Episode: 332 |
John Joseph :
Robotic Process Automation:


John Joseph

Robotic Process Automation

Show Notes

John Joseph is a Co-Founder of White Swan Automation, a firm that helps small and medium business implement robotic process automation.

Learn more at:


One weekly email with bonus materials and summaries of each new episode:

Will Bachman 00:01
Hello, and welcome to Unleashed the show that explores how to thrive as an independent professional. I’m your host Will Bachman. And I’m here today with john Joseph, who is one of the cofounders of white swan automation. JOHN, welcome to the show.

John Joseph 00:18
Hey, thanks for having me here. It’s gonna be great. I love it. I love your podcast.

Will Bachman 00:23
Thank you. So john, you were at Lockheed Martin, you with a couple friends are experts in robotic process automation. And you thought there was space in the market for a firm that would do robotic products, process automation for small and medium sized businesses. Tell me a little bit about the type of work that your firm is doing and the type of clients that you’re serving?

John Joseph 00:51
Yeah, love to. So a lot of our work is focused on mimicking human behavior. And that could be as simple as reading an invoice. It could also mean, taking data from one website to another website, or even making accounting entries into an accounting system, right? The complexity, that the robots that we use are able to perform and mimic certain human behavior has a lot to do with, honestly, what the human is trying to accomplish, and how we can flow that out. So although it is logic based, we’re able to get a little bit complex when it comes to making some of the decisions right, as long as they’re, you know, for castable. And there’s a certain level of thresholds that the company would like to have when it comes to having robots make decisions for them. But I think when we see a lot of our customers, a lot of them are, you know, small to mid range, maybe 20 to 100 employees. They’re, you know, not fortune 500 companies, but they’re companies that are that are suffering from a lot of intensive, repetitive labor, what we like to call pain and suffering in the office have 1000s and 1000s of invoices that are coming in from hundreds of different vendors, maybe if the office is like a shared service model, right. And they have employees that are manually reading these invoices, having to make sure they have the store number correct. And they’re having to make accounting entries somewhere to pay these vendors on time. It can get, there can be a lot of pain and suffering in the office, right. And through, you know, robotics, process automation, and then also OCR technology, we’re able to go in there and make a completely new workflow for them from the beginning to the end, to where it’s, you know, 95% hands off experience to where by the time they scan that invoice, or it comes in electronically, it could be in the accounting system the same day, with with minimal human effort. And it’s been great to see a lot of the capacity that these companies have been able to open up because they can retrain their employees to do a lot more what we would call more analytical work instead of just, you know, looking at a mundane, repetitive process. Now, they’re actually more more analyzing forecasts for the company or analyzing where they want to go next in the sector. And using a lot of that brainpower and education that their staff already has.

Will Bachman 03:33
Great. So let’s take a backup just a minute and robotic process automation. It doesn’t have anything to do with sort of the Ford assembly line and actual physical robots with the arms going out and putting in the rivet. Tell us a little bit about the history of robotic process automation. And some of the, you know, developments in that field. For folks that haven’t heard that term before.

John Joseph 04:03
Yeah, so you know, if, if I don’t, you know, a lot of your audience probably uses Excel, right. And Excel has been around since about 1987, after Lotus 123. And such, one of the things Excel brought to the table is macros. And a lot of people probably have experimented with recording macros in Excel and being able to do something repetitively inside their own excel sheet. And if you take that same concept, you move it to a different program that can not only interact with just Excel, but it can interact with the internet, it can interact with multiple accounting systems, it can send out emails, now you’re recording a process or a flow of a workflow. And you’re able to do much more complex things that Excel couldn’t do just by itself anymore. So there’s a lot of different vendors To offer different robots that focus on, of course, different workflows. The two large, the three largest ones right now are blue prism, automation anywhere and UI path. And those three are familiar to us. And each one of those robots have a different kind of scope, I would say that they do a lot better than the others. But that’s kind of the history is that you’re seeing, you’re seeing the same technology that we’ve had within recording macros in Excel, but now to a higher level, where it interacts with multiple systems. And the nice thing is, it’s no longer the IT departments that are maintaining these robots. A lot of it are the functional finance and accounting people who are able to establish these robots and also sustain them. Because the user interface of these robots, you don’t have to learn a whole new language anymore. A lot of it is much more simpler to create.

