Road to Prediction: Predictive analytics in HR

Nico Bitzer

‍

Hello and welcome to the Bots & People interview podcast. We talked to inspirational people from the automation space and today we talked to Christian Müller who's Head of Process Intelligence at Sanofi and Nikolay Goldovich, Head of Data, Process, and Automation CoE at Sanofi Business Services. We talked about a super interesting use case which is about predictive analytics in HR.

‍

It was about the job change process and how to build process mining into the living process and predict difficult use cases that can be then prevented by the business team directly from the beginning. That's a super unique use case to me it was new to me that you can use process mining for such a thing and I learned about it from Christian at a networking event. Also keep in mind we have some changes planned to the podcast so you can stay tuned that is why you also do not hear Louise at the moment but myself right now.

‍

And one more thing, you can meet Christian, Nikolai, and myself at the 7th Intelligent Automation and AI Conference that takes place in Hannover from the 19th to the 21st of June this year. The key event theme is to measure return on the west of your automation tools and investments and many more topics.

‍

There will be companies like Deutsche Telekom, Sanofi, Siemens, Bayer, Akir, and more and as said you can feel free to come there and meet Christian, Nikolai, and myself to talk about also this use case.

‍

And now the best thing is you can get a 10% discount on your ticket by clicking on the link in the description or using the code at the website of Intelligent Automation and AI Conference which is BOTSANDPEOPLE in capital letters. And now I wish you a lot of fun with the episode.

‍

Nico Bitzer

I'm super glad that you are here. Tell us who you are and what you do Nikolai, what's your role at Sanofi?

‍

Nikolay Goldovich

So my role in Sanofi is the following. I'm head of data processing automation CoE. I think I say this phraselike three times a day already so I'm already learning by heart so I'm responsible for three areas data, process and automation.

‍

Nikolay Goldovich

automation I'm responsible for RPA bots so we are doing RPA with automation anywhere. The last year we built a company-wide service to serve the whole organization so we are in the shared services but we built with our digital organization a partnership and now the service that is running in terms of RPA it's a service from the shared service and digital towards the whole company where digital is providing technical capabilities and shared services providing mainly process analysis and identifying the bots and building the governance.

‍

In the data area, I'm responsible for data in finance so it means that I'm an owner of a financial data foundation or several data foundations and what we're doing there we are trying now to consolidate the data for the finance perimeter and to start building analytical use cases as I think this is where the whole shared service industry is currently going quite intensively in terms of building data analytics and then start using AI models, machine learning, predictive analytics but on the data on the data side.

‍

What is interesting is that we're using a very modern technology stack we're using informatic and snowflake with the data sharing data lakes so we're building in parallel a service now data lake, an SAP data lake, and data lakes for our other co-applications.

‍

In the third area is a process and in process I have Christian who is Head of Process Intelligence in my organization and here we have two teams one is responsible for process mining and another one is responsible for process design. so I basically have two technologies on  board one is the signavio for process design and the second one is QPR for process mining.

‍

Nico Bitzer

Right and Nikolay I mean we are also in in conversation already for a long time and I know you are super well connected also in the process automation scene so probably we could send greetings to a lot of people yeah what attracted you to that field of process automation in general or digital transformation how did you get there?

‍

Nikolay Goldovich

So I think the journey started like seven years back I think when we did one of the biggest implementations of RPA in this part of Europe in Prague with Siemens. And since that time I'm only working in this area playing with the different technologies and working with the different technologies. So in principle, in the whole circle of intelligent automation, you will not be able to find the technology that I was not responsible for.

‍

Nikolay Goldovich

I was responsible for chat bots for intelligent document processing, for process mining, for data analytics, for AI, for RPA, things like that. So in the end of the day, now I'm closing the loop of the digital transformation being responsible for I think 90% of different technologies in the back office

‍

Nikolay Goldovich

I like is that I think that these technologies they are growing and we all know how RPA has grown in the last 10 years, which is quite amazing and I think that process mining and now process design also start growing and I think that we all know better than me history of Celonis, how they  have grown in the last I think 10 years. And the same as waiting for the data topic because I think the data was really undervalued in the back office so there were some finance people playing with the data with forecast and that's basically that's it. And now the data

is going more in transactional and operational work. And when I'm talking about data, I'm not talking about excel based KPI, so I'm talking about really like prediction analytics, machine learning, what is going on in the next step of the data.

