LIVE from International Conference on Process Mining in Rome

Welcome back to the "Mining Your Business" podcast to show all about process-mining data science and advanced business analytics. And we are here live from ICPM, it's exciting times, and we will be getting into a lot of very interesting topics today. And Jakob, how are you doing today? I'm doing fantastic Patrick, thanks. Great. The episode will consist of us talking to a lot of people here live at ICPM, getting their thoughts on what is happening around process mining and all the exciting things happening in this space. Stay tuned. Hello again, this is Jakob and I am talking to you, well, I would say live, but it's actually not really live, from Rome, from the ICPM conference 2023. It's actually my and Patrick's as well, third conference, we started two years ago in Eindhoven, then last year it was a volcano in Italy, it was actually lovely in the Alps, it was beautiful weather all the time, so it was really nice. And this year we are in Rome, for me it's actually the first time in Rome and we did the yesterday as we arrived to a conference, we did go straight to see a Colosseum, and then it was just pouring, and it was really, really bad weather, but we still managed to see some bits and pieces off our Rome and we really had fun, Patrick, how are you doing? Doing very well, I mean other than the fact that we got absolutely drenched yesterday, we did manage to see a lot of beautiful sites and Rome is very, very beautiful and I'm happy to be here. Also, I should mention that normally we record these things online and this is one of those rare occasions, we actually get to see each other and converse face to face, so this is a rare gem here. It is very rare and to be honest, if the sound doesn't sound perfect it's probably us because we have covered the room where we are recording with tablecloths and it looks quite funny. However, let's get to episode which will be a little special as well for multiple reasons. I guess the main reason is that we will bring a lot of guests, some of which you already know, some of which you might hear for the first time today and we will be interviewing them on the, essentially what they have a lectures here during the ICPM conference on. So we bring the interesting guests from the industry day, which is the part of the event where the leading experts in the field but also academics are talking on different topics that we've covered quite a bit already now, we're showing previous episodes and we try to give them a little bit of time to really get you up to speed with the latest trends, latest news and developments in the field of process mining. Yeah, because speaking to a lot of people here, there are some cutting edge things that are happening, especially the field is very hot right now with new innovations and especially AI and all these topics and we can talk to the researchers actually moving that front closer and closer. So we're gonna be hopefully talking to a lot of interesting people and getting those topics to you. Alright, so let's start with the first one. Hi, this is Max from the front of a center of pros and intelligence. I'm one of the co-chairs of the industry track at this year's ICPM and I'm happy to answer your questions. Max, super happy to finally see you in person. We did record an episode about ICPM with you roughly two or three months ago. So my first question is, what has changed if anything has changed and how excited you are that it actually came into reality? Yeah, well, from a conceptual perspective, there were not too many changes, but we made everything become reality now. So we have so many excellent speakers here, both from the vendor side, from the research side, from the adopter side, there are so many people here interested in doing boundary spending between research and academia. And this was also most part of what changed since we talked last time. What is boundary spending? Boundary spending means that, for example, in our context, research talks to industry and industry talks to research in order to make sure to get most out of it. I have a question regarding the keynote because that's gonna be Marlon Dumas. What is your stance on this AI entering the world of process mining? Because I'm sure a lot of people are asking you also what it means. And I feel like we're gonna hear more and more about this topic in the future. And since we also will talk to Marlon in a little bit, I wonder what your opinion on that is. So first of all, I'm super happy that we got Marlon for the keynote because he's a boundary spinner between academia and industry. And I also think that he's a very well-known thought leader on process mining. And he's also one of the advocates of artificial intelligence in the role of artificial intelligence in process mining. So from my perspective, the key thing about artificial intelligence is that it helps us take over cognitive tasks of people, from recognizing, to deciding, to acting, to reasoning. So we call this cognitive function lens. And the cool thing is that now with artificial intelligence moving on to generative AI, there will be many more activities that we know from business process management and process mining that can be augmented, that can be supported and automated and this will enable entire new value propositions for process mining and automation, for example. Now we have seen a lot of the technical talks and a lot of research papers about the technical aspects. Object centrism is obviously in the forefront here. But what other than the technical aspects of the ICBM, are you looking forward to? So first of all, I strongly believe that object centric process mining and all of the image and AI topics are two of the major streams that will shape how process mining will be done in the future, in industry and what research will look like as well. Well, that