PegaWorld iNspire 2024: Process Mining in Action: How Siemens Takes a Data-driven Approach to Process Optimization

Welcome to our Process Mining in Action session. I'm really happy we have all of you here. I know it's the last session of the day, the last session of the event probably. So I'm the one separating you from getting some drinks downstairs. So thanks for being here. Really appreciate it. Uh, before I start, I usually do this, so I would like to be to do a quick show of hands. So please raise your hand if you don't know much about Process Mining. You're here to learn.

So the first time you talk about it. Cool. Uh, please raise your hand if you have been exposed to Process Mining but still haven't applied it to a project or an action. Big hands. I promise I won't make you pump. Okay, Good. Which ones have used Process Mining? Uh, you know, in your companies. In your enterprises to do something.

Please raise your hand. Okay. And some people don't want to raise hands. That's fine. That's okay. So today, uh, you'll have me and Andreas. Andreas here, sitting here will be. Join me in a second. So we would like to welcome you to the session.

This is what we plan to do. Basically, uh, for the ones who are new to Process Mining, I'll spend ten minutes giving you a crash course in Process Mining. Then I'll invite Andreas upstage, and he will share some of his experience at Siemens using Pega. Process Mining. Uh, and the the most important piece is the last one. We'll open it for Q&A. So ask us anything. You know, we want to talk about the good, the bad, the ugly. It doesn't matter.

So it's about sharing experiences and sharing knowledge. Okay. And by the way, I hate speaking, doing monologues. So if you have questions, raise your hand. We've got mics. You can ask the questions. Let's make it super informal. You know, there's no point in being formal today. Sounds good.

Are we good? Awesome. So why are we here? So what is Process Mining? Why should we care? So I got this. This from the last report from Gartner about Process Mining basically by 2026, which means two years from now, only 25% a quarter of all enterprises will be using Process Mining to pave the way to autonomous business operations one quarter. So it's upon us. It's not in the future anymore.

It's something companies are doing right now. So what is Process Mining? Process Mining is a discipline. It sits in between data mining and process management. And just like data mining, it uses data to explain things to help you make decisions and to help you predict the future. But it's different from data mining in the sense that it uses data about your processes. It feeds itself from event logs, from the execution of processes running in different systems, including Pega Platform, but not only right. So it's about using data about the execution of your processes. So you can do make better decisions so you can optimize your process.

You can automate it in a much better way. So far so good. So if I had to explain what process mining is, I would use this slide. Because if I would ask you what your process looks like in your organization, in your department, in your group, you would probably explain it like this beautiful paved way in the park. And and that's across the board. That's all over the world. We as human beings have a very idealized vision of what the process looked like. We we simplified things. Right.

So it's what we think is is is happening. But what Process Mining does, based on the data, it will show you the reality, how your process really works, not how you think it works. And that's a big difference. And that's great because if if there is a difference, it means that there are opportunities to automate. There are opportunities to make it more conformant, to make it more homogeneous, to get rid of friction, to reduce rework and to make it faster and more efficient. Right. That's where, uh, the, the the advantage lies. So how do we do it? Well, I told you before, we do it based on data.

What kind of data? Data coming from systems. Systems like Pega Platform, ERP, CRM, ITSM systems or even legacy systems. Even something you you built in-house, right? What do we have to have so that we can mine a process? We need an event log. So basically an event log looks something like this. So each line is one event. And as long as we have three fields case ID timestamp and activity we can start mining.

So case ID is the unique identifier of that process instance. It happens to be the same name that Pega calls it. So case ID timestamp is when something happened. An activity is what happened to that case ID at that time stamp. So far so good. Awesome. If you have that, you can mine a process. And what can you do once you mine it? Well first one most obvious one is discovery.

So event log comes in and then a model of that process comes out automatically. You don't need to draw it. You don't need to interview people. You do it based on the data. The second thing you can do, you can check the conformance of the process. What is that you feed the tool with what should be your your process model what you should be doing. And the event log comes in as well. And you compare the two and say, you know what, I should be doing this, but I'm not. And this is where the difference lies.

