Bots. They’re everywhere, proliferating fast, and evolving their capabilities. Most of us are familiar with them in the form of chatbots, crawlers, and of course RPA bots, but what about an emerging class of autonomous software programs called Digital Coworkers? They’re not just next. They’re now, and are already impacting the future of work in verticals such as the insurance industry.
To learn more about Digital Coworkers and how they’ll interact with their human colleagues, we talk with Chaz Perera of Roots Automation. As the former Chief Transformation Officer of AIG (America’s 4th largest insurer by assets, as of 2019), he sought a better way to deploy robotic automation in enterprise operations. Chaz explains to us why Digital Coworkers succeeded where other bots failed. Along the way we’ll learn what the magic number is of automatable processes organizations need to have in order to justify establishing their own Center of Excellence, why a bot’s greatest value might be freeing up staff so they can spend more time with customers, and what a future with Digital Coworkers might look like.
Guy Nadivi: Welcome, everyone. My name is Guy Nadivi, and I’m the host of Intelligent Automation Radio. Our guest on today’s episode is Chaz Perera, Co-Founder and CEO of Roots Automation, which claims to have the world’s first self-learning, zero integration Digital Coworker bots. There’s lots of debate about whether automation in the form of bots will replace more people or augment more people, and that’s one of the subjects we want to hear about directly from an expert. So we’ve invited Chaz to join us today and share his thoughts with our audience. Chaz, welcome to Intelligent Automation Radio.
Chaz Perera: Hi, Guy. Thanks for inviting me on. I’m excited to share what I’ve been learning over these last few years in the automation space, and yeah, hopefully your listeners will take a few bits and bites from it, and help them improve their businesses.
Guy Nadivi: Well, fantastic. I think the first thing, Chaz, that everybody would like to know is what was the path that led you to start Roots Automation and your focus on Digital Coworkers?
Chaz Perera: Sure. I, at one time was the Chief Transformation Officer of AIG, a large multi-national insurance company. We were, at the time, we had been leveraging RPA technologies for roughly five years. It struck me as odd that we were spending, in our case then, many millions of dollars running an RPA program, and we were sort of challenged to get to value. I started to ask people, my peers at other companies how they were sort of dealing with that challenge, and I was coming across the same sort of problems, which is that the traditional RPA deployment methods, which really mean you need to spend a lot of money on a variety of different skill sets, you need to spend a lot of money on different technologies, and you need to spend a lot of time, typically many months, building a bot, and that is different from what RPA technologies typically promise, which is speed, and short time to value, and so I wanted to come up with a different way to deploy bots in a working environment. The second thing that struck me and the reason that we are focused so much on Digital Coworkers, as opposed to just bots, is at one time, I also ran shared services for the company. I noticed that when we were running shared services, often, people here in the U.S. or the U.K., for example, would often complain that the folks in the Philippines or in India, wherever our shared services entities may have been didn’t understand the business. They weren’t helpful, and so people would rather just keep the work onshore and do it themselves. I noticed that when we started to introduce bots into the operating environment, people had the same responses that the bots weren’t doing the job properly. They weren’t as complete as they needed to be, et cetera, and so it struck me that we needed to create bots that had more human-like qualities that, where humans could feel almost a fellowship between themselves and bots. Otherwise companies would be making huge investments in robotic technologies, automation technologies, but not getting value from them because people were essentially opting out of leveraging these technologies.
Guy Nadivi: Now, you just mentioned shared services, so I want to ask about an article you recently posted about the importance of standardization for automating processes. Before automating a process, in addition to it being standardized, it’s also a well-accepted best practice that it should be optimized, as well as thoroughly documented, so I’m curious, Chaz, what percentage of business processes are you seeing that are standardized, optimized, and documented before you begin to deploy your automation?
Chaz Perera: It’s a really good question, Guy. I think the reality is when you start to talk to a prospective customer, you find more often than not, and by more often, I mean, 70, 80% of the time, their processes are not well-documented. Their processes aren’t optimized. Their processes may be standardized in their mind, but not necessarily the way we, in the world of sort of process improvement, think of what a standard process looks like. All that being said, one of the things that we see at Roots Automation, and this is the advantage of sitting outside of a company and seeing the business processes of lots of similar companies, what we find is that the processes at one company are often similar to the same process at another company. A very simple example would be the accounts payable process within a finance team at one insurance company, is often very similar to the accounts payable process at another insurance carrier. In fact, what we’re finding is that for those back office processes, the similarities amongst those processes are closer to 80%, so what we end up doing is thinking about that remaining 20% and how we can build our Digital Coworkers in such a way that that 20% becomes something you can configure as opposed to having to always build from scratch. Similarly, as we move into the front office of insurance or the middle office of insurance, so underwriting and claims, we find similar things that the underwriting processes have a lot of commonality, insurance company to insurance company. The claims processes similarly have a lot of commonality, insurance company to insurance company. I suppose a long-winded answer to your question, even though these companies may not have standardized processes in their mind or they may not have documented these processes to the degree that we would like, we see enough of these processes to be able to say that there is standardization, there is commonality amongst the processes that we can leverage and take advantage of.