‍

And in process, I think it is far away from being super mature on the market because I think that a lot of people are doing process mining but the way that they're implementing sometimes some kind of weird because either they implement process mining because they heard or they liked the idea but they did not pack it  properly in the company so there is no the whole structure of continuous improvement of people who can take it and start doing something or they do not know there is this process and they cannot really, they do not know there to be process and they do not really able to compare where do they go with this. So they of course, they optimize cycle time, they optimize automation but they never ask and answer the question why, where I'm going and why I'm going there and not backwards.

‍

Nico Bitzer

I like that and we will also get to the topic analysis paralysis later on with when talking about the use case. What I particularly like about your approach is that you bring those three topics together data, process intelligence, and process automation. I think this is super strong on could be also a blueprint for many other companies.

‍

Nikolay Goldovich

Leave it in tthe podcast, I will send it to my boss.

‍

Nico Bitzer

Yeah, okay, okay. We’ll send a link and it's interested to benefit from it. I think I have metsome companies that have similar approach. And also what is interesting that they are also then started the automation center of excellence or the center of excellence from a shared service center perspective. So maybe that's also good good approach. And so as far as I understand you're leading those three topics and Christian, you are responsible for the topic process intelligence and leading that practice.

‍

Christian Müller

I'm Christian, head of process intelligence and as Nikolay said, responsible for the topics of process mining and process design.

‍

Nico Bitzer

Maybe you could take a little bit deeper and how have you set up the whole process mining initiative at Sanofi.

‍

Christian Müller

Yes, so we started with process mining in 2019 with the first use cases, which was accounts payable process for P2P. And we already had in mind that we like to scale big. So it was our first pilot, very proof value for process mining. And after that pilot, it went very quick so that we implemented process mining projects also for other service lines.

‍

Christian Müller

And we were able to do very good analysis here. And at a certain point, of course, we wanted to use the functionality of process compliance. So we were very curious about the “as is” and the “to be” comparison.

‍

Christian Müller

that was the point where we said, okay, let's look into the service lines, how the processes and the core models are designed, how they are documented. And here we figured out that each service line was on themself on how to do it. So there was no central support on how to document the process in a standardized and structured way.

‍

Christian Müller

that was the moment where we started to look into the process design topic to really support our service lines to document and to manage their process according to the Sanofi standard and not letting all service lines finding out by themselves on how to do it. So we leveraged this experience across the Sanofi business services.

‍

Nico Bitzer

And how did it started? So you said that, Nikola, you have this three streams in the center of excellence with data, process intelligence and automation. Was it then planned? Okay, we have to build up that three exact teams. Or was it that people got enthusiastic about it and grew more organically to care about this topic and to split the team like this.

‍

Nikolay Goldovich

I think, first of all, it was growing naturally, but from the very beginning, so I think the story is the following that we started with RPA, and then after some time of doing several boats and stabilizing  RPA service, and Christian was not part of my team, he was preparing, he was alone, by the way. In the certain period of time, he also joined, because we said that, okay, let's consolidate all our digital initiatives together, or transformation initiatives, so at least not all of them, at least some of them.

‍

And then, as long as we started working on the process mining, it was clear that data is playing more and more important role in this area, and we started looking into the KPIs, we naturally started looking at KPIs, and some KPIs doing manually, somebody is doing it with some company, and there was no harmonization, and then we understood: ok, we need to look deep into the KPIs, and then, then idea came to start building service for data, and this is how the data topic started growing. So in the end of the day, this was naturally, and then, and then when this started growing, we started looking with Christian at our co-models, I think very important topic is to mention that, in Sanofi, we very much into the driving of the co-models with our service lines, and the co-models structure was already predefined, so we know, and we have also Sanofi Proces book, where we documented all our entrant processes, not all, but at least the majority of the processes, and the definitely for the back office operations, most of them. And then, we started thinking how to drive harmonization of the co-models together with all these tools, and this is how Signavio came into the picture, and this is how we started growing all together.

‍

Nico Bitzer

interesting, that it grew so naturally, and Christian, if Nikolai says he were alone before, what exactly have you done, and how, maybe you can explain how you came then together and joined the

‍

Christian Müller

yes. Yeah, so I have a digital background and HR background, so that means I made digital stuff for HR, so it was like implementing workday and these kind of projects, and also in designing co-model, so I was part of the design team, and I was leading the project of implementing the co-model on

‍

Christian Müller

I was project manager for this, and

‍

Christian Müller

the automation team, the topic of process mining will be assessed,

‍

Christian Müller

say, finished with my project, I was looking for a new project, and when I hear process mining, I volunteered to be the project lead for the first process mining initiative. And that's why it

‍

Christian Müller

standalone project to prove value, and when it was realized that there is value and there is potential, I was moved to the digital team, to the automation team.