was, thank you for this question. This was one of the very reasons why we conceptualized the industry track because we see that process mining is much more than just technology. When you're used and when you adopt process mining in organizations, we talk about socio-technical systems. So there are people involved, there's culture involved. So there are so many different facets that need to be mastered in order to be successful with process mining in industry and that's why we have a lot of different focus sessions for example in the industry day and yeah, so be prepared to learn something. Max, thank you very, very much, enjoy ICBM and also thank you for hosting it. Thank you. So I am Marlon Dumas, a professor of information systems at University of Tartu in Estonia and Chief Product Officer Ara Promore and I passionate about improving business processes, everybody knows me for my motto, everything is allowed except killing, doing something on ethical or illegal and not improving your business processes. Well, I think that makes three of us who are passionate about improving processes and Marlon. First of all, welcome for the second time on our show. So for everyone who's listening, just like it was the case with John Mark, you can go back and find the full episode with Marlon on the topic of prescriptive process mining we recorded with him about a year ago. However, today Marlon, your topic is you are a keynote speaker here and you will have the topic of walking the way from process mining to AI driven process optimization. What does it mean? Absolutely. When I was thinking about this title for this very important keynote because you know the keynote at the industry keynote that ICBM is one in a lifetime opportunity, I thought what could be the first word when it comes to AI driven process optimization. And I said no doubt it's walking, yeah, because if I think about the biggest mistake when it comes to absorbing AI innovations and many other innovations, not only AI wants, is wanting to jump or wanting to run towards it. It's very tempting. You see this latest toy out there and you know you can ask some questions and it looks like it can answer them and you immediately think that that is going to transform your business process. But as Socrates said it very well, he who is not content of what he has will not be content of what he could have. And what happens is that if you are not able to improve your business processes with what you currently have in your plate, maybe with some effort, but if you cannot improve the processes with what you have in your plate with all the capabilities that process mining offers with all the capabilities that existing predictive process monitoring and simulation technologies offer, you know, you are not going to get substantially more. It's not that because this technology is a is 10 times more capable that you're going to get 10 times more than before. So and that's what I call walking the way and I'm going to be basically showing a pyramid of capabilities and how you have to really see this pyramid. You have to build the foundations on process mining very, very strongly so that you have the right data to be able to move up into predictive analytics into prescriptive or suggestive analytics and into proactive or augmented analytics, which is what we have with we see coming now. So with the advent of tools like chat GPT and all these predictive AIs, there's been a big boost or a big interest in this AI topic. My question is how is AI changing the world of process mining? It is changing it and is going to change it in two ways and as a process owner, as a process manager, as a driver of process optimization, you can choose this to pass up to you. As Ford said it, if I had asked it people what they wanted, they would have asked it faster horses. And that's exactly where I think that tool vendors are pushing process owners and process analyst tours. They are kind of trying to tell them that now, you know, instead of clicking somewhere or instead of looking at a picture or instead of writing some query, that now they can type a question in a short language. And if you think about it, what kind of questions you will ask matters a lot. If the questions you're going to ask is where is the bottleneck in my process or you're going to ask the tool to write a data preparation script at best, you will get a faster horse. You're going to be doing the same analysis you are already doing in a different way, possibly faster in the best case, possibly cheaper, right? But that's not the bottleneck. If your bottleneck for improving your processes is that you are finding time of a process analyst to do a four hour session over your data to discover where the friction points are, then look elsewhere. That's probably not where your bottleneck is, you know, your bottleneck is really in estimating the impact of changes in deriving a asking questions you didn't know you had to ask, for example, around compliance. So the other path is to see a lens as an opportunity to do something better than horses. The car and LLMs are and in that, it's a matter of the questions you ask, yeah. LLMs allow you to, the LLMs are incredibly good at something, it's recognizing the context. What process are you talking about, what activities are you talking about, what KPIs are you talking about? Therefore, the questions where they will add you more value is in terms of thinking about like what else should I be considering, what other improvement avenues should I be considering in my in my process. So that is where you should be heading if you want to take maximum value of this technology. Now if you have not built your foundation strongly, then you will not be able to go there. So for one of the reasons is because LLMs are if you want fast hallucinating horses and you need to take away that hallucination out of them. And one way you can