So it helps you keep your processes, uh, executing the way they should be. And third, and probably the most important one, it helps you enhance enhances the Process Mining word we use. You can you can say optimize, automate basically make it better, improve it over time again based on the data. So if I had this is what democracy looks like right. So event log came in and then we can see the process as if it were a map like Google Maps, right? We're used to it. So basically you can see the activities, you can see information about frequency, how often they happen, how often they go left or right, uh, how much time it takes to get from here to there and so on. Or you can have more formal methods like BPMN over there that is learned from from the data that was ingested, or it can replay it as if it were a movie. You can basically gain more knowledge about how your operations are running.

Think about it. Until yesterday, maybe you were flying without instruments. Now you have instruments and you know what's going on. That's really important. Then you can do conformance checking. And as you can see there, you can see some lines with green green boxes and red boxes. What they're indicating is oh you see the three first steps happened as expected. However the green ones are. We were missing them.

They did not happen. And by the way, the red one should never have happened, although it did. So it basically tells you how you are, how divergent you are from your process model. And last but not least, you do enhancement. And when we do enhancement, what are we interested in? We're interested in slow transitions. We're interested in bottlenecks. We're interested in reworks. So in all forms of inefficiencies that we can deal with and we connect that to data in your cases, which means we can say, hey, we got a bottleneck here.

And by the way, it's associated with this service in this region or with these resources or whatever kind of dimension that brings colors to your analysis. So far, so good. Yes. Question. Would you mind using that mic, please. Hello! I love the clear explanation so far. My name is Nicole Booking.com, responsible for process excellence in customer service. What I'm very curious about if you use Pega and you then do data mining and you want to assess your happy path and the and the compliance to the or adherence to existing processes.

I'm wondering if you find anything if you have actually, uh, kind of curated all your, your, your path. So there's no way to deviate other than just following the workflow in Pega. So I was just curious to find out how that then, how it show up and what kind of learnings you will still get from a adherence perspective. Awesome. Thank you. So so what happens is even when it's running in Pega Platform, you might have given the application some liberty to go to one step, skip another or something like that, or even go back. So if you know where you're deviating. Maybe you can go back to your rules and reinforce that from here, you can only go there, and then you avoid that kind of detour you're taking. So you can definitely fit this back.

And one way of doing it is capturing an insight about the execution. And then you promote that, push it to a jail studio saying, hey, here's a change I think we need to make, and I think we need to make because this is the impact we're getting. So one of the things about Process Mining is that we're telling stories about our processes. It's really about storytelling, but based on data. So now if I ask you to do something, I'm giving you the reasons why that is important. This is how many clients are affected, how many cases are affected, how much time we're spending on that. Right. So so it's about that. But definitely you want to close the cycle of continuous improvement.

You will find an insight. You go back, do changes continuously monitoring, monitor it continuously, do more changes. And over time you become better and better. Fine. Awesome. So let me show you a quick video. Just just so you can see it in action. Uh, so in this video, what you will see is a claims process in action. So let's go go, go.

There you go. So basically you see the map with all the steps. Now we're changing the metrics so we can see time how much time is spent between them. You can replay the process which is like Netflix on your process execution. You can look for slow transition. So reviewing the claim is taking a lot of time in this process. For example, you can see the reworks and then you understand there is a rework about asking more information affecting 17% of my cases. When I find I look for a root cause, I see that the location of the incident was not provided. And then of course I need to go back, contact the client to fill in the information.

That takes additional time. It's friction in your process so you can avoid that by again changing your application and making your location mandatory so that you know, once you are speaking to the client, you make sure you capture the information right. And then you can always use dashboards so that you can put the execution process in the light of the attributes of the aspects of the business. Because again, it's, it's it's fine that you find a bottleneck, but it's much more interesting when you understand what's triggering the bottleneck, because that makes it easier for you to fix. Sounds good. Awesome. So the big question what's in it for me? Why do I do that? What are the use cases for Process Mining.

So there's more than one. So some of the typical ones are these. The first one is process optimization and automation. Basically saying I understand how I am executing this and then I am very precise about the way I fix it. You know, I will do this change. I will use a Process AI in here. I will put a robot there. You're being, you know, surgical about how you make your intervention in the process. Second is operations management, which means you're continuously monitoring the execution of the process over time so that you can be always on top of how you're performing, what you need to do to become better and better and so on.