Guy Nadivi: In another recent article you posted about automation and return on investment, you touched upon the topic of automation centers of excellence, which generally speaking, are highly recommended for enterprises as part of their automation journey. However, you state that, “Statistically speaking, only about 4% of COEs deliver a positive ROI.” Is everyone wrong then about the need for centers of excellence, or is there a better path to take in order for them to successfully deliver an appropriate return on investment?
Chaz Perera: It’s a really good question. I would say there is no one size fits all. I think if you are a very large enterprise, having an automation COE makes a ton of sense, because you want to standardize on a certain subset of technologies. You want to be able to take advantage of the scale, that leveraging those same technologies will bring you, but if you’re a medium-sized company or a smaller business, the requirements around talent, the requirements around the technologies you need to bring forward are so vast that a COE is not an affordable exercise for most companies. Just to give a very simple example, if you want a bot to be much more human-like, and you think about how processes often begin today, they often start with an email, you need to think about Natural Language Processing, and so you need to bring some AI, machine learning expertise to the table. How many companies have people with NLP as a specialty, and maybe you need to now start to think about OCR as well, because sometimes you need to actually digitize these documents before you can run them through your NLP, and that OCR expertise and the computer vision that’s required to get OCR to work at a very high degree of accuracy is very expensive, and so that’s why I say the COE model doesn’t make sense, unless you’re a large enterprise. Excuse me, what I would add to just provide additional color, some of the largest enterprises in the world don’t need to have a singular COE. They could afford to have COEs by function as long as those COEs are operating as a federation, where the sharing, the learning, leveraging the same technologies so you can take advantage of your scale is still sort of core to that federated model. For everyone else, the best bet is to partner with a third-party advisory house like EY or KPMG to consider systems integrators, or look for a company like Roots Automation, which provides Digital Coworkers as a service, where you physically don’t have to worry about anything other than, “Here’s my business process. Can I leverage one of your bots to run it?”
The last thing I’d say, Guy, on that particular topic, what we found in our research is that the magic number for needing to have your own COE is if you believe you can automate 35 processes at your company, then it makes sense for you to have a COE because you can get to break even on that investment over the course of a few, short years.
Guy Nadivi: Very interesting. I’ve never heard that specific number. Chaz, you write a lot of really good articles, and in another of your recently published ones, you talk about the importance of enterprises upskilling their employees. That’s actually a topic we’ve spoken about quite a bit on this podcast. Let’s say I’m one of those employees, and let’s also assume that I want to acquire the skills needed to ride this growing automation wave that’s digitally transforming everything. Regardless of whether my organization pays for my upskilling or I have to pay for it myself, what are the top skills I should acquire to take advantage of job opportunities in automation?
Chaz Perera: I think it is wise for everyone, regardless of the industry they work in to try to learn to code in some language. There’s a mindset, a way of thinking through logic that coding brings forward, and I think that’s important. If you’ve never coded before, learn to code in some language, and it doesn’t have to be deep understanding of it, but just understanding the basics. If you want to be in the world of automation and more specifically in the world of RPA, it makes a ton of sense to try to learn using one of the base, more common RPA platforms like a UiPath, or a Blue Prism, or an Automation Anywhere. I believe all the RPA software providers out there for the most part, at least, provide a community edition or some sort of free edition that you can use to learn. Then, I would say that, thinking about my answer to an earlier question, Guy, around machine learning, and specifically computer vision and NLP, I’m not saying you need to become a data scientist, but I do think it’s important you understand what the art and the science of data science actually is, so that you can speak to it with some level of understanding so that when you are sitting with a data scientist and trying to solve some of these complex problems, you’re able to work off of a single dictionary, and you’re not sort of two or three steps behind. I’ll quickly add one more thing that sort of comes to mind, Guy. In the world of automation, it’s really important to remember there are humans at the end of every single transaction, and so change management has to be a skill that we all have. Automation technologies should be scary. They’re not intended to be scary, but the way you can make them scary is by not being transparent, by not talking about how these technologies can really help to improve the experience for customers, help to advance business objectives. If you introduce this stuff without a little bit of fanfare, without the right change management, that’s when people start to worry about their jobs, and so I would say also focus heavily on change management.