do it is by confronting what the LLM tells you against the data. So rather than using the LLM to query the data, you know, use the LLM to give you ideas on what to query and then go from what the LLM is proposing to you to the data, then you're going to derive qualitatively higher value than what you can currently do with existing technology. I can already say I'm sad that we cannot spend more time with you Marlon today, but all I can say also is enjoy the ICPM conference. I'm looking forward to your keynote. I'm definitely going to be there and good luck on your on the exploration of the intersection between AI and process mining, also in your academic domain. And thank you very much for the air time. My name is Jean-Marc Iriot. I'm happy to be here. I'm part of the Manon Hommel group, taking care for process mining and execution management. So we basically started with process mining, I think three years back. And I'm happy to be here in Rome at the ICPM 2020 free. Jean-Marc, it's very, I'm very happy to see you finally in person again, although we already met a few times before, but also to see you on the other side of the table in front of the microphone. So welcome for the second time on our show for the listeners. If you want to know in detail what Jean-Marc is doing and how we end up doing what he is, go back a few episodes and check the full recording with him. I think it was a very, very good episode. And today, Jean-Marc is here at ICPM talking about business value aspect. What will you talk about? We will talk about basically, first of all, I will give you some examples how we are doing it at Manon Hommel. So we will talk a little bit how we approach value, how we try to estimate value at the beginning of an MVP. So before we are doing any developments. But this is just to give some ideas because later on we will then talk about and ideate together about what could be value categories, how to measure value categories. And then of course, later on also based on an example I am giving with Manon Hommel, is to talk about how is a governance model, for example, looking like how to set up a reporting. And that's basically what we want to do to ideate together and see different maybe also use cases and different solutions because this is really depending on the company's culture what is here, the suitable approach. Now the ICPM is always focused a little bit on the technical aspects and the new innovations that are happening in the space. But why is the business value aspect also super important? I mean, we are a small and medium company, so German middle shunt. And for us of course, doing these huge investments, we need to make sure that we are having a kind of a payback of course. And that's why we are strongly focusing on value. And what I learned I would say in the last year, two years basically that many initiatives are somehow struggling because they cannot really prove value. What kind of value you are heading for? And it's difficult of course for some to prove value and that's why initiatives might lose momentum and maybe then die later on. Let's end on an example. Could you give us one example on how you, let's say, calculated your ROI of a specific use case and we don't, as usual, we don't need any specific numbers. I'm just really curious about the thought process and about the steps you actually take to get you from okay, we have a use case to let's do it. Yeah, I mean, we started not with an ROI, we started with a full business case saying okay, we expect these license costs, we expect these implementation costs. We also put into the calculation the setup of a cent of excellence and putting this together gives you a good understanding about the cost situation across the next years. And by doing this and having at least a minimum return on invest, then of course you get an understanding about the value you need to achieve somehow. So this is exactly what we are doing. We are breaking down the values or the savings we need to have on the yearly level. And then we are tracking it month by month. If we are then somehow close to what we promised or what we want to achieve or not. And this is of course then giving us a transparency and keeping up in a positive way of course kind of a pressure to deliver. All right, let's end it on the last question which would be what is the one recommendation you would give to the listeners and active practitioners of process mining. I would say try to bring a couple of horses into the race means start with some use cases parallel in case one use case is dying because maybe the assumptions are wrong because this will help you then to show somewhere quite fast value or will support it to show value or that short. So spreading your eggs among many baskets. Yeah. Exactly. John Mark, thank you very much for joining us for a little session here today and enjoy the conference and good luck with your talk and lecture. Thanks a lot. All right. Next to be here, my name is Raphael Garcia. I am a principal director in Accenture leading the practice on process mining within the Europe region. We have special focus on the German, Swiss and Austria. My background is process mining since 2004 and happy to be here today. And we are very happy to have you on as well. Raphael, you are spearheading today a discussion about interface aspect. I guess me, Patrick and also all the listeners would ask, what is the interface? There is a very good question and interface means here the connection of process mining with other technologies and methodologies which are out there. So for instance, if you think about process mining as a data lattice thing, right, you can connect that to lean as a sigma to say, well, how have I been to the knowledge to improve processes? You can go to data analytics and go say, oh, yeah, I want to have predictive analytics. Right? And that's an interface to process mining. So