And to intervene when it's still timely to do so. Third is customer journey and experience, which means if you have a customer facing process and you know you can use process mining to put yourselves in the shoes of your clients. And this is powerful because you can begin to feel the pain they sometimes go through because of the friction that is in the process that was invisible to you before, right? Oh, they they leave and abandon my the, you know, the sales cycle here so you can work on that to convert more or they get confused in this step. Maybe my user experience is not good enough and we should improve it. It gives you a brand new perspective into what's going on. Fourth is process compliance. So it goes hand in hand with auditing. With risk management.

It's this ability to audit your execution not as not by sampling some cases like you would traditionally do, but by actually checking all the executions against your, your your your reference model. Okay. So far so good. Questions? Okay. So crash course now your experts black belts in process mining. Right. You're all there. So I would like to start the panel.

So we we want to hear, you know, the experience of somebody who's doing Process Mining for real. And I would like to introduce Andreas and I will get my Process Mining head over here. So Andreas, please join me on stage. Andreas works for Siemens. He has a background in computer science. He's been an entrepreneur. He has worked in the fashion industry. Uh, you know, a Renaissance man, if you will. And I'm happy to welcome on stage.

Please give it up for thanks for having me. Thank you. Are so. Awesome. But. So, Andreas. So we can start. Would you mind? Oh, I'll go back to my Clark Kent, uh, persona now.

So, uh, could you please briefly introduce Siemens? I know, I know, this would take two hours, so give the short version please. And what your team does at Siemens. Gotcha. So short version. What we are doing at Siemens. So we are roughly 300,000 employees working as a technology industrial company. We are divided into four different divisions Di digital industries focusing on automation of factories. Then we have smart infrastructure focusing on building smart homes but also building cooling for data centers.

Okay, then we have mobility focusing on trains and healthcare sector. We have Healthineers and that's basically our four divisions. Um. And your team. My team. Yeah. So we are GBS, we are a service provider and we are trying to support the business lines in providing value, um, trying to build digital solutions for them. So this is the future we want to move to. Um, that's why Process Mining is on our agenda.

G stands for, uh. General Business Services. Nice. Thank you, thank you. So there was life at Siemens before Process Mining, right? Long, long before that. And then there was process improvement initiatives back then. How did you do it without Process Mining? And what were the pains?

Yeah. So this is a really cumbersome process. Still is. Um, we heard today of a project where we gathered the cash collection part completely from Siemens into GBS. These kind of projects can take 3 to 6 months, even longer, depending on how big the process is to get it fully into into GBS. What takes 3 to 6 months? The analysis of the process itself until to understand them, map them, build this on charts and have a process description. and then, okay, we don't have them even on boarded yet. Okay.

And what in that journey. So you mentioned that it takes long etc.. So when you were looking for to use Process Mining, what were your expectations? Why did you want to do it? So basically like you described you're looking really for bottlenecks. So where the process takes longer, why it takes longer. Um, we want to reduce the clarifications. So where we have reworks in the process itself. Can you explain what a clarification I know what it is but it's rework.

Yeah. So whenever we get a request as a transaction, we don't have any information available to work on the request. So we need to get back to the requester to gather the last information to fully execute the process. And this is creating a lot of noise in the process. What we want to get rid of. Gotcha. So it's friction at the end of the day. Yeah. Friction.

Friction sounds good. And, uh, so when when when you thought about using Process Mining, you had some use cases in mind, right? So what were the use cases that you were targeting in the beginning of the journey? Um, you mean use cases in where we want to use it? Yeah, absolutely. So we basically have four work streams where we start OK these four work streams we take as a pilot. There was one team smart Infrastructure in Germany where we do service from Czech Republic. This is um, where we basically do service management for fire alarms. Police alarms.

It's contract management. This was the first target that we have 300,000 contracts per year where we need to basically execute and manage. Um, they are the biggest pain points from our perspective was always like, we get a lot of clarifications. Um, also the regions in Germany are working differently, so we wanted to compare them with Process Mining itself to see the differences, the variances, the similarities. Um, yeah. Then the other use case was for digital industry. There. We built our application, um, for the regions in digital industries where they basically manage the orders. So we rolled it out basically globally.

