#12 The Process Mining Overview Episode

When you get automation right, you can dramatically improve productivity, performance, and your customer experience. But what is the point of automating processes if you aren't sure where your processes were going wrong in the first place, or whether you are even putting the most suitable processes forward for automation? In today's episode of the Process and Automation Podcast, Arno and I will cover how process mining unlocks your automation potential.

Great point, great start as usual, uh, Sasha. It is all about, you know, what is the point of of all of these processes we are automating if we don't, if we're not really sure where things are broken in the first place, or even whether or not we are putting the most suitable processes forward for automation. So I think the key driver here is really to identify those processes within your business that really suits that kind of target need for automation.

Yeah, sounds really good. What do you think, what else should our listeners care about, or why should they care about process mining?

Well, process mining is of course a new term. Well, it's an old term, but it it's becoming more mainstream now. And I think process mining is really just an automated way to process discovery. And process mining is important because it can point you in the direction in your business where automation is most needed. So, and we're obviously still at the tail end of the pandemic. And in response to this pandemic, we are having, A lot of businesses around the world are digitally transforming at warp speed. We hear about these sort of 3 to 5-year digital transformation programs that's now been sort of squeezed into a sort of 12-month period. And also McKinsey Co., they recently did a study and they found that most executives out there have amplified and increased funding for new technologies. And, you know, companies have accelerated their digitalization plans, you know, by years, you know, as a direct result of the pandemic. And I think this is where process discovery is, you know, using process mining is certainly a catalyst.

Yeah, absolutely. So, so, you know, this, this potentially It represents a positive or definitely it's a positive and welcoming change in our world of automation, in our world of work, I guess. Yeah, and companies must really take care not to get swept up in this digital transformation wave and people get too excited, I guess, as well. And yeah, and they should really avoid rushing into automation just for the sake of it because it's now like a really trendy and sexy topic. I think typically this is CIOs, when you start a new role, where you would concentrate on, isn't it, on these new topics. Yeah, so I guess, yeah, you still need a clear plan as always when you kick off a project and what is sort of return of investment. I think all that stuff, yeah, it's not a silver bullet and you still need to plan everything as normal really.

Yes, and I think many organizations out there they don't really have much clarity on their as-is processes, the processes that's currently within their business. And they're not really aware of how these processes perform within their business. And without this knowledge, it can be really, really quite difficult to identify, for instance, where processes are working and where some are not working, for example, which problems require urgent attention, and which type of technologies are best suited to solve processing issues. And I guess if the listeners out there, if you're executive and you in that same boat with all of these where, which, and why questions around process discovery, I think process mining is the tool for you to look at. And there's really very, very good products on the market like Process Gold, Signavio, Celonis. We also use one called QPR. So I think, Sascha, let's jump into the topic of process mining tools next.

Yeah, absolutely. Yeah, so before the process mining software really emerged in the market, documenting and mapping processes was a big task and very often a costly undertaking. So that Really, yeah, I think lots, lots of people sort of put this a bit to the side and sometimes dived straight in into the, the actual doing of process optimization and doing manual process to digital process, but maybe not even knowing if that was actually the right one. So yeah, so this, this manual documenting and mapping of processes, yeah, involved gathering and collating data manually. Yeah, so you had to sit down next to people, speak to internal resources, and, and yeah, and sometimes bring in outsourced people and to really go through, through what, what go through in the what is going on in the processes and really sit there with a stopwatch and doing all these time motion studies, isn't that? And, and yeah, that was very time-consuming, very costly, and very often that even when this all put together, the big documentation, lots of documentation, sometimes no one was really looking at carefully and very often out of date after it was produced. So, so now, now we have process mining software that can solve these challenges.

