Kevin Montalbo

Welcome to Episode 70 of the Coding Over Cocktails podcast. My name is Kevin Montalbo, and joining us is Toro Cloud CEO and Founder, David Brown. Good day, David!

David Brown 

Hi there, Kevin!

Kevin Montalbo

All right. On this edition of Coding Over Cocktails, we're going to have another Tech Smackdown, putting forth two technologies, concepts or architectural styles in a friendly sparring match. Today we're debating between RPA vs. BPA, robotic process automation versus business process automation. Which approach to automation should we apply to our organisations?

Let me introduce our guests. On the RPA corner we have an advisor and board member to startups. He’s the author of “Artificial Intelligence Basics: A Non-Technical Introduction” and “The Robotic Process Automation Handbook: A Guide to Implementing RPA Systems”. He is also a developer of courses for PluralSight, including “UiPath: A Big Picture”. Besides his writings, he has founded several high-tech companies which include WebIPO, BizEquity, and Hypermart.net, which was sold to InfoSpace.

Joining us today to represent RPA is Tom Taulli. Hi Tom, welcome to the show!

Tom Taulli 

Hi! Thanks very much for having me.

Kevin Montalbo

Alright now, on the other corner, we have a strategic advisor, author and podcaster on Intelligent Automation.
He has been helping organisations to discover, develop, and deploy automation. He was recently the head of RPA consultancy at Fujitsu America, where he enabled the Fujitsu Digital Workforce offering. He also recently spoke about RPA vs BPA at the Business Process Management Institute.

Representing the BPA side is Nandan Mullakara. Hi Nandan, welcome to the show!

Nandan Mullakara: 

Hi, thanks Kevin! Thanks for having me. Hi, David!

David Brown 

Hi Nandan, thanks for joining us! I'd like to start off by doing some definitions if you don't mind, because I can understand why there is some confusion in the marketplace with the difference between BPA and RPA because we're talking about automation of processes in both cases. So I would really want to dive down into what the definition of each is and then we'll dive into some use cases. Tom, maybe you can start us off describing what RPA is and then we'll get into some use cases later.

Tom Taulli 

Sure, yeah. In my book The RPA Handbook, I have a section in there, the definition of RPA and I quoted different companies –  UiPath, Automation Anywhere – they each had different definitions of what they thought RPA was. So, there isn't a clear cut definition of RPA. Like if you had a relational database, it's pretty clear what a relational database is. RPA is more of a concept and it has evolved over time and it's been around for about 20 years and it primarily came about from just plain screen scraping.You know, stuff is on the screen, take it, cut and paste it, put it somewhere else and that saves you time. That automates a task. Not really a process but a task. 
But over the years, especially with companies like UiPath and so forth, they've got more sophisticated with it using computer vision and artificial intelligence and all these other technologies to make it more powerful. But the idea, primarily still, is that you might have what is known as “automation of interface.” So, someone spends a lot of their time working with CRM or ERP. And they're literally just cutting and pasting stuff all day long. Not the best use of your employees. 

In UiPath, you could create a bot and it just does it for you. And then there's other processes that are unattended automation or it happens in the background, so it just kind of does things while you're doing things. As a customer support person is talking, the RPA can provide you with some information to resolve the conversation. Or, you know, another big use case with RPA is just invoicing. Invoices are emailed to the company. It's scanned. There's a lot of  OCR with RPA and you know, figure out if this invoice looks correct. If so, we'll just process and automate it. If not, maybe we need to get someone to sign off on it and so forth. So, basically RPA is software that creates a bot to automate a task, not really a process. And it usually is someone sitting at a desk and just making it so they don't do tedious and repetitive work.

David Brown 

Okay well, Nandan, if RPA is task-orientated as opposed to process-orientated, maybe you can run us through how business process automation differs?

Nandan Mullakara 

Yeah, so BPA is more strategic than tactical. Right? So, Tom mentioned automating processes, automating tasks by mimicking actions at the computer. Whereas BPA does implement automation by carrying out the multiple steps of the process. And many people call it a long running process. 

So, a process where you have, let's say, in order to cache where you have multiple steps including people, systems, and different automation technologies. The automation technology is where RPA comes in, so the RPA orchestration is probably done by a BPA tool. And then the RPA does parts of it. So, how Gartner calls it is that the BPA runs the business process, whereas the RPA is like calling the business process. Right? So, it's more holistic. BPA is more holistic and does the end-to-end process automation.