And what we want to do then is compare each region. So how's regional order management done in Italy in comparison to us? Also see what's the difference there. Can they learn from each other? Maybe we bring them together. And to do that you had to put a team together to do that, right? So what kind of people were involved in this initiative? Yes. So basically based in the first place.

So we start basically with the base out of the business itself that currently developing together with the developers, the applications. Um, so this is really they need to have business knowledge at the end. Um, curiosity. Curiosity they have. Because at the end they were, you know, um, people out of the business. We built them up to be a so they are curious. Otherwise they wouldn't need learn Pega itself. Now it's basically for them the next step to learn mining with this tool. Um, and the other big thing, like you said in the beginning, um, what we need to learn is storytelling.

And I would say it's, it's rational storytelling in a way that you say you tell a story with data itself, that we really use it at the end, a tool to, to drive change management internally. That's that's the big target. Yeah. Storytelling. Yeah. That's always a topic. And in this journey where you were exploring some of these use cases, can you tell a little bit, you know, some of the findings. So what could you find. What did Process Mining.

Yeah. Yeah. Tell you. So. I mean, we are mine since two and a half months basically. So keep that in mind. Um, what we have, for example, is in one of the applications, we found out, like you said, it's always expectations reality. And they are surprised positive or negative. There we had our surprise where we always thought, okay, let's go with portals as a front end facing to the customer to get the request itself.

What do you mean, a portal? You mean a portal to the case management system. To do the external customer. So yes. Um, and what we realized then, because we could compare now the input channels, emails versus the portal itself. And then we realized that the process is coming via the portal took longer than the cases. By coming via the email. Say that again, please. Again.

So the cases that came via the portal took longer than the cases that came via via email. That was a. Did you expect that? No, no, that's what I said. I mean, this was a big surprise also. Yeah. I remember the day we found that out. I almost fell off my chair. Yeah.

And I mean, we don't have an answer yet. I mean, this is the idea behind Process Mining. I mean, we figured it out. I shared with you, but I can't give you today an answer why this is. So. We're still in that journey to figure that out. Why? That is the reason it might be. I mean, we have some ideas.

It might be that we have more automation currently in place via email. Um, and. You need to move them to the portal. Yes. When, when it starts with that channel. Awesome. Yeah. Very good. So this was something that was unexpected.

Yeah. What about the things that were expected? So what does Process Mining help you when you already expected something? So like I said, I mean, the one team we are doing, the contract management itself there we had the expectation that we we know clarification are our pain point. Yeah. Um, what we realized. Yes, it is a big pain point. Um, so in numbers, it was roughly that just 7% or 8% of our requests were with clarification. But the not not the manual effort, but the cycle time itself was like 20%.

So the impact was was really huge just because of the clarification. We had a lot of noise in the system because of that. Um, yeah. But and how do you how do you solve that? So you identify that, how can you solve such a problem? So what we figured out for now is, I mean, we have multiple ideas at the current stage. One idea is we dive a bit deeper into the data and realized again that we could because the clarification happened with a big spike in a certain point in time. And we thought, okay, let's, let's maybe flat, flat that thing and say, let's have the clarification over a period of a year so that if we have the spike, we have a huge backlog. They are overwhelmed.

They can't close the cases. So to flatten the clarification at least, um, what we want to do, because this is a seasonal business. So you need to imagine we send out the yearly invoices. Then they realize something with the contract itself is not matching. Then they come back to us and then we have these clarifications. And so what we want to do is inform them much, much earlier that we will invoice them that they need to think about. Okay. Um, keep in mind check your master data, update your master data maybe a bit earlier. Send us these requests.

Can you give an example of a master data? What does. Um, address. Yeah. So the address on the invoice is wrong. Just an example. And then they say, hey, please change the address. But then we have that huge spike where we need to, I don't know, 10,000 address changes within two days or so. It's too big for a team of 40 people.

Sounds good. So we want to flatten that. That's one idea. The other idea is now what we are trying to do is reduce the clarifications by put an external portal in front of the customer. Um, and here we build now slowly the story up, try to do the change management together with the customer here. So as I, as a customer from our perspective, um, to really give the customer the possibility to enter the address changes in the portal itself immediately. I think fuel numbers was something else that that was often, uh, one of the reasons why the invoices were, were, were challenged. Right. And awesome.