Yeah, and I think this, this process mining technology, it really allows you to easily capture, you know, information on your processes. Directly from your enterprise systems you work with every day. So you could, for instance, point it at your ERP system and actually discover processes that's inefficient, processes that takes a long, long time to execute, processes that's potentially broken, processes that has got inherent bottlenecks in them, processes that causes negative effects on your business. Without process mining, if you were to go on the traditional approach of process discovery, it would be very hard to have a shot in the dark, so to speak, and hit those pain points. So using this data mining capabilities, the software really inspects your systems, your line of business systems, your ERP systems, your event logs, and then it documents and it automatically documents, should I say, your as-is processes. And much of this procedure is automatic. So it means you don't have to set aside expensive resources from your workforce workforce, you know, to actually do this labor-intensive work. So it's a very convenient way to look at all of your detailed data that your processes produce, and you can obviously, you know, map out those processes. So you could then, you know, use this visual web-based process mining or the underlying mapping tools to create your process diagrams that that's really factual and very objective, and then people really understand and can work with because it will align with the data. It will be familiar to them. They would be able to look at this and say, this is the process we follow. And the beauty of the process mining is the fact that it will also apply some metrics and performance metrics across your processes. It will allow you to look at the best of its class. So you could look at the processes that execute the fastest, also the worst of of its class, you know, which processes are the worst performing ones, and also give you drill-down capability. So if something, let's say there's a process, let's say accounts reconciliation, that is fast in your German office but slow in your Spanish office, then you could investigate that. Maybe that's just because, you know, Germans are more efficient and Spanish take more siestas. Or it could be the other way around, right? So there's certain things that happen in your Spanish office that the Germans lag behind. So this is the thing that with conventional discovery process, discovery is near impossible, right?

Absolutely. And you can also see plenty of compliance breaches, not necessarily how long certain steps take. So you could even then find out in real time that, that so many processes, when we look at accounts payable, for example, that the invoice might be submitted after a PO, before a PO has been raised. So that immediately makes it clear that something is going wrong here. And then because the invoice comes before the PO has been raised, the process is sort of already inefficient and lots of people, yeah, they want answers. And if you spot those things, Yeah, you can take immediate action. And I think without that kind of insights automatically, those kind of instances are probably not really detected. And processes are lived like, yeah, completely, yeah, live differently than—

Yeah, I think the point is, you know, it gives you that clarity, right? So you've got a clear view on your process performance and it's all factual-based, it's based on your data. You can see all the steps and the systems that's involved, the people involved for each of these processes. And like you say, you know, you can easily spot anomalies and costly anomalies like invoices going out before purchase orders arriving, right? So that will cost somebody time to actually fix that and, and write back to the supplier and say you need a purchase order. Now that might happen just once and you could forgive that sort of process is broken, but if you can imagine having thousands of invoices without purchase order. That takes a great deal of very expensive resource in an accounts department to sort that out, whereas process mining can actually highlight these issues and not just highlight it, but you could also look at simulations, how we can fix this. If we take certain steps, how can we improve the process, right? So I think The key thing here really is, and I think you can elaborate slightly on this, but it is kind of making decisions based on data and data with inside your systems.

Yeah, absolutely. Yeah. So I think this is, this is the key, key factor to get as much data as possible. I think this is potentially one of the key things you have to look at that you get a really good end-to-end process view of that data to really then get a meaningful picture. Sure, it makes sense to just look into the ERP which is great, but there's so much stuff which happens in an end-to-end process, which is far— there's so much more stuff happening. So maybe you have an SAP system, but the end-to-end process includes maybe this invoice portal and this supplier system and another system. So if you don't capture that data in all these other systems, you get only half of that truth. So there's a big, big effort to be not effort, but you should, should concentrate a lot on, on these other activities as well. So this is why lots of effort has been put into building these extra solutions where you can capture pretty much everything what happens in an end-to-end process. So like little applications would sit on, on a desktop which is recording pretty much everything what an employee is doing. Within, within that process end to end. So that really helps you to, to get the real picture. And yeah, yeah, once, once you have that real data, you can find out the real bottlenecks and inconsistencies. And yeah, it's, it's super important to, to get the whole picture in.

Yeah, and like you say, you know, when you have all of these intelligent tools, you can really go to work and analyze that data, those clicks, those tasks, and you can even sort of identify trends in the data patterns, and like you mentioned, bottlenecks and these sort of things. Now, again, if you don't deploy something as process mining, process mining software, it is very, very hard to get to that level of granularity in a timely fashion. And also, you know, the process automate, you know, like we say in all of this, with process automation tools, you could make transformation more collaborative, right? So you could collect feedback from all of the key stakeholders and the process pain points, ideas for improvement on these processes based on factual information. Because I can guarantee if you speak to somebody in accounts and you tell them that we think there's a problem with with invoices going out before purchase order. I'm pretty sure somebody will say, yeah, we have actually seen this, but we didn't thought about mentioning that in our traditional process discovery workshops because we just dealt with the problem, you know. And for this, it's— we bring that to the forefront and say, let's bring that out and put the evidence on the table and be a bit more collaborative and proactive and present to people the broken processes. And, you know, I bet you top money that a lot of people will very quickly prioritize those as the ones that's super broken and the ones you should focus your attention on, right?