And so, this is usually preceded by an organisation first analysing the process and understanding what's to be done to automate an interim process. Whereas RPA many times ends up being tactical and “just let me go and address that particular task.” Though now, everyone is stepping back and looking at the strategic picture and then bringing in the appropriate technology. 

So, what I would also add is that these are both complementary technologies as you saw. The BPA does end-to-end and the RPA does the tasks, but both are complementary and both are working together in automating a process.

David Brown

Interesting. So, is it fair to say that RPA is a subset of BPA?

Tom Taulli

It could be, yeah. But I also think it could be like your first step into automation is RPA. You know, you start with the task, you start with some simple stuff and as you build those automation muscles and you understand how automation works within your organisation. You get a little bit more ambitious and you think , “Maybe we should automate the whole process and then we need to bring consultants in.” I think one of the reasons RPA has grown so fast is it's easy to get started with and you get results.

So, I can automate someone's desktop work using a CRM and get almost immediate results. Automating the process takes time and thought. A lot of the time, projects easily just fail if it's not planned correctly. And BPA or whatever you want to call it, it's been around a lot longer than RPA. And it's been more of a slow growth business because it takes a lot of effort and investment from the company to make it work.

David Brown

What I'm trying to understand is, in the business process, you have a model. Typically, when I think of it, I think of a flow chart, with steps in a flow chart of a business model, with each step in that flow chart as effectively a task in a process. So, that's where I said, “Can RPA be a subset of BPA?” Because it seems to me like you are calling tasks in a business process. 
So, if in a business process, it requires me to, for example, read and write to a database. Now that has traditionally been one step in a business workflow or business model. But is that RPA? Or is RPA only when a bot comes involved, where you're scanning an invoice or translating or something like that?

So, how can we differentiate between a step in a business model and business process between what is an RPA step and what would just be a regular step in a business process?

Nandan Mullakara

Yeah, so what I would say is that RPA comes in where traditionally BPA has lacked. Right? And so like Tom said, the whole BPA space has been getting a bad rap because it takes a lot of time for you to implement this thing. Whereas RPA is a quick win, and you can do it right? Now, the thing with RPA is that it leads with UI integration or it leads with using the user interface, which is great to do things quickly. 
But [this] also has its downsides because you are using a screen to get across your data and the screen may change, or the application may change or upgrade. And so RPA has its downsides and its upsides too. So where RPA is strong, like whether it's using it for a legacy system or where there's no API available, RPA is the best fit for automating those things. 
And, mind you, there are so many such situations and so that's where RPA has been becoming very popular. But, you know, from an API perspective, from involving a human in the loop, from involving other systems, BPA is still good. So, what has happened now is that RPA actually made it easier, made it more, if I can say, sexy to automate.

And that has made it possible for the reemergence of technologies like BPM, which is now being rebranded as iBPMS or intelligent BPM. And BPA also has seen an emergence and you know, vendors have been adding things like low-code and also AI and process mining and things like that to make it more useful. So, it's become more and more as we proceed with what Gartner calls “hyperautomation”. We are seeing that all of these complementary technologies are actually helping people. Beyond all these terms of BPA, RPA and the fight in the turf wars. I think there's a lot of value in automation as a whole, whatever term you call it. And that resurgence has been dead by RPA. But BPM and BPM has its role. As I said, they're all complementary roles.

David Brown

Well, Tom, you mentioned that RPA has been around for 20 odd years with screen scraping and the like. I guess a lot of us think of it as a fairly modern thing. The hype around that has evolved more recently. What has caused this hypergrowth of RPA?

Tom Taulli

Well, you know, it used to be called screen scraping. And what is so very sophisticated about screen scraping? So I think, how you describe your category or industry is very important. If you describe it as something that sounds boring and not too interesting and not sexy, you know, people don't want to buy it. So, in 2012 the ambassador or the advocate or whatever you want to call it, at Blue Prism said “we need to rebrand the industry” and he came up with RPA, robotic process automation. And everyone thought, “That sounds great. Where do I get that?”

Funny thing is it wasn't robotic because there's no robots and it really wasn't a process because it's more task-oriented. So, I didn't even really describe what it was doing, at least in the early days. But they got it right in the sense that they made something sound really exciting. So, I think half of what makes technology rise above the crowd and the noise and so forth, is just having a good term or phrase to a good story, you know [like] Steven Jobs. “You need a good story to tell.”

But I think what really has been the key here with RPA and what has made it a success is UiPath. UiPath was a company that was on the verge of bankruptcy in 2012. It all kind of happened during that same period of time in 2012.