‍

Nico Bitzer

interesting, and now I know also Christian, that you do some networking at least, and we met in some networking events as well, and you are a real ambassador for digital topics and also for process mining. What would you say is Sanofi doing differently from other companies that do process mining? So

‍

Christian Müller

to others who are doing process mining, I hear very often that they have good analysis, good results, but then it's the question of how to get these results and these

‍

Christian Müller

into actions and into value. And I had a very nice exchange, and we have a bit with someone, and during that conversation, it came out like with process mining, we figured out something like smoking is not healthy, so no one will question that. But just the information that smoking is not healthy does not stop anyone to smoke, and with process mining it's sometimes a little bit the same, so you show it, look here, this is your process, there are the loops, there are the bottlenecks, everyone is super excited looking at that, but nothing is happening

‍

Sanofi here is very good to bring this also then to the process level and make the connection with continuous improvement to turn the insights we get from process mining into

‍

Nico Bitzer

very interesting, I have this one app on my phone is tracking my calorie intake, I know it's all the steps and every evening I think, oh no, 1000 steps were missing to reach the tenser or no I ate 3000 calories today because of that cake in the afternoon. That's what you're explaining right so tracking alone is nice and could also side and action. Very interesting. Not

‍

Christian Müller

but is the transformation.

‍

Nico Bitzer

that gap between in. then I would say the topic of the podcast is road to prediction. So let's also a dive a little bit deeper into that

‍

Nico Bitzer

very interesting use case. We also talked about it in one of our leadership roundtables. And also I received some messages from people after the roundtables saying that was a super interesting use case. So maybe you could again explain what it was about. Then we can dig a little bit deeper

‍

Christian Müller

Yes, absolutely. So it was about process mining prediction. And we were really looking for a use case. We hear very often this prediction topic from the market, but we haven't met anyone who did it. Let's say so we said, okay, let's look for a use case to do this

‍

Christian Müller

And we went to our HR services colleagues and we looked into the job change process. So we hear that there are some difficulties with a job change process. And job change process means all kind of changes during the employee lifecycle. I

‍

Christian Müller

executed by this process. For example, promotion, demotion, let's remove to new position these kind of job changes. And we hear that the main issue with that process is the effective date.

‍

Christian Müller

imagine I will receive a promotion by beginning of next month to me as an employee. It is super important that my business title is updated and my salary is decided of course. So there is a huge risk on employee satisfaction.

‍

Christian Müller

the process in the system was such low that many cases went late. So that means the completion of the process was after the effective date. And to mitigate this, there was a team was following up this manually. So the processes were monitored and it was ensured that business title and payment was updated. And this could cause additional manual work for the team to get this

‍

Nico Bitzer

Okay, so to to rephrase it, what job change means someone gets another job title and this has related to the payroll or another title. And sometimes this needs to be evaluated by a lot of people, approved by a lot of people. And this is in total job change process and this can be delayed.

‍

Nico Bitzer

of course, I can imagine that also people get really upseif they super long until their payroll gets updated or also the title because it pt robably shows up in a whole bunch of systems. I

‍

Christian Müller

the team was very good in avoiding these kind of situations. So we have not seen a lot of people. So we have never seen someone with with the wrong

‍

Christian Müller

our job title was yeah, followed up very closely, but again, manually.

‍

Nico Bitzer

And now how did you implement this process mining piece into the process? Could you explain that a little bit more? Because I think that's that's key,

‍

Christian Müller

Yes, yes, of course. So the underlying system was workday and it is important to know that

‍

Christian Müller

of that job change process took

‍

Christian Müller

some years ago and it was before the captive approach before centrally before centralizing of the processes. And that's why process was configured as is seven years ago. Let's say and that that was and at this time when the process was implemented, there was no sense of harmonization or standardization.

‍

Christian Müller

that means at this time the process was implemented, our sonobi was working. And that was now causing the issues. So we yeah, sit together with our colleagues from HR services, we looked into the process and we saw okay, how how to take this, how to extract the data and how to do the modelization.