And, uh, what else. What else could I tell? What else findings? Um, I mean, another interesting finding is that we I mean, I said we have multiple regions in Germany, different kind of business in these kind of regions. And we realized, for example, in one region, um, that we have a lot of clarifications to a specific government department. Um, so the government, of course. Yes. And in other regions we don't have that. So the idea would be now this would go into a digital solution, might be it's just a clarification needs to be performed with that, uh, government part to you know what, why do we get these requests always from them or not from the others to, to figure that out.

So it also helps you to to change the process, maybe in the direction of outside world, not only internal processes. And what role did Process Mining play in identifying those findings or in quantifying those findings? How did that impact the way your team worked upon, uh, the data and upon the fines? Let me put it another way. Could I just have a report? Run a report and find the same things? Yeah. So what I think we realized here. Really?

It changed the way you think. In a way. So really, how you see the world. Because suddenly you have a different picture in your head. How the world looks like. If you see the process in front of you and you suddenly have completely different discussion, what you maybe wouldn't have if you just have a Excel sheet in front of you. Um, and the interaction to interact with the tool, it's like like it's storytelling, but it's imagined, like real mining. I mean, you really mine go deeper and deeper and deeper. And this is how you saw in the tool, right?

That you click on certain activities, then the activities is opening. Then you dive deeper, um, look into details. Then it's like, ah, maybe not that the right way. You dig back, maybe in another direction. Um, so it was really, really helpful for that for that session to, to find things. Awesome. And, uh, so you've been through you've been down this road. What advice would you give to somebody who's starting that journey? What advice would you give?

Um, first of all, I would say get your stakeholders on board. Um, because, I mean, you want to use the tool as process change for sure. So to change your process, get the stakeholders on board that are involved in the process. That's the one thing for sure. Um, define the problems. Define the problems for sure. Clear. You want to mine because otherwise you you are lost in a way if you don't define them well enough, if. You don't know where to go, any, any way will take, you know, any road take you there.

Right. So. And Foster creativity. I mean, the guys needs to be really open minded. Um, because, I mean, the idea is the picture showed it really well. You had in the beginning. This is the way we thought the process is, but it might be not that way in the reality. And then you need to be open minded, creative to accept that reality in a way. Otherwise you.

Yeah, it will not make sense. You close your third eye in a way. Sounds good. Well well, well. So, uh, I would like to open for questions, so let's make this more interactive. Yes. Whoever wants to go if you can approach the mic. There are two mics, one on the right, one on the left. Here.

Yeah. I have a question for Zlatan. So, um. Yeah. So, uh, my question is, um, how did you obtain, uh, buy in from your executives on adoption of the Process Mining tool. Uh, how did I do it? Yeah, I mean. Let's say, I mean, I convinced him with the story of we need to increase our data quality. So we have that vision really to go in this autonomous enterprise direction.

Um, and he knows that our processes have a lot of noise in it, and noise is not a good thing if you want to develop, uh, AI in the future. So that was one door opener, for sure. Um, and the other door opener was a high level business case. So also that helped here that I know the costs, what it will cost me one year. I know the processes already. Um, what effort lies behind the process? And Then if I say okay, I find 7% productivity gain per year. If I use Process Mining, then it was for us a door opener in a way to say, okay, let's use it. That that kind of was the price.

So the 7% was the price. We got to find 7% cost reduction opportunities using Process Mining a year. Did you meet it after? No. I mean, it's just two and a half months for now, but we will mine. We will not give up, but I'm pretty confident that we will find more. Yes, yes. Hi, my name is Subhajit. I work for Accenture in the Netherlands.

In the recent past we did a proof of concept for one of our clients there. Mhm. Um, so my I have two questions. Anyone if you can answer it like that. So one is uh, we found it very hard at the beginning also to define what we want to find out from Process Mining because it's a concept. You you understand that you need to gather the noises. You. The terminology used is really nice. So you want to understand that.