But what happened was that there's the BPO industry, business process outsourcing industry, usually offshore. You take a lot of stuff and you go to low-cost areas and you do that and these BPO organisations handle that. 
And at this turn in 2012, these organisations are looking at ways for them to automate their operations and, you know, to get more margin for profit, because it's a really tough business. And they just kind of stumbled upon this UiPath software and screen scraping and realised this could be a big help for our business. And Daniel Dines gets on an aeroplane to India and he spends three or four months there with this client, this BPO client and they really reworked the system in a way that understands the typical corporate tasks that need to be done. And I think a lot of it had to do with that product-market fit.

You know, they had a customer pain point. They had software. Initially, it really didn't really work perfectly for it. But by talking to that customer, they got closer and closer to that fit and when the fit happened. I mean I think when they transitioned to more of this RPA way of doing things, I think in that first year, revenues went from five million to 200 million within a year for UiPath and the rest is history. I mean it is a dominant player. So, I think a lot of it is that they were smart to work with some of these BPOs. And then eventually, consulting firms, understanding the client pain points and developing software that can solve those pain points.

And it's not necessarily the most sophisticated technology that wins either. You know these are pretty simple bots. And it's all Windows. It was all Windows software. It wasn't even in the cloud. People kept calling UiPath a cloud company even though it's still using Windows software. So, you know, they realised that a lot of companies have Windows software, the Windows tech stack, so build it using that and make it so their technology works more efficiently in a more automated fashion.

So, I mean there's a lot of factors why it became a success and why RPA is a high growth industry. But I think a lot of it had to do with that product market fit about 10 years ago. And then UiPath and some of these other companies built on that.

David Brown

How important has been the emergence of machine learning algorithms to RPA processes? Nandan, do you want to take that?

Nandan Mullakara

Yeah. Okay. So, machine learning or AI has been a big addition to RPA and automation in general, right? So, where machine learning helps is in specific areas. What an RPA does, an automation in general does, is automating work and moving data from one part to another. And most of the data is still stuck in documents. You know, unstructured data in documents now, most of them are digital. They are in PDF formats and things like that. But still, to read the data, you need to read the data out of the document. It's not in a database. So, that's where, let's say for example, AI, ML helps, it can extract the data out of the documents so that the automation can use it. So, that's one use of it. 
The other use of it has been in terms of computer vision. So, the computer vision helps the RPAs see the screen better and extract the data from it better. And also, chatbots use something of an AI called natural language processing. So, you can understand, supposing you are at a bank and you're chatting, they want to understand what you're saying. So, natural language processing is done through AI. So, all these AI-related technologies are helping our piano and automation in general.

David Brown

I liked that you mentioned the exchange of data between systems. I'd like to dive into that a bit more. So, Nandan, in a business process automation where you want to use an RPA process as a task within that larger business process, how are you exchanging data between that task and the rest of the business process? You mentioned, for example, a machine learning algorithm which would create structured data out of unstructured data and some other system may be dependent or looking for that data. So, is the process exposing that as an API? Is there something like that that the business process can then leverage?

Nandan Mullakara

So, RPA… what it does is it extracts the data and then it exchanges this data with other applications. Now that I'm thinking about… it doesn't expose it as an API. So it basically would hand off the data. Suppose there is a long running process and you need the data and you need to take the data, pull out the data, and put it in an ERP system. So now there are multiple ways to do it. The RPA itself can put it in the ERP through screens. 
It can also use APIs. And also it can hand it over to a long running process within BPA or BPM, which can take the data and then it, in turn, calls an API or you know it does what it needs to do to call a system like an SAP to put in the data. But where a BPA process comes in handy is where it is long running because it maintains the state. And so, let's say if you need an approval, it can send for an approval and wait for the approval coming in before it goes to the next step of let's say, sending out an email that this transaction was successful. So, you know the RPA works seamlessly with the BPA process in implementing or in executing or automating an end-to-end process

David Brown

Tom, is there anything you'd like to add to that?

Tom Taulli

Yeah, I think that he did a really good job explaining that. Also, a thing, going back to the AI question. You know, in data too. So these ERP systems, CRM, you know, they create all these log files, different transactions. And there's just tons of it that's created every day and there's a lot of that within the organisation. Now, the interesting thing about data, there's a huge explosion of data and most organisations get and use a tiny percentage of that data. There's so much data created that companies don't even leverage on it. They don't even know how or what to make of it. So, a lot of data just goes wasted. 
But with these log files, you can apply machine learning to that, to understand what your current processes are in your organisation. And this has led to this category called process mining, which has grown very quickly, and has grown very strongly over the last 10, 20 years. And there's a lot of open source solutions and some proprietary solutions. UiPath has its own. They bought a company for that. And what it does is it maps out different ways your organisation operates and how they handle certain workflows and it can find them and it'll say, you know, there's bottlenecks here. You can reduce the bottlenecks here and make this process quicker. You can make suggestions and it's not just a one time thing. It helps you overtime because companies' processes and ways of doing things are an evolution. They change and so the process mining, looking at those log files, we'll take advantage of that and see better ways to make your organisation faster.