‍

Christian Müller

And while discussing all the problems they have, we had the idea, okay, that is the great use case for the prediction analysis. So our idea was to build something like a weather forecast. So to apply historical data on current cases and to predict how they will go in the future.

‍

Nico Bitzer

so it means that the team also got access to the process mining software to look at this use case or have you from a central point of point of view, then nudging them or you send some automated notifications or how's that

‍

Christian Müller

the team access from the very beginning. So they have seen the dashboard from the very beginning and contributed to that. So the first version of the predictive was not that reliable. So we really needed to come back and now needed to ask the team for their feedback. So what is really the risk from your point of view and how to bring these risks into the prediction analysis. So we really needed to be trained.

‍

Nico Bitzer

And what are eventually? cases or attributes of a case that can lead to that kind of a longer job change process that would be super

‍

Christian Müller

For example, huge number of approvals. So depending on the countries we have seen a to eight approvals during the workflow. And also when, for example, there was compensation change, the process went to the manager's manager for approval. And what indicator was when it is already more than two days waiting for approval, it is very likely that this case will fail in the

‍

Nico Bitzer

Okay. Interesting. And maybe a question to you, Nikolai. I mean, that's HR. So isn't there also we apply predictive analytics to HR? And I know it's a bit of a provoking question. Isn't there anything to look at GDPR and also clarify the things with, it's the second thing maybe with the workers council, for example. But we did it.

‍

Nico Bitzer

So you and how that's I think that's interesting. Because probably some companies still hesitant because predictive analytics can of course be also critical.

‍

Nikolay Goldovich

but first of all, the

‍

Nikolay Goldovich

you analyzing? So are you analyzing employee performance or are you analyzing the process? And then the question is, if you analyze employee performance, do you look at it on the aggregated level or you see it on the names level, on the people level.

‍

Nikolay Goldovich

So for sure, working council will never allow, especially in Germany or in France, working council will not allow you to make people or human performance analytics something like that. So this will no go. So that's why even in the system, we do not capture the names and the ideas of the people.

‍

Nikolay Goldovich

So everything is unanimized. And we strictly, strictly prohibit to all the use cases to bring personal identification in the systems, all where unanimized them and mask them. Because otherwise you will really, really go into the troubles with the GDPR and with working councils.

‍

Christian Müller

As the only German speaking colleague, I had the pleasure to present this to the works council.

‍

Nikolay Goldovich

That's what I said. That's what I told you. We did it.

We did it. Yes. Christian did it.

‍

Christian Müller

Yeah. Yeah. When to the work is cancelled very early. So even before we made the first pilot, we showed to them what process mining is, what we plan to do. And there is of course an agreement. It's no question on that in ins and off. It is absolutely a no go. We will never ever load personal data. Also, the analysis is our focusing on process on the global process. And we're looking at two years data from the past two years with no possibilities to drill down to individuals.

‍

Christian Müller

And in general, the work council reflected that they really appreciate that we plan to improve our processes. And we also agreed that the work council can come around at any time, at any working hours, and can say, okay, so we are now here. We didn't inform you before. Please stop whatever you do and show us a little bit around in process mining. We like to assess you.

‍

Nico Bitzer

just I once had a pleasure in an unknown company at an unknown time to implement their travel travel and expenses software as a project manager. And then as soon as workers, cons of our hearing, the word artificial intelligence, they would love to set up. a lot of meetings about this and what it is about.

‍

Nico Bitzer

It's always a bit critical traffic. That's why also I like Nikolai, how you also split it and said, it's about the process or just predictive analytics in how the process is running and how there might be bottleneck next and not how the performance of an employee would be in future. So that I think that would be really critical, right? If you would say, based on the data, we gained probably this employee will be not that good anymore

‍

Nikolay Goldovich

But this is the in honest, this is one of the topics which we as a company or we in the back office are not yet there. But this is one of the biggest problems that in general, we're all who are working with the data have because in the end of the day, the borderline or the red line between what we do good and what we do really bad is a very, very thin and

‍

Nikolay Goldovich

to go to the dark side of the story and to really start analyzing things that you really do not want to analyze. And of course, I mean prediction of the performances very extreme, maybe it will not be that easy to predict somebody's performance. To evaluate somebody

‍

Nikolay Goldovich

based on which articles he is reading in intranet and how often he's doing something this is super easy at the moment. You do not need to be genius for that. Yeah, it's also. Yeah. So I think that this are all the topics in in in front of us as long as the digital 3 is to see what is inside and how

‍

Nico Bitzer

it's

‍

Nikolay Goldovich

will be always people who would like. not always in the right direction. But this is also applied not only to the corporate environment, the same applies for our daily life. I mean, you never know, okay, you will not name any company, will not point out on any company, but you never know whom you digital tray, where you are scrolling or where you are searching in the internet, will land on which table.

