

AU: Yeah. I think that that estimate is actually an underestimation of what the market size actually is. What I’m hearing from the marketplace is, companies like UI Path, companies like Automation Anywhere, are looking to IPO, right. We we’ve seen Blue Prism, which was one of the first RPA companies to IPO, when they went public, their stock went gangbusters, right. Because a lot of people understood the value of what the software was bringing. If I look up the stock today, [it’s worth] 1.64 billion pounds. That’s just Blue Prism alone as one company within the space. That just goes to show that I think a lot of people haven’t yet realized the scale at which this could operate because you’re going across multiple departments. You’re going across HR, finance, IT—you’re going across multiple industries. I’ve seen it successfully be implemented in banking and telecommunications.

The good news is that adopting this technology is a very short timeline. It can be a matter of as short as three to four weeks, depending on the layout of your IT, and kind of the decisions that need to be made, three to four weeks to go from identifying a process through, to publishing it in production and having the automation run, and actually performing the tasks they need to.

MT: I mean, everything you’re mentioning sounds like a benefit to the business, right? It sounds like you’re saving time, small investment upfront, big savings on the backend, if you think about it. Why isn’t it being consumed more? Why isn’t there more adoption around it? Cause it seems like the silver bullet, right?

AU: Yeah. Great question. Fear of what’s coming in and changing things around. That’s definitely a critical aspect of it. The other few aspects to consider are that you’ve got business processes like onboarding new employees, like your finance processes, like your IT processes. They become very critical to the benefits and the business case that you have outlined. In order for that to successfully be implemented, you can’t start with just one or two robots and just go from there. You really have to have a broader plan of “What is your operating model going to look like? Who’s going to support the technology and production? Who has ownership of the designs?”

I always tell clients, “You shouldn’t just rush to automate because you can, like, there may be a very good process that will yield thousands, if not millions of dollars of benefits by automating it, but just because you can automate it doesn’t mean that you should.” There are ancillary considerations around it, which require some assessment, which requires some planning.

The final thing I’ll say on this question and why RPA is not a silver bullet is without that cognitive element of it, RPA is just automating very simple steps, right? Those simple steps can yield a very big business case benefit, especially when you’re looking at onboarding, you look at IT and finance processes, but many of the aspects of a business operations today that want to be automated require some level of cognitive capability, some level of machine learning. You’re getting more and more capability with Azure, AWS, Google Cloud Platform to use machine learning models, to use data and analytics, to make those decisions. Once you start to pair that up with RPA, which we’re seeing with Microsoft’s Power Automate, Microsoft’s Power platform, we’re seeing a lot of that integration with cloud technologies, with advanced analytics models, as well as RPA all coming together to give you this very intelligent piece of software that can run autonomously, truly autonomously, where it doesn’t require someone to come in and make a decision or a judgment. It constantly is learning by itself.

MT: In your experience, when does a client know they’re ready? Is there an automation readiness that they need to achieve, a level that they need to achieve before they start doing this? I know you touched a little bit on it before, but could you dive into that for us?

AU: I always like to tell my clients: “Never automate a broken process.” It’s one of the mantras that I go in with. We never start with the complex processes. We always do an assessment, first and foremost, as part of our discovery to make sure where the automation readiness is there.

Finally, we have to make sure that the data, what I alluded to earlier, the quality of the data is there to support. If you need a machine learning model, if you need some sort of analytics engine to decide which step to take that the, you don’t automate a process and then work backwards to try and figure that out.

I think you really need to figure out the model first, if you’re going to go after a complex process and then figure out the automation piece next, because arguably that’s the easier part than some of the cognitive AI stuff.

MT: Yeah. I mean, I think that’s very important, the assessment part, because you’ve got to know exactly what’s happening with the information, what’s happening with the process, and really dive into that. I like that.

So, these always feel like a cliche question, but it’s just where we’re at, right? We’re in a recession. Many people around the world are struggling. COVID-19 has affected businesses financially and we have to find ways to do more with less. We just keep hearing that phrase over and over again. What does this mean for automation technologies and the future of?

AU: Yeah, it’s a tough reality. There definitely seems to be a huge impact on the economy. As a result of COVID, we’re finding that many of our clients are looking for solutions that will lower their operating costs while also maintaining (if not improving) the customer experience.

Arguably, automation is the perfect candidate for that, because you’ve now got this very low cost technology, $15,000 for a license and an automation that could do 3x the amount of work that an FTE can. While at the same time, you’ve got a lot of advancements in machine learning, so that if you’ve got customers that are calling in, you’re able to transcribe what they’re saying in real time, take that intent from that real-time transcription, and then start back end processes to help turn over a sale. And that’s one example of how you can use automation to augment how you can service a customer’s need and in a recession, going back to that main point, what it allows you to do is it allows you to serve your customers with a lower cost.

It allows you to serve your customers with really some of the latest technologies. It builds this loyalty and it allows the employees to focus on risk mitigation, to focus on the customer experience, to focus on building better products. So going back to this idea, a recession is the perfect time for an organization to really reinvent itself.

MT: Couldn’t people also think, “Okay, well that’s gonna take my job, right?” Because there’s going to be fear. The more artificial intelligence is implemented into companies, the more people are gonna say, “Well, my job’s going to go out the window” Is that true? Are those rumors true?