Guy Nadivi: Okay, so let’s talk about the humans in those automation transactions. Your company, Roots Automation specializes in Digital Coworkers, and yet, that’s a term I imagine might create some anxiety in people, and perhaps even trigger some resistance from employees fearing job loss, or radical changes to their job. On this podcast, we refer to that kind of resistance as robophobia, and it’s been known to create friction for enterprises deploying automation. Chaz, what would you say to someone experiencing robophobia at the thought of working alongside a Digital Coworker?
Chaz Perera: We were very careful when we thought about what Roots Automation’s product would be and how we would offer it to the world. The reason we chose the term, Digital Coworker, the emphasis on “Co”, is because we wanted people to recognize that our bots are not simply there to take over their work. Our bots are there to be one, an extension of their team, two, to essentially step in and do the types of work that people typically don’t enjoy doing, don’t get much satisfaction from, and by being able to engage in that more mundane and rote work are, you as a team member of this Digital Coworker, you should now have the ability to engage with customers, to work on projects, to do things that we all hope will create more value for the company than more of the transactional work that bots or Digital Coworkers are just great at. All that being said, transparency is really important, and so when we are talking to customers about implementing one of our Digital Coworkers, we think it’s really important that they talk about where these coworkers fit in on the team, the types of work that the Digital Coworkers are going to be doing, the types of work that the humans will now start to do. Fundamentally, in our platform, Guy, you talked about how we provide this self-learning bot. We don’t pretend as though our bots on day one will be able to do all the work that a human does today, and we also expect that they will be imperfect in that exercise, and so what we encourage is an interactivity between humans and bots, and in our product, that means that as the bots come across things they’re not sure about. If they’ve come across data that they’re not sure about or data that’s missing, or they can’t triangulate data, what they will do is they’ll stop, and they’ll actually ask one of their teammates, the human on the team, a set of questions. Through that interactivity, the bots are starting to learn. What we found is that our customer’s employees, the people that are teammates to these Digital Coworkers, really enjoy that experience. What it’s allowed us to do is to dial down very naturally some of that fear, some of that trepidation that people have had.
When our Digital Coworkers are introduced in one part of an organization and people start to get excited and start to feel like these bots are really part of a team, they naturally start to talk to other parts of the organization about the experiences they’re having, the excitement they’re having, and it helps to, again, dial down some of those fears, and so nothing is better than one employee talking to another to say, “Hey, it’s not what you think it is.”
Guy Nadivi: Chaz, can you share with us some of the outcomes, and particularly the ones that created that excitement from Digital Coworker deployments you’ve overseen?
Chaz Perera: Yeah, absolutely. Typically, what we are able to do for our customer, because we provide these pre-trained Digital Coworkers, what we’re able to provide is a coworker that’s ready to work in a customer’s environment typically between three to six weeks. Because of their ability to learn and engage, they get to productivity quite quickly, and as a result, our customers are seeing a break-even on that investment in as little as five months. If you extrapolate the benefit, because one of our Digital Coworkers is as effective as four to eight people at a company or four to eight people on a team, what we’re seeing is our Digital Coworkers will get a company to about a 250% ROI over the course of a five-year cost benefit analysis. What’s also interesting, Guy, what we’re seeing that’s less to do with the financials and much more to do with the feeling on the ground at a company, our customer’s employees really do endear themselves to our bots. They give them names, they give them personalities, they talk about them as though they’re real people. In fact, we regularly get emails from customers saying, “Hey, any chance Roxy could do this? Any chance Claire could do that?” That is probably the best indicator of success, that we know we can make the CFO happy by getting them to the financial value, but making sure that the line staff and the line managers are just as excited is really what we’re striving to do.
Guy Nadivi: Personalization is very interesting. Chaz, given some of the radical changes to the way people have worked since the start of the COVID-19 pandemic, what role do you envision Digital Coworkers can play going forward?