‍

Nico Bitzer

Christian, to get back on also to the use case, how do you ensure reliability of the model and does it still need some effort from your site now to make it better and better? Or would you say it's now set up and running and we are we are done with it. So

‍

Christian Müller

so how we ensure reliability is we are looking at last two years of data and we are refreshing this each day. So that means we always look about the last

‍

Christian Müller

that's why we can really rely on the data and we can really see how we are improving. The lessons were

‍

never ever use process mining as a KPI tool

‍

Christian Müller

Analyze processes and not generate some process KPI's for management

‍

Nikolay Goldovich

And don't build a data model inside of the process mining tool that it will not be reusable for the other kpis

‍

Nikolay Goldovich

I think this is a very interesting topic and I can tell you that one of our common friends, Nico told it to me when we were chatting.

‍

He said, I told you, don't build anything in the process mining. The process mining is only for visualization of the results and to show and to show your how your data model based on case attribute and based on timestamps looks like. But for now, nothing else. Don't do this mistake. Who did it

‍

Nico Bitzer

Yeah, but I think still, I mean, you are pioneer in that case and I find it super impressive that you went out with such a use case and such a usage of process mining. And if then, of course, it needs some iteration. I think it's also quite normal.

‍

Nikolay Goldovich

that when do we process mining, I think this one of the topics that when you do something like this, you need to have in your mind the big picture of everything what you try to do and this was in that time, it was not available. It was very use-case driven. So tool-imp lementation-driven. And now when we started digging into the data, we see that it should be done completely

‍

Nico Bitzer

Okay. And now as you reworking some of that parts, but also the use case itself was super successful for HR. Do you plan also to roll out similar use cases like this? Or have you planned to identify potential in other departments like not only HR, but procurement, finance and all the other interesting

‍

Christian Müller

I think it is also important to mention that the result of the prediction analysis was complete redesign of the job change process. So when we started to show up all the problems of the system and be predicted all the potential issues, we also had the idea to, yeah, think about what or to think about emails, to think about automated workflows to get these issues mitigated. But that was really the point where we said, okay, let's stop, let's say automating a process who is not good.

‍

Christian Müller

So let's cut this down and redesign the whole process. And here we had really very great improvements. So I check I have the life, there's point in front of me.

‍

So we have improved from 16 days in average to three days the cycle time with the new designed process. And we have now created the dashboard where we can see how the process was executed before with numbers. And now the more and more growing data with the new process and how the improved opportunity was to answer your question on further use cases.

‍

Yes, we have some use cases we see in front of us, not in the HR and HR services environment. So here we also plan to do further process mining analysis, but we have not seen predictive use cases here. So we really like to do classic

‍

The case use case we have seen is an internal control. So we have internal controls where on a quarterly basis data is extracted manually into Excel to identify some compliant breaches. Let's say very manual.

‍

And here we see the opportunity to make an alert right in the moment we see that the process is deviating from the path from the path which is compliant. So there we definitely plan to use this opportunity, this technology. And we also look forward to look into the service now topic. So to see okay background here is the user experience. So what are the root causes for a bad user experience? What are the any kind of patterns we the user experience is

‍

Nico Bitzer

identify and we can highlight before? Very interesting. And then also to talk about predictive analytics, and also this approach predictive analytics and process mining in general, where do you see in general potential? Not only for Zanofi, but also for other companies in using that kind of technology or use cases.

‍

Christian Müller

You mean for prediction? Yes. Yeah. All processes where data is available and where data can help to predict cases for being bad. I need to say I do not have concrete examples, but I would say all the processes where ongoing cases are at risk, and this can or need to be pointed out. The predictive analytics is the right technology to look into

‍

Christian Müller

a different here is when you do classic process mining projects and you go to the customers, everyone is kind of relaxed because you're looking at data from the past. Whatever you do, you cannot change it. And this is a totally different attitude when you show the top five cases which are likely to be failed this week, but it's not too late. And so the motivation is totally

‍

Nikolay Goldovich

I think that in general, prediction is the next step for the every process mining project, because in the end of the day, you can improve your process forever. But the ultimate person who is managing his process is not to improve the process, but to really ensure that it will not happen again in the future. Otherwise, you are constantly retrospectively changing something. So this is not...

