Paramita: Hello everyone and welcome to PwC Luxembourg TechTalk. Today we continue with our Data and AI series and have a quick chat with Daniel Trautmann on process intelligence.
Paramita: Hello Daniel.
Daniel: Hello Paramita.
Paramita: Welcome to our show. Like I had explained to you, we are doing this series on data and AI and we are treating all the different subjects and topics that we can think of within this vast area, zone. Now before we start, just before coming downstairs I asked two of my colleagues can you tell me what is process intelligence. Who knows process intelligence? One said it's a buzzword.
The other said process intelligence is intelligence about processes.
Daniel: Not so bad the answer.
Paramita: And what she said later on probably makes sense because what she said was that apparently you gather intelligence to make processes better.
Does that sound good? What is process intelligence Daniel?
Daniel: Well yeah I mean again the answer was not too bad.
What we try to achieve with process intelligence is to address actually a problem we all know. It is to address a problem to identify the gaps between the promise and the reality of a process. And so let's imagine you go to a restaurant and you have in your menu card a lot of pictures of the different dishes available and then you order something and what you get is actually not what you expected to get. And that's what we try... That's what you also see within processes that you think you know your processes at the company. But then when you look at your reality it's much more diverse and much more different than you thought it would have been.
Paramita: So what kind of processes are we talking about?
Daniel: Well we can talk about any kind of processes. I mean that's the beauty of process intelligence that we could talk about the purchase to pay process at a company. We could take a client on boarding process, HR processes, IT processes, change management, stuff like that. So any type of process which is actually supported by IT systems where you have data available... we're talking about specific event log data. And well if you have those systems, if you have those data then you can apply a technique called process intelligence to actually identify those gaps between the promise and the reality.
Paramita: Between fish curry and sushi...
Daniel: Yeah. Yeah exactly.
Paramita: So it's basically a technique to identify the gap between what is expected and what the reality that you're getting and how will this... how being intelligent about your processes, how will this help businesses?
Daniel: Well imagine you are a COO of a company or you're working in the compliance function on the internal audit function. I mean no matter where you work you would want to understand how your processes work and what you have defined maybe 10 years ago in a policy or procedure is actually what is the reality in your systems. So and well beside understanding those processes you would like to identify maybe outliers, maybe transactions which do not follow the process like you imagined it would be. Like you have what we call a lot of rework activities for example. So you approved the transaction and then it goes back because there was a mistake and then you have to re-approve. And you want to be sure that then after modification has been done it really gets re-approved or the system maybe just comes to an end because it was approved initially already. So to identify those outliers or compliance breaches, also breaches with segregation of duties, to also identify maybe inefficiencies in your process because I mean if a process takes two days longer than expected it is a cost for every company, to benchmark against other processes for example.
Paramita: It's like an audit...
Daniel: It could be an audit. You could apply an external audit on internal audits and of course we're trying to use much more data and also process data to do audits but it's much more than that. Because you can analyse process from different angles. So as I said you could analyse it from an efficiency point of view where we are not interested in an audit but that helps of course business departments to optimise their processes, to be more streamlined, to be more efficient.
Paramita: You mentioned data and when I started doing research on process intelligence, I came across terms like data intelligence. So are they the same thing? What's the difference between process intelligence, data intelligence or data analysis? Because what you just explained process intelligence kind of seems like data analysis. I mean you gather the data and you analyse what is missing and how to make it more efficient. So what's the difference? Why so many different terms?
Daniel: True, there are a lot of terms out in the market and maybe everybody has his own perspective. I have mine. Process intelligence is very specific because we really talk about processes. Data intelligence you could apply to any type of data set because not every data set refers to a process. Let's take an example of a general ledger. Here we talk about transactions... It's data but it's not a process. Because a process would then refer to how to create a transaction, how to approve it, how to modify it and stuff like that. So you really have these activities and this is how process intelligence works is that you have a list of all these activities and also a timestamp. So meaning when they have happened. So you really have the right order and also the difference of days or hours within those processes.
And one of your questions was what is the difference between analysis and intelligence? It's that you can analyse a lot. Probably all companies have a lot of data. And the data can tell you a lot but you also have to use it. You have to listen to the data. So there the intelligence part steps in where companies actually use the data to make decisions, to make sure you have sustainable growth in your company.
Paramita: And how do we become "intelligent", how do we become "process intelligent"? I have no idea whether that makes sense...
Daniel: Yeah. That that makes sense because it's a long journey. Some companies are closer than others. Some are already there. And I mean of course you first need to find your processes. So you have to write them down in a policy and procedure. I mean most of the companies are doing that. But then of course you would like to support your processes by IT systems, by IT applications. You want to avoid that you do everything on paper. And then there is the issue of data quality. I don't know in the last podcast if there has been any reference to that. But every time you work with data, you have data quality issues.
Paramita: We spoke about that. We actually spoke of data governance and I learned that because you know when you have all these terms and terminology... I thought the data management and data governance are probably the same thing. It turns out that data governance is part of data management and data quality is another part of data management and I hope to talk about it in detail with another colleague of mine and probably in a later episode.
But yeah you were talking about data quality within...