So, I think that's the other part of RPA. And it's also BPA. It’s to try to take more and more of this data and understand your current processes and provide recommendations on how to make them better.

David Brown

Yeah, great. We've talked a lot about RPA. I wonder if we can dive into more of the use cases of BPA because Nandan, you mentioned maintaining state is an important element of a business process automation and I think that's probably a big differentiator from robotic process automation, task-orientated events. So, can you run us through that? And you mentioned, for example, approval processes. Can you run us through some use cases, typical use case for business process automation? And then maybe we can talk about where RPA might sit in some of those use cases?

Nandan Mullakara

So, BPA is used for multiple automation. And there is some overlap of where BPA and RPA is used because, BPA has been a technology which existed even before RPA. And it has taken care of the simple workflow automation earlier. And the whole idea of BPA was to remove people from the process and that's kind of what RPA is also aiming for. So, BPA and RPA overlap a bit. And in the sense of the simple workflow automation, both can do it. RPA does it through a UI integration whereas BPA does it mostly with APIs. And you know, automating the workflow with the systems by invoking the systems.

Now the other place where BPA is involved, which I've been talking about, is complex business process automation which is long running processes that require you to persist the state while you wait for the other tasks to complete. So, this process is usually cut across organisational boundaries and usually end-to-end processes so things like order to cash procure to pay. These are all end-to-end processes and complex business processes, processes that BPA is used for automation.

And the third place where BPA is generally used is for case management. So you know, case management is where there's a unique situation with complex interactions: people, business applications, regulations and things like that. So, case management is another place where BPA is used quite a lot. For example, in healthcare, to manage a certain case, a complete end-to-end case. But the whole thing is that it does more of an end-to-end process automation.

David Brown

Tom, is there anything you want to add to that?

Tom Taulli

No, I think that's a good explanation. But I also would say too that, this is especially for RPA, the impact of COVID has been significant because companies all of a sudden had processes come under pressure. They may have had employees that are remote or maybe their employees you know had to be laid off or what have you and they had to find ways to automate because they just lacked the people and we're here.

In the United States, we have very low unemployment and it's just hard to find people to hire. And so, I think what's happened is that automation is filling that void, but it was definitely accelerated with COVID. But I do think what's going to happen is that because it happened so quickly and a lot of these companies just rushed to buy and implement this software, now, they have a lot of spaghetti code and legacy, just technical debt built up in a very short time. Just like we built up a lot of financial debt in the last couple of years, we've built up a lot of technical debt in the past couple of years and I think the next few years it's going to be more about figuring out, “Okay, now what do we do with this?” Because we probably built a little Frankenstein inclusion and you know, we may not be getting the kinds of outcomes we should be.

So, because things are moving so fast, we couldn't think about that in the last few years. But now as the pandemic is fading away, these companies can do that. So I think that for consultants or think people in this industry, I think that's going to be another opportunity because I think there's a lot of implementations that we're probably not done properly.

David Brown

Yeah, that's an interesting perspective, that obviously the pandemic accelerated a lot of digital transformation efforts. But you're saying that as a result, we've ended up with a lot of technical debt and that's going to be a lot of cleanup process going on now. Does that mean that maybe we're not going to be innovating and transforming as much but perhaps refining our processes and doing some cleanup?

Tom Taulli

I think that that's correct and I think that's kind of like, the stock market and the tech stocks going down is because companies have bought already so much of the software already that you know, I don't think there's a lot of appetite to buy more of the software because they’ve got to figure out what to do with what they already just bought in the past two years. So I think yeah, in terms of the digital transformation, those types of efforts, I think it's probably more about maintenance and reworking and restructuring in the next few years.

David Brown

I'd like to conclude with the common misconception or otherwise that RPA and BPA are there to take people's jobs away from them. So, Nandan, you pretty much said the same thing. The aim here is just to reduce the amount of people required to run a process. Tom, you had an interesting perspective that actually, people weren't available and there's a demand for labour which can't be fulfilled. And so this is filling the gap.