‍

Yeah, so it should not work like that. So in the end of the day, if you have a process, if you have any process, then your task should be in future to predict the failure and not to change it retrospectively. So that's why the prediction in the process mining should be standard. It is not

‍

Nikolay Goldovich

this, you need to have something very sophisticated or something like that. I will give you very, very straightforward and still with example, constantly using. If your supplier sent to you nine invoices and the whole were rejected, believe me, 10th invoices will also be rejected.

‍

Nikolay Goldovich

Yeah. If you will not... If you will not predict it properly and will not identify the root cause and will not send the email and will not ensure that something is done, but it will be the same fail. And

‍

Christian Müller

all. So it is a totally different usage of process mining. So usually process mining is used as an analytical tool like an external measurement instrument and using predictive is using process mining as a part of the system of the process. And what I really find interesting here is then how to deal with the findings and how to

‍

Christian Müller

Does this really need to be done by human or maybe a pot can be activated to do

‍

Nico Bitzer

Yeah. And also what I like about it and also why I find it so fascinating is not... It's not too common now.

‍

So there's still... From my point of view, and we are also talking to a lot of companies more that analytics part in the focus and not the predictive part so that I like a lot about it. And that also you put the business more into account with that information.

‍

So it's not just a central department looking on the process and thinking, oh, we need to change XYZ. But it's rather giving the business the team that cares about the process, a real indicator and, hey, you see this could be late this week or this could fail. And really... Yeah. Yeah, go ahead. And really one way out of that analysis, parallel analysis to refer on that, what we said in the

‍

Nikolay Goldovich

Yeah, I think it's really... Yeah, we do not have a right answer to this question. Should it be driven from business or should it be driven centrally?

‍

Nikolay Goldovich

So we also kind... We think that we need to give more power to business, but we know if you give all the power to business then sometimes topics are not moving. So that's why it is... It is not that obvious how to do properly. And I think every company should find their own way and their own... based on their culture and based on their organization. So I think that we give a lot of power to business and we... when we... as a team, we try to be more on the

‍

Nikolay Goldovich

And sometimes it is playing really against us, like the honest, but this is how it was at the time, it was so interesting that time flew by very fast. And we still have the last part of the podcast, which is our question-wending machine. But we have three interesting questions to you where you can answer spontaneously. I will ask one to you, Christian. I want to unify. And start with Christian.

‍

Nico Bitzer

Yes. Christian, what is the thing that you should never automate ornnever process in all variation should not be automated. automated? that's a very good thing. So we have this question often, I like that you stay business. Some people say, are there a relationship to whatever my children are? But I like that you have a concrete example of a process that should never be automated

‍

Christian Müller

It's nice. It's impossible to automate the conversation towards my children. I don't have any kind of hope in that

‍

Nico Bitzer

see. Let's look into a good question for Nikolai, which one should I take? Nikolai, what historical person would you love to have a coffee

‍

Nikolay Goldovich

Considering current political situation or

‍

Nico Bitzer

can consider whatever you want to prepare.

‍

Nikolay Goldovich

would go for Churchill. I like this guy. So this guy was a real master. So what he did and how he did, I would really go and ask him how he was able to do all these things and to go with his country in the level during the war where he came.

‍

Nico Bitzer

probably a very interesting coffee talk. Yes, cool. Then, Christian Nikolai, thank you so much for the conversation. I think there were a lot of insights for our listeners. I super tuned to hear how it goes on with your predictive analytics journey, but I'll sustain touch with you. Thank you.

‍

Nikolay Goldovich

you very much for the podcast. It was also very interesting to discuss with you. You also have a very good overview of the market and also to tell what is going on. So it's always a pleasure. Thanks for

‍

Christian Müller

for having us

‍

Nico Bitzer

Thank you. Bye. Bye. Bye.

‍

Nico Bitzer

Hope you had a lot of fun in listening to this episode don’t forget that you can meet us three at the Seventh Intelligent Automation and AI conference, taking place in Annova on the 19th to the 21st of June this year. Besides super interesting topics like how to measure the return on invest of your automation initiatives, we will also as bots and people have the honor to host a panel discussion about the impact of ChatGPT in enterprises. If you're interested, visit the event and you can also directly use the link in the podcast description to get 10% off your ticket. Thank you!

‍