Daniel: So of course data quality is I mean you have to make it right so that the data doesn't tell you something wrong. And then we talk about data quality. However there was quite a famous mathematician in the 19th century. I have to admit I forgot the name but he said something... Cannot repeat the exact words... But he said something like the errors you make with wrong data is better than not using any data at all. A mathematician said that in the 19th century and it's still applicable today. So even if we use data with poor quality with techniques like process intelligence or data analysis you can identify those quality issues when you try to analyse the root causes when you have issues. You can identify those quality issues and then you will be able to resolve them and to make the quality better in your data. And then you can step in and further analyse and really get insights in your processes. But of course to have clean data without any quality issues that's already a challenge for all companies.
Paramita: Yeah of course. Just out of curiosity... We have heard about all these you know are more or less I mean people who are in this sector, you know like us who work in this sector, we have heard about data analysis.
But we rarely hear something of process intelligence. Why is that? Is it not something that is considered important in businesses? And I'm sure that it's very important given how you explain what it is.
Daniel: I think it's at the moment it's more popular in operational companies because of course when you produce certain products, you're really keen that the processes run smoothly because every hour of delay is a real cost. In Luxembourg as we're talking about a big financial industry, it might not be that popular in there. But still I'm thinking of on boarding processes about AML or simply HR or IT processes. Also credit loan process. I mean every company has its own processes which are important for them. So it might not have been popular in the past but in my opinion it's a very powerful tool.
Another problem which could be a root cause for that is that the data is not available because when we talk about event logs which you need for process intelligence not all systems are keeping those logs. Because we are of course talking about massive data. It doesn't mean you cannot apply other analysis types but process intelligence would not be my choice.
Paramita: OK. And coming back to the how to... So there are tools right and tools and methodologies... There are standard methodologies? Or there are industry specific tools?
Daniel: I would not say that you have one standard approach because every company is different. So you should not try to squeeze processes within a standard approach. Of course, some processes are more homogeneous than others throughout several companies. So of course there are processes out there which we can imagine how they would look like. What we see on the market. But maybe companies say no that's not how we do it. And we think it's better like that. And then together we will try to find out what is the best approach and what angles from a process we would like to tackle. Because with those tools also you mentioned of course we are using tools for doing that type of work. They're very powerful and you can analyse processes from a pure compliance perspective if dedicated controls have been applied or what users have been involved in a process but also the efficiency angle… how fast or high efficient are your processes.
If you met certain KPIs at your company so you can also develop your customised dashboards with those tools of course. Segregation of duties I mentioned and also what party at the companies interested. So are we talking about audit or we're talking about really business operations or senior management.
And then sometimes when you apply such techniques and the objective is not really defined you're getting lost because the tool is so powerful that you do a lot of things. And then yes sometimes you get lost and you really have to concentrate on the objective. One of my colleagues said the tool we are using at the moment it's like a Ferrari. It can do a lot of things, it looks very nice. It's very fast but if you just drive it in the city centre...
Paramita: It's a shame...
Daniel: It's a shame. Exactly.
Paramita: OK great. And just to wrap this thing up is there any new development in process intelligence that we should look out for?
Daniel: Well I think the concept of process intelligence or process mining will stay the same. Of course, there will be new tools coming out which will help you to do it more efficiently, more effectively. But for me what will be very interesting is to also include it in other topics for example in robot process automation and robotics. Because before you want to automate your processes using robots you want to know if it's really that standardised that you think it is. And what process variants you might have. So you maybe don't need one robot but maybe 10 robots to do 10 different things. And also how AI for example could help us to even do better process analysis to know which option of a process it should follow or not. What would be the best choice. And to integrate it with services like that will be very interesting. Also what a lot of companies are telling me is that well their processes are not supported by IT systems or not the full part of a process is supported by IT systems. There is still a lot on paper. But now we also have the technology to transform this unstructured data which you have on paper to a structured file like an Excel file and then analyse it through techniques...
Paramita: Who are these companies?
Daniel: Well there more than you expect. Of course, everybody is trying to move away from paper but I mean every company has a printer in their open space. So there's a lot of paper being produced and yeah that's the reality but we're getting better.
Paramita: You used the expression "process mining". Is it the same thing?
Daniel: Yeah we were again talking about terminology here. I prefer process intelligence because as we said the intelligent way of using it is then to base decision out of it and to really have an action plan ready based on that analysis. Mining would just be to gather the data but then not doing anything or maybe not doing enough with it. So I prefer the term intelligence. But yeah when you Google process mining you will also find a lot of things which relate to process intelligence I'm pretty sure.
Paramita: Well perfect. Thank you so much Daniel for making me a little bit more intelligent about process intelligence.
Daniel: Well it was a pleasure to be here and yeah I hope it was not too technical.
Paramita: No I don't think so because I understood so thank you so much. It was a pleasure. Thank you.
Daniel: Thank you very much for the invitation
Paramita: So that was my conversation with Daniel. Hope you enjoyed the show, especially hope that you learned a little about process intelligence and we hope to see you next time for a new episode of TechTalk.
Commmunication, PwC Luxembourg
Tel: +352 49 48 48 5821