So, I'd like to conclude with both of your perspectives on, is BPA and RPA going to make people redundant? Tom, maybe you can start.

Tom Taulli

Yeah, I think it depends on what time frame you're looking at. Short term, I don't think so. Again, the next few years will probably be my opinion about cleanup and that usually requires a lot of knowledge and I don't think AI is at that point where we can have this digital consultant come in and a digital eccentric person come in and figure these things out. I don't think that that's going to happen realistically.

So, I don't think automation is going to get rid of people. I think it's going to be recessions or the usual suspects that get rid of people. And I don't think automation at this point. Although, I think companies do want to. They don't want to say it, but I do think they like to cut costs and over the long term, as it gets tougher when labour gets more expensive and labour gets harder to find and qualified labour gets harder to find, long term, companies are going to be looking at making more and more investments in automation and to me, that will impact job opportunities.

And I think AI will get to the point where it's going to start replacing, maybe 10 years from now, CPAs or attorneys. I was at an event over the weekend. It was one of the first events I've been to since the pandemic. There's an attorney there and I said, why can't I just have a piece of software that when I get a contract, I just upload it and it just tells me the red flags, the things I can negotiate, all these things that my attorney would tell me.

He goes, “No, I don't think you can do that.” I said, “I didn't ask you whether you want it or not. I just, I want it. Obviously you don't want it because you bill me lots of money.” So I wanted to, you know, I could just pay $30 a month and I get my attorney and maybe I can get 90% of whatever I need within a second. And then for the crazy scenarios that I have to think about, I can get an attorney for that. I think that's going to make a huge impact. I don't think it's today. I think it's years down the road, but I do think it will happen and I think it's going to have a major societal impact. I don't think we're prepared to deal with that.

David Brown

Right. Nandan, what are your thoughts on that matter?

Nandan Mullakara

Yeah. So, I just want to clarify that what I said was that BPA’s intention was to replace people and that's where it had a bad rap, right? I think that messaging with BPA, as well as RPA has been wrong. It’s that you replace people. So, what you actually automate with RPA for example is tasks – it's not jobs. It's not an entire job even if it's a process. It's not jobs, so we still need people.

“I think that messaging with BPA, as well as RPA has been wrong: it’s that you "replace people". So, what you actually automate with RPA for example is tasks – it's not jobs. It's not an entire job even if it's a process. It's not jobs, so we still need people.

But, let's not sugarcoat it. We can reduce or we can improve the efficiency. So over a period of time, as Tom said, we will see efficiencies where we can have a lesser number of people working. And it's happening to some extent already. And the good thing about that is that the jobs that are getting automated are the repetitive road jobs that people anywhere didn't like to do. And you can see this happening with the great resignation. People don't want to do this. You know, these kinds of tasks or jobs which are menial.

And so actually right now, there is a shortage of labour in spite of automation. So what it shows is that it will come from both sides. We will have people not taking up those jobs, for example the Millennials or the new Gen Z, they will not want to do these repetitive tasks. So like Tom said, there is efficiency, so why would I need to pay an attorney a full fee when you know you can do a lot of it through automation?

So, I think it will all come together in a nice manner where you would actually have machines or boards or automation or you may call it doing the work that people don't like or things that bots can do better and there will be jobs which will open up for people which are more aligned to what people can do. So, I think we were doing a lot of bot work which will be taken off and we will have more human work going forward to interesting times.

David Brown

Thanks guys very much for the debate today! I'd like to inform our listeners of where they can follow you both on social media and the things you're writing and talking about. Nandan, what are the best channels to follow?

Nandan Mullakara

So the best place is LinkedIn. That's where I am usually found. You can reach me over chat, connect with me. You can also find me on Twitter though it's not such a big following, it’s my first name and last name, “@NandanMullakara.” And I do have a blog website. It's Nandan.info, where you can find all the places where you can connect with me. And thanks for having me.

David Brown

Thanks, Nandan. And Tom, where can our listeners follow you?

Tom Taulli

I'm on Twitter. Now, who knows what will happen to Twitter in the next few months after Elon Musk gets there? But my handle is “@ttaulli.” And then the same as with my LinkedIn, you can find me there and then I do have my own website at TomTaulli.com. Any of those. I have my website, there's a contact form on there and you can just contact me that way. I'll get the message and thanks again. It was great being part of this discussion. It was a lot of fun.

David Brown

Good stuff. Thanks guys! I think it was really informative to highlight the differences and synergies between BPA and RPA. Thanks for your time!