you're using the same data, you're doing the same thing, but actually connecting to process mining as we know it as an explorative fashion, right? So this is what we're going to be doing there. And our goal, eventually, is to understand actually what is missing today, right? Because we always speak about interfaces, we always use process mining. Oh, yeah. What is missing? All right. I'm going to go to that. That's what we're going to do. We know about interfaces from programming and why they're so important, but can you tell us a little bit about why interfaces in that domain are also equally as important? Right. This is always always a good question to answer. We're not talking about the interfacing at the technical level, don't we? It's not like we're used, for example, the same databases or how we structure the data in an object-centric way. We're talking about methodology interfaces, right? So how the topic of process mining as such connects all the topics, it could be from modeling to improvement over automation and data science. And what is the value generated by those interfaces, which are currently not that day? So just to give you an example, right, today, I mean, today, not today, but we've been thinking about process mining and simulation. So it's like a combination made in heaven, it should work. But if you look at the current tools, you don't have that many power to simulate things. So what happens at the end of the day is that you have one tool for process mining and another one for simulation. And you don't have a tool chain, which crosses all of them together. So, you know, we'll present to a board thing, for example, right? You need to switch tools and go here and there, et cetera. And this is what we want to explore today. You know, what is missing on those interface? What you would like to have, as you don't have today, and maybe we can put in a bridge, in a kind of letter, and send out to Salonis or someone else, for them to do some features on that. You asked yourself and the audience here a few times, what is missing? So what is missing according to Raphael? What is missing, in my view, these are two things, right? So the first one is a stricter, or an ostrichter, but a connection between processing movimental dollars and process mining. I'll give you a reason for that. So most of the engagements that we've seen in the past, firms doing this, at some point the guys in a forum and say, yes, I do have a data model, it's very solid, it's representing my process, ends of what, right? So how can I bring that to life? So I start doing some things, but it's more like a try and error thing, right? And some of them are more advanced in terms that they use, like they've been doing process mining, not process mining, but process improvemental dollars for a long time, and they understand the power of data. And these guys have from 0 to 100 in seconds, the odd ones take a lot of time to get there. So this connection could be way better, single out, be in the academic world, which we're here today, right? But also the technical world, right? So we can have dashboards, how we can get tools, which have you doing this, you can just maybe any eye to do this, right? To come up with interesting questions to answer and maybe ask those questions and answer them too. This is the one thing. The other one comes to, you have very solid database, right? In process mining you have everything, which is the good and the evil at the same time, right? But then let's go to the good thing. So why don't you use it more and more for predictions? I mean, we have, we learn, we could learn, let's call it this way, 10 years of transactions way in the past. I mean, if that's not a good training basis for process mining that I don't know what's going on the best, right? We have all the decisions in there, we have everything, all the parameters. So actually to design future process and also to manage the current processes and executions, it will be super good to use this database, not take the same decisions and wrong decisions again and again and again. So why not using this? And if you see the current landscape and people using it, it's not there, right? So this is what we're trying to do today here. Raise awareness of what is there, what's not there, what could be there, and what must be there. Well, Rafael, I'm sure there must be a lot of things that are interesting ahead of us. And I know that there is. So thank you very much for joining us and enjoy and good luck at your sessions today. Thank you very much. My name is Doc Farland. I'm associate professor at Aintov University of Technology. I joined at university almost 13 years ago and since then I've been doing research and process mining. Derek, welcome to the show. Very happy to have you here. I actually think that the first time we met or at least we've seen you was two years ago. I see him in Aintov when we were slowly, slowly exploring the amazing growth of process mining. So welcome to our show. And Derek, your topic and the theme that you will be discussing today is a collaboration aspect. And I was wondering what type of collaboration are we really talking about? So we're talking about collaborations between researchers and industry. Now industry has many different players. On one hand, we have the software vendors who are developing process mining solutions and analytic solutions. And on the other hand, we have the practitioners who have the problems who kind of look at everybody to help them. And the question of collaboration is how can we facilitate these three parties together to work better together to move the development of process mining solutions forward? Now is this a question of a technical challenge or what are you looking to address here? There's a lot of challenges. Anything everybody first looks at the technical ones but these are when you try to get the