Will Bachman 06:09
Now, there’s some even more consumer friendly types of robotic process automation. I mean, Xavier comes to mind which at Umbrex, we use for some tasks, as well as even simpler if this than that, or IFTTT. Tell us a little bit about the various, you know, consumer friendly process automation tools that, you know, consultants, my listeners of the show might want to get started with if they’re if they have some processing on automate, what are some of the ones out there,

John Joseph 06:45
I would probably suggest for the number one overall, I would say, from my own personal experiences, UiPath is a great one to I guess, get your feet wet. Maybe look up some YouTube videos kind of look at what UiPath robots can do. Their licensing model is pretty easy and straightforward. Automation anywhere is another one that is pretty nice with their licensing. And also, you can get your feet wet with that. But I think UI path has a game up because UI path has their own kind of Academy and certification all in like one little place so that from the ground level, you can take certain roles that you might want to have maintaining that robot or sustaining it or develop it, and then go through their courses. And it’s all user friendly, very easy to learn. And I think for most people out there, UI path would be a good connection with the systems that they have already been placed with our IT infrastructure.

Will Bachman 07:53
Yeah, an order of magnitude, how much does it cost? If you wanted to use UI path? And maybe for listening to the show, many of them are independent consultants, you know, how much would it cost to get just one license for, for one person if you want to automate some of your own tests?

John Joseph 08:11
Yeah, so I think with, you know, with different robotic vendors, right? The, the licensing is going to depend on how much you buy up front. With the different licenses of the would probably say, robot. But you know, with UI path, I would say, with a robot working per year for a license, it’s probably depending on your volume around 4000 or so. But that’s a 24, seven, capacity of a resource that you’re able to use. So what we see a lot of times is don’t think of a workflow that you’re trying to optimize. And say, Well, I have one workflow, I need one robot. What you should start thinking of is how long will the robot take to complete my workflow. And then if it only takes two hours, that means you have another 22 hours of capacity you can use on that license. So it, it can be multiplied to around one robot is doing 15 to 20 different workflows for you. Every single day. You can schedule it. That’s kind of where the cost structure comes in. And the way kind of think about it when you’re when you’re wanting to dip your toe in the water.

Will Bachman 09:37
If someone doesn’t, you know, if they’re running a solo practice, and they, they don’t, you know, need even 24 hours a day. Is there. Is there more like fractional licenses where you can say, look, I just want it for two hours a day. That’s it.

John Joseph 09:52
I think there are some I don’t I don’t know. I know that some robotics companies are looking into that I don’t have anything off the top of my head right now that I can suggest for that one. But it’s just kind of like, just the way it used to be now that you can rent cloud data or server cores, I’m sure that their RPA group is going to have something like that as a service offering.

Will Bachman 10:16
Okay. And for the more consumer focused ones, like Zapier, what are the other ones that are more comparable to Zapier? That might be more in the, I don’t know, 50 or $100, range, something like that per month?

John Joseph 10:31
Well, I would think that automation anywhere would probably be around that kind of price range for some other more simpler products. But I don’t have any other suggestions than that one right now. All right.

Will Bachman 10:48
So let’s walk through a case example of of your firm. And I’d like to get into the details. Now, you can certainly sanitize it. But can you give us an example of a company that you’ve served, maybe just tell us a little bit about what whatever company it is, again, don’t even share their name? And you know, how you how you help them?