But it's not always easy from a business perspective to get there. Like one of the business cases was, I want to know how many cases are going straight through. There could be a really simple way of finding out by using a report, not using Process Mining. So how did you go about defining those use cases, like what I want to understand from my Process Mining exercise. That's question number one. Another question number two is also connecting other systems which have other ticketing systems and stuff like that, which is not built in Pega. So processes not built in Pega and connecting them to the Process Mining via an API or an extract or those kind of stuff. This always is a question. I heard it around, uh, this PegaWorld multiple times.

How are we doing it? Is it really easy? How much effort do we need to put into that? So if you can enlighten us on that part as well. Sounds good. Thank you. I'll take that one. So, uh, I'll start backwards. I'll start with the last question.

Um, is it easy? So Siemens is an example. So the process is orchestrated in Pega Platform, but it's not completely executed there. Right? They have SAP systems. They have other systems that they go to in order to complete the execution of each case. Right. So what we did, we used Process Mining on the Pega Platform on the case of Pega Platform to give us a first look into how that's going on and break down the execution into pieces. So we knew where we were, you know, where where the slow transitions were taking place, where the rework was taking place from from a more generic perspective.

From there you can dive and say, oh, let me, you know, get some. This system is important. Let, let let let's bring it in. And, uh, and here's the thing. Process Mining can manage can can mine Pega processes and it can manage non Pega processes, mine non Pega processes. So that's possible. And for that we have different connectors connectors to you know files different file formats to databases to RESTful APIs so that you can put that data in and mine it as well. Uh so so I think in their case the first approach was using the connector because it was easier and data came in and then we could identify, uh, you know what? If we need to bring new systems in, uh, the first question was.

Identifying use cases that you want to define using. Oh, yeah, that's a good one. So. Well, we had a clear mandate. So five 7% cost reduction. And you know do that. So and and that was interesting. So we did a workshop session in London and, and uh, Christian, who is uh, Andrea's boss was there and he's the main sponsor and he's the one with the 7% number. And, uh, and we were kind of saying, hey, what about the clarifications?

And, you know, all all this notion that we need to reduce that friction. And he came from a different perspective. He said, yeah. What about the ones that don't have clarifications? Why are they not automated yet? Oh that's a good question. So that gave us another brand new road to to to to follow. So yeah we can automate these ones. We can reduce friction over here.

And when you break down into pieces and look for the root causes and the opportunities, that's how you build your, your, your case and, and try to go after your objectives. But start with an objective. I think that that's that's one of the, the takeaways. Like in that case it was clearly stated what they wanted. But it's not the only thing you can do with Process Mining. It's not just about cost reduction. We have clients who want to increase the conversion of sales opportunities, for example. It's another use. Or they want to optimize working capital.

So shorten financial cycles. So find opportunities for automation. So it depends on what what your business is asking of you. Yeah. So the proof of concept that we did also was a successful one. And I think the example actually is spot on because we did a use case of we have two contact centers which operate in two different models. So there is no way for them to normally compare to contact center. How are they performing. They don't know that.

So we did that side by side comparison and it was a success. A lot of different families and it was a big. Awesome thank you. Yes. Hello again. Um, so Process Mining from Pega is relatively new. So how long have you been? Have you been using the application? So we started last year in August.

Okay. Roughly, I mean, just for eight weeks, let's say eight weeks. Quickly dive in, get a feeling. Pega gave us access. That was also a good door opener because there we already got a good feeling to say, okay, looks good, let's do it for another year. Then we paused the project until roughly December. December. We then said, okay, let's restart the project. Um, build now the connectors.

Connectors, what he described. Because what we did last year was manually uploading it. What we can do now is basically press a button and have always refreshed data. Then we had some training, took another three months and then we started mining for now, three months. Basically that's the time we are mining. Um, and he also set the expectations right. We thought really that we could mine faster. Um, but it took us longer. Okay.

So then one follow up question. This was very insightful. So for how long have you been on the Pega Platform and how did you before check if the workflows were running as expected, not having process mining. So we are now for five years roughly. Um, and if they're running like expected, um. Users complaining one thing um, or we check internally with KPIs reporting. So we had reporting. Yeah. Reporting.