collaboration going, usually the ones you come to last. So it often starts with organizational, cultural aspects. Researchers think very differently work, very differently than a software vendor who has a tight schedule and wants to ship products, whereas a practitioner who is in his own world of problems that you need to understand. So this is one of the problems that we see takes quite a lot of time and you can't solve through technical means but just by talking, listening, understanding what the other side wants and say some form of empathy and kind of giving the other side some freedom to do what they think is necessary to do and engage with it in a right way. What does a perfect collaboration look like in your eyes? I think I've seen now several collaborations going quite well. So I see a pattern that I can suggest a bit and there's a time where you have to start from some mutual interest so you have to kind of see each other in the eye and say, okay, you have a challenge that I find interesting I want to work on. And then it actually takes about a year of talking to each other regularly to understand what is the problem that the other side really has, what are the common interests to also understand the language that the other side speaks to understand which words they use and why do they use them and what are they after. And also what, at some point, technical solutions come in, what that does mean. So one of the things I've learned at some point is whatever fancy technological solution we invent on the research side, somebody will come with a question and where is the business value now. And it is surprisingly hard to answer as a researcher. So this is one example where you need to need this time of regularly talking about the ideas to see where you meet in the middle and in the middle is not a compromise but where the things come together to be productive. So this is one aspect of a well-working collaboration, the other is of course you need access to data. And for practitioners, it's usually difficult to share data. It's also difficult to understand the data. So what works really well if practitioners are willing to invest in paying somebody from the research side that can be a master's student or PhD student to embed them inside the organization so that they become part of the company, have access to the systems and work with the team's data to have the problems to understand what actually needs to be done. And then you can solve the problem within the environment that needs it solving. So these are two very important ingredients I've seen working well. Right. So it's putting them in the context of an actual business value, then context order, understand what the research actually is supposed to do. Yes, exactly. Exactly. This is what this does. This from the business side, something else is required, then typically researchers will come with a technological solution that introduces technology elements. The company has not worked with before. So the company also has to have the willingness to engage with that, to absorb that, integrate that into their existing way of working, into their software stack, and there's a lot of other challenges that come with that. But this openness has to be there, otherwise you generate an idea, but then you stop with okay, what do we do with it now? Derek, if you would have to give one recommendation or advice to practitioners, such as ourselves, consulting companies or companies that are applying process mining in their own way of working, and the advice would be how to engage with academia. What would you tell them? Ooh, this is a very good one. So academia likes to talk about the things they're doing, right? And when you as a practitioner want to find a partner who can work with you, I think you have, it's a bit like dating. So you have to go out, talk to the people, what are you doing? Find out whether there's somebody who's doing something where there is a matching interest and then kind of point that out, right? So you have to pull on academia a bit to share how the way of thinking is. Yeah, I think that's one gives me almost an idea on developing an app like a Tinder where you just match with the academic topics and practitioners that would need it. But to be fair, it needs a bit more talking than just looking at the picture and swiping. Derek, all I can say is thank you for joining us and I wish your session to go very well and that you make some breakthroughs there as well. Good luck. Thank you very much. So my name is Elham Ramazzani. I actually lead, I have two hats at KPMG, I'm from KPMG, I have two hats at KPMG, I have a global role and a local one. As my local role, I'm based in the Netherlands and I lead the data analytics team for KPMG platforms, which includes Microsoft, SAP, ServiceNow and Salesforce. And as my global role, I lead the Centre of Excellence for Process Mining at KPMG. So our process mining community globally. Elham, it's a real pleasure to have you on the show. We've just finished the, basically, we did the conclusion of the workshops that were happening in ICPM, there were multiple streams at the same time, organizational one, cultural one, IT and so on. You were together with Lars Rankamire from Salonis hosting the organizational one. And there was a conclusion, basically, of three bullet points. The first one was that the process mining is not a self-fulfilling property. I do like this point. Can you elaborate on what is behind this? So process mining, no company actually starts process mining because they want to do process mining. And that also means that there is a whole lot of other things that are ready to be in place, that needs to be in place to make a process mining initiative successful. And so, no technology really excels and increases the adoption if it's not governed properly, if it's not linked to business priorities. If the value out