Even though this week was relatively a short week for me (as I was fortunate enough to spend some time with my family visiting the great city of London) I still managed to find a central theme in the various meetings I was having in the Nordics (where I was the rest of the week).
We all know that the process mining is a very hot topic (still is, even despite the fact it has been on the market, theoretically for as much as 24 year, but practically speaking for about 10 years now) and also process management is gaining traction again as a topic and the synergy between these two is what has been front and center in most of the conversations I’ve been in this week.
Let’s dive into this a bit, shall we?
For those who have been missing out of what process mining is, this is the technology to gather and analyse system log files and reconstruct the actual execution of business processes from these system logs. Process mining confronts you with the harsh reality of “this is how you have executed your business processes” (.. as far as we can tell based on the system log files, so excluding any manual activities or system activities that are not logged). Still, even with the disclaimer taken into consideration it almost looks like magic. It is incredible to witness the sheer mess organisations can make out of something as “simple” as executing a business process. Of course I am over exaggerating this a little bit, but the so-called “spaghetti” diagram do speak for themselves (blurred for data privacy reasons of course).
Next to the sheer discovery of process execution, process mining (at least the top platforms in this industry) also provide action engines, orchestration and proper analytical capabilities for process performance and conformance management.
On the other hand, we have process management. This has been around the block for almost 5 decades by now and has seen it all. From having process maps as it main (and only) deliverable, to becoming a single source of truth and on its way to be the baseline repository for business transformation, compliance or operational excellence. It is here that you can document, manage, maintain and govern your business processes as they should be executed. This is the on-paper equivalent of process mining’s ‘as it really happened’ process.
It’s not hard to imagine that the should-be, or documented process will look much cleaner and more understandable compared to the spaghetti like picture from unsorted/unfiltered process mining, and there is good reason for this as well. Documented processes are often used for consumption purposes throughout the organisation and you don’t want to confuse the living daylights out of your end users. They just want to know what they need to do and what role they play in the process. To refer to a great episodes by Die Drie Prozess Philosophen, I am on Daniel Matka’s side here that it would be useful if everybody has a common understanding of what a process actually is.
The magic starts to unfold when these two forces meet in the middle. This usually can take the shape of either one of two things, or both. You can
(1) overlay information gathered from mining on the documented processes in order to provide additional and highly relevant information on the state of certain parts of a process, or
(2) combine the documented process with the mined process and provide transparency on the topic of process adherence. In other words, how well has the organisation executed the business processes in line with the documented (or should-be) processes.
I will not go into more detail on both now (and will leave that for one of the following episodes) but this has certainly been the central theme for me for this week.
Finally, we all talk a lot about AI and I am a firm believer that if organisations want to leverage the power of AI effectively, they need to make sure that all of the necessary contextual information about the processes, combined with the insights about the actual execution is available in the data set that AI is using. This is, to me at least, the only way to make AI come up with remotely relevant and accurate suggestions.
Anyway, thanks for putting up with again this week and see you next time.