John Joseph 11:10
Yeah, so there’s a lot of examples, you just pick a good one here. So you can think of a large company, 100,000 employees or more. We’re talking, you know, billions of dollars when it comes to a lot of their different departments. And they have employees, definitely, you know, 100, plus employees working different types of payroll, hundreds of employees doing different AR, AP, all sorts of things. And what we’ve been able to do is pretty much take a look at their processes and their workflows and their pain points and get a list of different, you know, things that they know that they’re doing manually today. And I think a good example is, if you’re in a corporation, every it group probably has a long list of asks that the functional team is asking for either customizing a program, making it a little bit easier, running reports automatically for them, you know, a lot of those asks are put to the wayside sometimes because of either security concerns the company might have, or there’s just bigger fish to fry. Right. So the functional teams are usually having to just get used to doing something manually, because the IT department is somewhat expensive. And it’s hard to get their time. Well, there was a there’s there’s a small group, you know, within that large organization that had about about 40 employees, that were billing out different vendors that they did work for, and they needed a way to find out what invoices that they need to send out, because according to the contracts, you know, you had a certain you only could build them once every while right, so you had a rolling kind of 30 days to send out an invoice. And what our robots were able to do is start from the very beginning of the process where they would log in, at, say, two in the morning, when no one’s at work, right, they’d log into the accounting system, they would run the specific reports, to then find out which vendors they had to go and Bill out. And then they would take that information and store it in a database and highlight the ones that were the most sensitive or the most high value invoices that they would want humans to take care of. And a lot of that pre work was all done before six in the morning. So now when the team logs on to their computers, all of that pre work is done for them. They can focus on invoicing those customers exactly the way they needed to. And we’re just talking about the invoices that had to be sent manually. For a lot of the invoices that were pre approved with the customer. Those were kind of on a whitelist with our robot to where now the robot just sends those invoices out automatically if they’re pre approved. So the volume of invoices that humans had to process went down by about half and hundreds of hours were saved every week, because of all that pre work and all of that invoices that gets sent out automatically where it used to be very manual. So that’s kind of just one example. And that can be repeated. A lot of different sectors where you Have people doing a lot of manual pre work just to do manual after work? where robots can kind of help and step in for whatever the threshold The company has confidence level wise with their workflow.

Will Bachman 15:12
Okay, what’s another example? Is there another use case other than invoicing and accounting?

John Joseph 15:25
So, sometimes we all want to gather information either because of auditing, or if there’s an auditing, internal audit or an external audit. And sometimes external auditors or internal auditors want a sample size of a certain pieces of information out there in the systems, right? We see that asked for a lot where there might be, you know, an auditor comes out of nowhere, they say, hey, I need 100 of these different parcels from from this different program. Well, you know, we can get a robot set up to where whenever that ask is there, the turnaround time for that can be much faster and much more accurate, because, say, their contract numbers, or it portfolios or customer information, just any, anything that’s already out there hanging around somewhere in your systems on a database somewhere that you have, or even screenshots, a lot of times we have human beings, still manually going out there taking screenshots of different things for audit purposes. You know, we can design robots to go out there and do that automatically. So that when you do have that asked for either corporate internal audit, external audit, or anything that can already be set up to where it’s a quick turnaround time. And before you send that back to whoever asked for, you can always have humans look over what the robot did. But all that manual work that you already know that you had to do, we can get that set up to where you don’t have to do that anymore.

Will Bachman 17:04
It’s amazing. Tell me about the process of one of your engagements. So how do you, you know, is there like a diagnostic? And then is there a standard kind of, you know, three or four steps that you go through phases of one of these projects? Walk me through how a typical project works?

John Joseph 17:26
Yeah, so I think a typical project works when the company realizes there’s enough pain and suffering to where we might want to get rid of it, right? If it’s taking one screenshot a year, no one’s really gonna complain about that. But if you’re having to do that every hour for day, it kind of gets up there. So we get engaged when people realize, hey, we don’t want our employees doing this particular process anymore. We think that there’s better technology out there, but we just don’t know what it is. So what we do is, we usually have a meeting with them, there’s usually a live demo to kind of help us understand the complexity of what they’re looking to automate. And from just looking at the workflow, we’re able to establish what it tool would be best to fit their workflow. But not only that, but their culture, how much how much it infrastructure do they have? What programs are they using? What complexity or what kind of guesswork is needed for this? Are there or is it just kind of, you know, step a to step B to step C, you know, we’re just trying to gauge the complexity. And from that, we can understand what technology would be best for them. And then from there, we write up a business case, we kind of show them, usually a demo of the technology that we’re suggesting. So that they can kind of see with their own eyes, what the future can look like for them. And then from there, we start working on the project. And then we complete the project. And then everyone’s happy. We like to work with companies that definitely want to learn how to do this technology themselves to a lot of the technology we use, you don’t have to be an IT infrastructure to maintain. A lot of these programs are able to be used by accounting by finance people, business managers. And we take the time to help train them so that if there is small adjustments that need to be made, they’re self sufficient, that they can take care of it themselves. And they can grow with any future workflow changes or anything else that comes up that they might want to take a crack up. So that’s kind of how we work.