of that technology is not measured, is monitored and if there is no ownership for it. That process mining is also no exception and it takes a lot of other organizational aspect, cultural aspect, collaboration to make it a success and harvest the value out of it. So it doesn't operate in a vacuum? No, that's it. So one of the bullet points that you had is science works, organizations don't. What do you mean by that? Yeah, because we're always thinking that the technical challenges and the technology edge is basically all about technical challenges. But when they come to organizations, there are so many other aspects that are required to make that work and therefore as a scientific problem, it's a technical problem, scientific problem, there is a procedure, there is a way to tackle that extended border of knowledge and extended border of science and tackle those. But organizational are very complex phenomena and I think that that's the reason behind it. The last point you had there was organizational aspects are crucial to move beyond insights towards action and value. Yeah, that also relates to the previous point as well that insights by themselves are very nice to have but then always comes the question of so what and to make that more tangible, making them actionable, of course there are quite some technological solutions which is great, which is helping. But in addition, there needs to be a governance structure in place, there needs to be a proper demand management, there needs to proper knowledge management, community building. Delivery should be best practices out of delivery and launching, rolling out use cases needs to be captured and there should be governance over your data management or data pipeline. And the way dashboards are prepared, the way analytics are created. So all of that is required to have a sustainable growth, to have a sustainable and scale up. If you don't do that, process mining stays as a very small initiative, but to make sure to scale up, help with the adoption of process mining, these elements needs to be there. Yeah, and of course the, for me the one of the most important things is, which I see at my clients is that they are process mining initiatives scale when they link it directly to business priorities, look at the company, what are the burning platforms, what are the biggest initiatives enterprise-wide initiatives, and how you can hook the process mining initiatives to those bigger programs and use that momentum to scale and, yeah. Alham, thank you very, very much for coming to our show and yeah, good luck on the further journey with helping organization adopt and achieve the value with process mining. Thank you, and for having me. Okay, hello, everyone, my name is Michal Arosik, I'm a former CPO at Minet, a process mining vendor that was acquired by Microsoft, so right now part of the, part of my role at Microsoft is being a PM architect, so kind of shaping the product in the direction to help drive value for the customers. Michal, welcome to our show, it's really, really pleasure to have you here. So you've been talking today in the workshops about the IT aspects of a process mining initiatives, of a process mining project, and as a result of discussion with academics, but also practitioners that are participating in the conference, you came essentially to three conclusions. The first conclusion was that a good data engineer should tell a story, could you elaborate on what you mean by that? Yes, so this was really a beautiful and fruitful discussion in the workshops or in the focus sessions, and the storytelling was one of the topics that we encountered, basically not only data engineers should be storytellers, but we are all here, we are seeing a lot of researches and use cases where we hear stories, basically the data should tell stories and everyone should participate in telling the story because you need to interpret the data to the end customers and they need to find the value in there. So, when I started in the process mining community like 10 years ago or more than 10 years ago, we've been hearing that sentence that was mentioned at the workshop as well, where you show some outcome to the customer and the customer says, "Oh, this cannot be true, this is not my process." You have your data, bad, you have your algorithms bad, but it's actually their data. So we were really talking about how to persuade the customer, how to line the data to the outcomes and to move forward to something that we see in the other areas like BI, where this is not a question, like you look at the sales report and you do not question the data, you do not question the outcomes. And storytelling was one of the solutions and maybe even using the newest technologies like generative AI and so on could help here because these are storytelling technologies and the possibility to do overcome. I'm glad you brought up AI because that leads me to the second point that you have outlined. But Gen AI and LLMs make a good sparring partner, so sparring with whom and for what? Yes, so in the session we had the nice mixture of people, really, from academia, from consulting companies, practitioners and as well as software vendors. And I think it was one of the remarks or quotes basically from a person from a consulting company that we see all of this hype, you know, like Gen AI, basically, LLMs used everywhere. Like we want to push it everywhere. But we need to wisely think about where does it make sense, right? And one of the topics was that if this person as a practitioner is driving the process mining initiative and doing some consulting with the customer, it's kind of looking at a process that might be very, what's a custom for that or specific for that customer. Sometimes these people do not know where to start, what kind of questions to ask. So the sparring partner was someone or some technology that could guide