Absolutely. And I think the beauty of having all that data available as well, so you could collaborate with all stakeholders. What you can then do is with this data, you can set then automations in to sort of perform actions based on that data, maybe in some systems to sort of disallow certain activities based on that specific behavior that process mining has just uncovered. So it is really, really good to use all that insight and work on the priorities. And there's definitely lots of capacity, business capacity to gained back really. So it's all about business capacities, and that's if you're wasting time here and there, and just to bring them back onto the table to get stuff done and maybe to release funds as well to get more innovation projects going. So it's a very important topic.

Yeah, I would agree on that. And I think that the end goal of investing in something like process mining is to really agree or understand how these broken processes should be fixed and how they should run, the kind of end state. So the as-is state is the broken ones, the to-be ones is how the process should really run. And that's a great way to agree a way forward to fix it because we can lay it out as it is, we can understand what's broken, And now it's clearly visible for everybody in the team, you know, where we need to focus our attention. It is really then easy for the organization to agree on the processes and the process components that needs to be improved. So, you know, and now we've got a way forward. We've discovered the broken ones. We've understand the to-be state. We've got a nice priority list of things that we need to fix. And now we can start to automate. And what a great way to start automation with having the right things in place to automate. And that's what this software brings you. It tells you the areas that you should focus your attention on. And now we bring the tech into the picture, the automation tech, you know, the robots, the workflow and case management, the low-code applications, the AI, all the things that we feel we could not just automate this process, but really transform it, right? Because we know the ones that we should be, that we should target. And this is why we say process mining supports automation decision-making, right? Because that's the key of it. It tells us which ones we should target.

Yeah, otherwise you just concentrating too much on just workflow, workflow. Although, yeah, the key problem is somewhere else. It might be a good idea still, but it's not necessarily the best one on and the top problem to solve for the business.

Yes. Well, I also think it's, you know, if you look at different problems with inside a business, you can build a very small robotic automation that from a return of investment probably saves a lot more time and resources than, for instance, automating a wrong end-to-end workflow. And I think this is the key. And I think vice versa as well. You could choose the wrong process to automate using RPA when you miss an opportunity where a workflow process could have solved another problem that avoid these invoices coming in before purchase orders going out, which cost a lot of money to remedy. Absolutely. So I don't think it's necessary that the tech, although, you know, things like AI, RPA, all of these progressive things are really powerful. But you could spend powerful technology solving the wrong problem, whereas, you know, just sort of to the left, there's a problem there that's really expensive that could be solved, you know, using a bit of workflow that's a lot more cost effective, you know, saves a lot more time and, you know, produce those types of results that unlock people's time so they can actually spend more time looking at other broken processes and prioritizing those things, right?

Absolutely. So shall we, shall we drill in, do, do go, go deeper into the different technologies, or was this a good, good example?

Yeah, let's look at it.

Okay, so, so one of the first automations we can obviously look at, and I touched on this already, is robotic process automation, RPA. So we already had a couple of episodes in our podcast talking about RPA. So if you're not too familiar with RPA already, so yeah, please have a listen to all these episodes as well. Yeah, but otherwise, yeah, RPA is really good suited for tasks and processes that are very repetitive and high volume and rules-based in your business. So like processing spreadsheets manually, invoices coming in manually and entering that information into the ERP system. So these are very, very simple and good examples where RPA might be very beneficial. So yeah, that means, in other words, the process that runs the same way each time at high volume with no need for exception handling. I think that's what RPA is sort of in a nutshell. And yeah, very, very powerful as Arno mentioned, and is now connecting legacy system without API connections. So that's also a very, very good use case, a very, very popular and attractive one.

Yeah, and it's generally, you know, for the types of automation where, you know, you, you don't really need a high degree of exception handling, right? Now, where you need sort of that, that sort of human touch still, this is where we sort of enter our workflow and case management technologies. And this is ideal for automations where we require some human input and exception handling. Handling or judgment or decision-making. Now, this sort of technology using workflow and case management, for example, it's very common to deploy that next to RPA as well. So you can almost see that the sort of the human workforce works with the digital workforce. The digital workforce might do some repetitive and mundane activities, hands it over to the human to make a judgment call and/or to sign off and then hand it back to the digital worker or RPA to actually execute the transaction. Now, this this sort of workflow and case management technology, and, you know, this, this, this can be sort of orchestrate or orchestrated using orchestration tools, workflow orchestration tools that you could, you know, govern the flow of work between different resources within your business or within different systems, for example. And like I mentioned, you know, this could include RPA robots as well. Task management, assign a robot a task. Yes, it will do it very fast, but at the same time, this workflow orchestration can also assign a task to a person, right, to, to do action and make a decision in order for the process to, to move on in a timely manner.