Chaz Perera: I hope that because of the pandemic, companies recognize how critical their human workers are to keeping customers apprised of what’s happening at the company, keeping customers highly engaged, keeping people loyal to brands and how important employees are to solving those substantive problems that customers often have. And so really, I hope that Digital Coworkers are enabling that by continuing to leverage these bots to not just handle your low complexity and mundane tasks, but because in the world of intelligent automation bots are starting to learn that they can move beyond the more rote tasks that companies often use RPA to do, and start to move towards the tasks that are slightly less sort of defined and structured, because then, you really can free your people to focus almost entirely on engaging with your customer, having conversations with them. Ultimately, that’s the thing that will allow you to build your brand and keep customers on your books in perpetuity.
Guy Nadivi: Chaz, as you know, the pace of innovation in our field can leave your head spinning, given some of the advances in automation, AI, and other digitally transforming technologies, but I’ll ask you this question anyways. What do you envision will be some of the biggest disruptions we’ll see in the next one to three years with respect to Digital Coworkers?
Chaz Perera: I think a couple things. The first, in the context of process mining, I think a lot of the process mining technology we have today is very good at developing a solution designed to document, providing you with that process map, but I imagine that over the next few years, we will be able to go from an exercise in process mining to an actual working bot, this concept of no code actually occurring in the world of bots, and so that would be a huge leap in terms of getting to value quickly, simplifying the exercise of developing and maintaining what are quite complex technologies. I think that would be a fantastic shift in something I do see coming. The second thing would be GPT-3, starting to get bots to not just be able to converse more naturally, but read more naturally. That will allow these bots, as I was sort of talking to earlier, start to move beyond the more mundane, rote, standardized tasks, and start to move into those tasks that require more and more judgment. Then lastly, and this is something we pride ourselves on at Roots, when we talk about Digital Coworkers, we want our coworkers to have very human-like qualities, and so today, our bots learn, our bots communicate, and they can do that on Slack, they can do that on email, whatever it might be, and our bots have this ability to sort of anticipate things. In our parlance, two examples of that might be that our bots, Guy, might recognize that you have your cup of coffee at 9:15 every morning, so that’s not a good time to bother you, come ask me questions at 10:00 instead because you’re more likely to provide a good answer. That’s an example of how we see bots being more human-like. What I hope to see over the next few years is that bots, our Digital Coworkers start to anticipate more. Imagine a Digital Coworker participating in your daily huddle, listening in on the conversation, hearing that the manager who’s leading that team saying things like, “These are the 10 things that are going to be the priority for the day.” “Everything else comes second,” and the Digital Coworkers actively reprioritizing work as a result of what it’s hearing. That’s how we get to a more cohesive office environment that has this happy balance between humans and Digital Coworkers, and certainly that’s what we’re striving for.
Guy Nadivi: Intriguing. Chaz, for the CIOs, CTOs and other IT executives listening in, what is the one big must-have piece of advice you’d like them to take away from our discussion with regards to implementing Digital Coworkers at their organization?
Chaz Perera: It’s interesting, I’ve heard a lot of IT leaders at companies of all sizes talk about agile, and that being the way they deploy technology quickly and create value quickly, but what’s often missing in those conversations are business people, so the IT part of the organization has adopted agile, is trying to move fast, but they’re not doing as good a job educating their business counterparts about how they need to operate, they need to contribute, how they need to learn in this agile-operating environment. You can’t have a true agile environment unless all parties are at the table with an equal understanding and an equal ability to contribute, so I would say that is something, thematically speaking, I see often that some people are left behind in that conversation. Then, the other thing I’d sort of throw out there, bots don’t fail gracefully, and so I would encourage IT leaders across an organization to think long and hard about the interactivity layer that needs to exist between humans and bots so that you don’t leave a human stranded. The thing that, going back to that original question, Guy, you asked about, “Why Roots Automation?,” one of the things that absolutely drove me nuts about the deployments of robotics at AIG, we would give our business users an Excel spreadsheet at the end of the day that said, “Here are all the things the bots did. Here are all the things the bots couldn’t do.” “For whatever reason, you need to pick it up.” That is not a great user experience. That is not a great customer experience, and so I would say spend time thinking about that interactivity layer and how you can create a better human experience for people that now have to work with bots.
Guy Nadivi: Setting realistic expectations, always a good idea. All right. Looks like that’s all the time we have for on this episode of Intelligent Automation Radio. Chaz, we love having automation innovators on the show, and it was great hearing your thoughts about the innovative role Digital Coworkers will play in the future of work. Thank you so much for coming onto the podcast.
Chaz Perera: Guy, it was a pleasure, and again, thank you for the opportunity to share my two cents at least with your audience.
Guy Nadivi: Chaz Perera, Co-Founder and CEO of Roots Automation. Thank you for listening, everyone, and remember, don’t hesitate, automate.