the person basically inside of like in the analysis, make the route in the analysis or in the process optimization initiative kind of being someone who will ask questions also back in order to find the right insights. And the final conclusion of the workshop was that the data lineage should be baked into process mining solutions. So what can we imagine under this term? Yeah, this is kind of connected with the first topic. So this is exactly the, let's say transparency between the data itself and the outcomes. Because of course, all of the vendors are showing the visualizations and so on, but the main idea is to be transparent also in this field. When we look back at the AI, it is always good if the results of some AI algorithms or some machine learning algorithms are explained are kind of augmented, so that the user knows. And this is the same case here. So we need to know where the outcomes are coming from. Sometimes it's hard because there's so much data that and we were tackling this topic as well that we can look at the technologies, we can look, we can kind of test the outcome and say, okay, this makes sense. This is really something that that is correct. But as we see the, how the technologies evolve to what extent we are looking at the LLMs with billions of parameters, are we not going to jump over a cliff like where we just need to trust the technology. So that was one of the questions. But for now, the data should be lined and the data lineage should be inside the process mining solutions because that's how the adoption of the customer would benefit. Mikhail, thank you for joining us. I know I can say it's best to flag with further developments of the Microsoft solution. Thank you very much for the invitation. So you've heard it. Those were our guests. I hope that we brought you some of the latest trends and that you enjoyed the unusual episode because we sure did not only hearing everything that they had to say, but also being able to meet them face to face, it's quite rare occasion for us. As you probably know, we do record remotely. It would otherwise be a very expensive show if we didn't have to go to every guest that we have. So it was real pleasure not to only do the recording live here in Rome, but also to have a little bit of chit chat with all those people before and after recording because those are the moments when you, you know, you get to know the person and it just feels different than doing this remote live. Yeah. And I should also say you get to meet people not just because of their work and the talks that they hold, but also, you know, on a personal level and, you know, having just fun conversations with everybody, what they're doing out and also outside of process money. There's some really, really interesting things that you're here. So this ICPM has always been a good experience for us just to put a face and a body really on all these researchers and guests that we've only been able to see in small corners in our laptops. Yeah. Unfortunately, we have to end on a little bit said news and said note because we've got some news for you together with Patrick. And first of all, it's been quite right. We've been producing quite a right here. We've been producing this podcast for almost three years. That's true. We brought you roughly a little bit over 70 episodes, which means like 70 hours of content filled with process mining, golden nuggets. And I'm really afraid that this ride has come to an end for at least, you know, for now. Because we have decided to go on. So we will, this will be the last episode for some time and there are plenty of golden nuggets, as you said, in our previous work that will still remain where it is right now. And that will be our catalog for the foreseeable future. I want to thank everybody that has listened so far. Everyone's been incredible and very supportive, should be said that, yeah, you guys are great. We're listeners, dear listeners. And yeah. Bye for now. Feel so almost emotional. Yeah. I can almost seen Patrick's tears because he's sitting on the other side of the table. Thank God we don't do video podcasts, because otherwise. But as we said, it is what it is. It's been really, really fun. We are proud. I can say I'm proud what we've built. I'm proud how many of you this show, this podcast reached. And you know how it is. All good things must come to an end. This might not be the last time you hear from us. We might come back. We just don't want to put any pressure or deadline for us. And we simply want to recharge. And overall, rethink the whole concept and hopefully come back stronger than ever. Absolutely. So watch out for whatever we do next. And here, maybe see you around. Yeah. If you have any questions or just want to reach us out, you know, you can send us an email on mining your business podcast at gmail.com. We're also on LinkedIn. So we'll be very happy to hear from you or, you know, comment on the content we've created so far or ask about the guests anything really. Because overall, this is about you. This is about the community. We grew together with you and we are proud and happy to see where we are right now. So thank you again for any time that you really listened to us. It was actually pretty cool because we met a few people who knew us, who knew our voices and told us about their stories and where they are listening to us. And it's really crazy to see these people in real life and see that there are actually the people that are listening to the show. So these have been really lovely occurrences. And once again, thank you very much. So this is the end. This is the last episode. If you liked it, as I said, just text us, leave us a review. And well, I would say usually I say, talk to you in the next episode of mining your business podcast. But this time I just say bye-bye. And thank you. Thank you for everything. God, that's awesome. [MUSIC PLAYING]