We have to blame Terminator for this. As soon as Terminator came out and Skynet took over the world, people started to fear this technology when it’s actually being built to help us. Could you address those rumors and those feelings?

AU: You know, automation has been referenced as the fourth industrial revolution on the jobs market, right?

It’s really going to transform the way we work. In fact, our partner UIPath, one of their quotes and their lines is they are representing the future of work. If you look at some of their videos, you’ll see that what they’re alluding to is the fact that before, if you had a process and a company that was growing and demand was really swelling, you really have to throw people at that, you would have to hire more people. You would really have to make sure your workforce is there to serve customers and is there to transact all this stuff that needs to happen in the backend.

Now, however, you’ve got this new capability of building out your virtual workforce. [It’s] that the idea of having a virtual workforce that can pretty much infinitely scale at a fraction of the cost. That’s really what’s driving a lot of the disruptive force behind automation in the labor market.

Even though there is a disruption happening in the jobs market, where automation is taking over certain things that people used to traditionally do, it is also creating new jobs [for] people who are now designing those automations, supporting the automations while they’re in production, and also working on the AI and machine learning elements of those automations. Those are new jobs that are being created from the effort.

Frankly, there were jobs out there that aren’t suitable for anyone to do, and there could be safety aspects related to it. There could be the fact that the job was so repetitive and cumbersome, that it introduced risk to the organization. We would rather have employees focusing their time on how best to serve customers and how to mitigate risks that may be coming down the pipeline. And in order for them to really focus their time on those value-add activities, we really have to have them focus away from the manual repetitive work, which is perfectly where RPA and automation really fit in.

It’s a very powerful pivot to go from fearing this technology to now saying, “How can we use this technology to actually accelerate our growth?” I do believe that automation [is] becoming even more prevalent than it has. It’s going to give us as workforce, more time, more energy, and more capability to be more creative in the products that we develop, in the way we have customer experiences, maybe even reimagining what our work weeks look like. We’ve structured society Monday through Friday, and then the weekends around work, right? Maybe that doesn’t need to be the case anymore. Maybe now that we have AI and we’ve actually proven with COVID that everyone can work remotely, right? So that’s already disrupted education, it’s disrupted healthcare. It’s disrupted many service businesses, and unfortunately it’s disrupted places like restaurants, but everyone has adapted very quickly in this new setting. I think the same thing will have to happen when AI really becomes self-aware.

That is going to be equivalent to like a COVID event where all of us are going to have to completely re-imagine how we use technology because all of a sudden the technology will be aware of us, right? It’ll be aware that we are using it to perform our actions. So I think it’s not a matter of if it will happen, I think it’s a matter of when and those companies that are investing heavily in this space today, while they may have higher capital expenditure against their peers, maybe not, I think they’ll end up being the forerunners and the ones who come out in front, especially when the growth cycle continues and those that don’t invest like those companies that weren’t able to pivot during COVID, they would have to shut down.

MT: Well, mate, I tell you, if you tell people it’s going to extend your weekend, I think you’re going to see adoption rates go through the roof, right? I mean, as soon as you said that my eyes lit up!

What’s the most important thing you want people to take away from this conversation? You know, whether it’s about the future of work and automation, like the trends we’re going to see, what’s the most important thing people should be thinking about after listening to this?

AU: I think one thing that we haven’t really touched on that I’d like people to take away is: even if you aren’t a developer or coder or software engineer doesn’t mean that you can’t benefit off of these tools. Many of these tools are being built in a way that are intuitive for a business user to use and their user interface is built in such a way that allows you to drag and drop in, allows you to really work your way through a defined workflow to build an automation and the training for RPA, for example, UIPath, they offer it for free on their website.

So if you want to go and learn how to use the UIPath software, you can download their free software. You can go through all their training programs for free, and you can go ahead and start automating within a short period of let’s say even two weeks, what a lot of tools are doing is that they are, they are implementing smart software that can auto-detect the right model that you need to fit a certain type of data. You don’t need that data science background to say, “I’ve got this quality of data, therefore I need to use this model.” Going back to this idea of the key takeaway, it’s becoming accessible to everyone and this idea of having a democratized set of technologies, where you’ve got citizen developers in any organization, coding, building, innovating.

That’s what the future of work looks like. It’s no longer that, “Hey Matt, go process these thousand invoices and we need to get it done like yesterday.” It’s like, “Hey Matt, we’ve got a backlog of a thousand invoices. Can you come up with a technology solution that will automatically process this with low risk, high accuracy, and a fraction of the cost that it would have taken you to complete?”

MT: Wow, I’m excited for the future. Cause, you know, that’s just so important, especially now, but you know, we do have to wrap this up. I know you’ve got to go. Thanks again. Really appreciate this. And I hope we can do this again, I enjoyed this.

AU: I would love to. Thanks Matt. Bye.

MT: Thanks very much.

DigitalNOW is an original business and technology podcast by Logic20/20 that is released on a monthly basis. In each episode, host Matt Trouville interviews a new expert to learn about industry trends, fascinating new tech, shifting customer expectations, and the steps every business can take to stay ahead. Check back here for future episodes, OR you can find us on all major podcast sites, including Spotify, Apple Music, Pandora, and more.