Yeah, so this is sort of really taking care of this end-to-end processing, isn't it, the orchestration?

Yes, that's right. So by means of an example, a case management, you might have a a mortgage application, for example. So there's going to be a lot of sort of workflow processes with inside that. And, you know, that's something that you can map out using a workflow tool. And, you know, once you're done with that, you might actually identify some of those activities in that workflow to be robotic. So you might then decide to plug in some robotic processing in there. You might not do it on version 1. You might just have an end-to-end traditional sort of workflow/case management implementation and then bring on robotics a bit later.

Okay, yeah. And then we have our low-code application development technology, which is sort of helping us to build our workflow case management and our RPA solutions lightning fast, isn't it? So that's, yeah, with low-code we can bring in this extra functionality super fast. Supplying, or not supplying, providing digital forms very quickly in the solutions. And yeah, build the whole solutions around low-code and RPA 20 times faster, for example, than with traditional coding methods. I think that's a very powerful method. To get your business applications live.

Yeah, and they always, they always say, you know, with low-code, this is a way to, to turn piles of paperwork, disconnected documents, spreadsheets, and confusing email communication chains into digital capabilities and digital forms. And, you know, and you want to do so very fast. Now remember, the context for this conversation is process mining, and the good thing here talking about low-code is the fact that we have got the output of our process mining now. So we are actually using, or using low-code workflow and case management and RPA to build out the right automations. So it further amplifies that benefits all of these progressive technologies provide, right? Because you've got the right hammer for the job, you know, and you are applying the sort of right solution to very strictly identified problems.

Yeah, absolutely. And the list goes on. We haven't mentioned artificial intelligence or other things in this mix, so you could throw chatbots in and all that stuff. And as Arno mentioned, yeah, we're picking the right tool is really important to get this automation underway. So yeah, the list goes on, and sometimes process mining uncovers all these inefficiencies which do not really need any technology to make the process better. So if it's just a procedural problem you uncover, you just tell someone now going forward, this doesn't happen anymore, that invoices or purchase orders are not sent. And then this problem is potentially minimized or completely eradicated straight away without any RPA, any low-code, and any workflow. So yeah, The key thing is really to understand, in order to optimize and streamline process from end to end, automate anything that can be automated. You may need to combine multiple types of automation. So there's not one specific thing, just combination of that will get you there.

Yeah, exactly. And I think, as you can understand from this podcast for our listeners out there, Once you understand, you know, the sort of the glitches in your processes, um, you know, in, in this way, then you can have informed decisions as to what technology, you know, you should apply to, to, to fix those processes. You know, if, if you scramble and just to go ahead and automate the processes without, you know, looking in depth at the processes first, you might actually choose the wrong technology as well for for the job. And this will be very disappointing, not just for yourself, but also for the morale of your transformation team. You know, nobody likes to fail at things. And I think, you know, in this typical instance, we are already succeeding because we're choosing the right problems to target. And, you know, and once you've implemented your automation technologies, you can continue to use process mining tools to monitor the process performance. You don't have to stop there. And, you know, you could use the data data that comes out of the automation to drive continual process improvement. So that's a great bonus in all of this, that you don't have to unplug all of these things because you're in your ideal state now. I think it's a kind of constant thing. And I think this really gives your business a very critical advantage, you know, at this time where business agility is all important, is really a key, you know, for success. In the current business climate we are conducting business in, right?

Absolutely. Yeah. Thank you, Arno. And yeah, as always, remember, we said it before already, technology alone is not necessarily the silver bullet. You need technology combined with digital automation and the industry expertise really to maximize return of investment. And yeah, this is why working with leading partners and vendors is really key. Key to your success as well. Yeah, process mining can be a big start of something really big in your business. And yeah, we here at the Automation Guys, we can't wait to help you on your way and to get you on your way. And yeah, we hope you found this overview of process mining useful and hope you— we hope you can take a few points away from this episode today. Please visit us on theautomationguides.net for resources. We have certainly planned a few more episodes, Arno, I think, on process mining over the next couple of months. So please watch this space for more on this topic. Thank you all for listening again and see you all very soon. Unfortunately, that's it again with This episode of the Process and Automation Podcast. If you like this episode, please give us a 5-star rating and don't forget to subscribe to this podcast so you don't miss any upcoming episode. We hope you will tune in next time, and until then, let's automate it.