Um, we used and then we checked it on a monthly basis, weekly basis dependent on what we want to check how many cases we get by week? Do we get no suddenly more or less cases as an indication that, you know the tool is not used well. Um, this is how we did it before. Okay. Thank you. Very clear. Let me make a comment. Because this I remember a conversation I had with Christian when we were there, and he said something that was very interesting. He said, the fact that I have Process Mining now and that I can quickly monitor how I'm executing it, can help me speed up the development process, because if if the portal is not behaving as it should, I can kind of capture it very early on and course correct.

So, so I can iterate faster knowing that there is a safety net that helps me course correct in case something does not go exactly up to plan. So that that was a kind of a another benefit that that he saw. What else? Yeah. Hi. I'm George from Singapore. Government. Okay. Uh, I want to have two questions.

One is about the conformance checking you mentioned about using event logs and models. I'm just wondering how you generate the models and fit the information to the tools to say that. What is your expected? Okay, so the other question that I have, uh, or you want to answer. Okay. Go ahead. Sorry. So the other question would be, uh, in practicality, right, to do the process mining how often that you should do this kind of checks. Uh, on the process mining.

Is it, like three months? Six months? What is your experience to use the the process? Both good questions. So I'll start with the first. That is easier. So uh, for us to do conformance checking you provide the model in BPMN notation. So if you have your model in BPM, you you you you upload it to the system and then it will compare that model into the execution that you're having about the second one. How often?

It depends. Like all things in life, it depends on what you, you, you you, you want to to do out of it. Uh, there there are. If you think about use cases I showed earlier, if you're about operations management, you might update data every hour because you want to know what cases you might make a change. Like it's really about the operations. If it's about auditing, maybe you do it every year because that's the auditing cycle. So it really depends on the use case you have. And sometimes it also depends on, on uh, the average life cycle of your case. So how long it takes.

So how often do you want to assess that and that that sort of thing. Right. You will often. But so I would say think about the cycle of of process improvement. So you found some insights like we did there. So we have a list of things that we can do. So you execute them how you know, how soon do you expect to see benefits. Because what you want to do after you do some automation improvement etc. is to put side by side what happened before the automation happened afterwards so that you can follow through the improvements you are you are executing right?

Yes. I'm Shubha from MUFG Bank, Ltd.. I had a couple of questions. One was we have implemented workforce intelligence and I want to know if we can layer Process Mining on top of that or is there, uh, you know, I also heard about task mining. So which do you think is the better alternative? And if we can just take the data from workforce Intelligence and push you to Process Mining to get, uh, you know, uh, the identify the opportunities if that would work. And the other question was, with respect to the change management, how did you actually implement, uh, you know, make your, uh, the workforce adopt whatever you're doing. What was the strategy you employed for that? I'll take the first one.

You take the second. Yeah, but we had a third one. There was a. Third one third. Three one, two, two. So. So about task mining Wi-Fi. Yes they are. They go hand in hand.

We usually say that process mining gives you this ability to see things from beginning to end. So you go very wide. But sometimes you want to understand what people are doing at some steps and even outside Pega Platform or in other systems. And then you can use task mining to really see what's going on. And you go deep into what's happening. Right. So, uh, it's so close and so important that we are currently bringing the two products together so that you will be able to do process mining and task mining in the same platform, the same tool. And if you want to see that in action, there's a task mining booth in the innovation hub where you can see some of that. Um, yeah.

So the idea here is think about at the end you tell a story and then you break down the story, maybe into epics, and then you break down the epic into user stories, and then you have a good feeling, you know, how much effort it is, how you want to implement it. When do you want to implement that? Um, this is. So there is no rule of thumb, you know, that you can do it within two weeks. One week. It really depends. How huge is your story? Um, the one story where we probably need to change something with the government will take maybe years. The other thing where we want to implement now, uh, automatically email reply, something that will be just done within a Sprint two weeks.

Yeah. So break down the user story into the story you are telling into an epic, into a user story. Then you have a better feeling about the effort, and then how much change management needs to be performed for each of these stories you build up to. That's how we do it now. Guys, I think we've reached the top of the hour actually one hour late. So I would like to first of all, thank you very much for being here. Please visit the booths in the Innovation Hub. So there is one for Process Mining, one for task mining, one for process intelligence. You will see more of what we had here in action there.

Uh, ask the questions. Uh, Andreas and I can stay a little bit longer and take some questions personally and, uh, enjoy the rest of the show. Thank you, thank you.