Will Bachman 19:53
And then these robots, I mean, they’re logging into all these different systems and so forth. So How do I know? People may be concerned about security and passwords and so forth? Talk to me about that.

John Joseph 20:08
Yeah, that’s a good question. So with a lot of different companies, they have like, either an IT team or an HR team that kind of handles login information and passwords. You can look at these robots, the best way to easily set them up is you just pretend they’re almost like a new employee, they have their own ID so that when they logged into these systems, there’s an audit trail, you know exactly what the robots doing at all times, or when it’s signed into something. And then a lot of our robotic programs have what we call vaults where you can store that password for that employee and or for that robot. And anyone who has access to that vault, has an audit trail who logged in to see the password, was it for development, or sustainment? Was it for a production fix or anything like that. So the people who definitely have been making a lot of these robotic process, robotic process automation tools, definitely have security in mind. And these robots aren’t like invisible in the shadows or anything, they’re going to be treated just like if a human was logging on to a system to access information. So many of the IP controls that are already established within these companies, they rarely ever have to change. Just because a robot enters the workflow. It’s usually mimicking the human behavior anyway. So all the it can still still say, Listen,

Will Bachman 21:40
do you have any rules of thumb on when a company should start thinking about, maybe we should automate this? I mean, we can say at one extreme, if, if an employee if one employee has to do something that takes one hour per year, you can pretty firmly say, okay, probably don’t need to automate that is probably not going to be worth the time to pay your firm to come in and automate that. If there was, you know, 20 employees doing the exact same task for 40 hours a week, every week. And it was a very straightforward algorithm, you know, algorithmic thing, then you’d say, well, that’s definitely you should, you know, probably automate that. Absolutely. Where is sort of the place where it’s starts becoming going from, you know, no to Yes. Like, is it? You know, if you have, you know, effectively sort of 10 person hours per week, is that the break point, or you need, if you have 20 employees doing, you know, one task that takes them an hour every day, like, Where should companies think about saying, Okay, now we should think about doing automation?

John Joseph 22:48
Okay, so, if I had to take that very philosophical question and pack it into a rule of thumb, I will, I’ll probably say that when management wants to grow faster than their payroll can handle, that’s when you need to look at automation. And hiring humans to do more work, is something that has traditionally been the solution to more work. But that’s not a good rule of thumb anymore. If you’re ignorant in the ideas of what robotic Process Automation can do, or OCR, I would say, give us a call or go look at some videos at what other companies have done for their own processes. Because as soon as you see the light of how much automation can help, or what it actually can do, and how complex it can get for you. Work workflow wise, you’re going to see that a lot of the things that you just kind of took for granted, and just thought, Well, you know what, as an accounting or finance major, or doing taxes or anything, this is just the way the workplace is. This is how we’ve always done it, this is kind of the way we need to continue to do it. It’s that’s not how businesses are going to grow in the future and survive. I can say that. I’ve been part of a lot of council meetings with like, Walmart, Disney. See where a lot of the major companies that are in Florida right now, Siemens, they’re all looking into every part of their business, just how can we automate even the simple tasks because once our employees start thinking about automation, they’re going to give us suggestions on what would help them the most. That opens up the capacity for us to put them to do more analytical tasks. Therefore, we don’t have to hire as much anymore, just because we have harder and more work. So the rule of thumb is, if you’re, if you’re if you, if you can’t keep up with the payroll because of the work you got, you definitely want to look at automation and give us.

Will Bachman 25:10
Fantastic. And john, where can people find you online?

John Joseph 25:14
So our website is white swan automation.com. We have all of our little BIOS there. And all of our digital different experiences there. And we have contact emails, we have our cell phone numbers there. We love to meet with anyone online. And we can set up meetings with you look at any processes you want to take a look at. And we’ll just tell you up front, you know, what kind of technology would be best suited for you and what that would look like.

Will Bachman 25:44
Fantastic. Well, john, thanks so much for joining listeners. We’ll include that link in the show notes. And let’s go think about what we can automate. Thanks, john. No problem. Thank you.

Related Episodes


AI Project Case Study

Paul Gaspar


AI Project Case Study

Astrid Malval-Beharry


AI Project Case Study

Julie Noonan


AI Project Case Study

Markus Starke