What Process Mining Actually Reveals About Your Infor System

Transcript

Jeremy Stoltzfus
Hi everyone, thanks for joining us today. Today we are going to be talking about Infor Process Mining. My name is Jeremy Stoltzfus. I’m a principal technical consultant with RPI Consultants.

I’ve been with RPI for just about 9 years now. Overall, I’ve been in the Infor Lawson space for over 25 years, working my way as system administrator on the client side for a couple of different clients, and then moving into consulting.

With that, just a couple of things that we’re going to go over today. So we are going to do a Process Mining overview, discussing what exactly it is and what it can accomplish.

We’ll look at the industry product catalog, which is a very important component of Process Mining, and then process insights. That’s where we’re going to spend a lot of the time today, kind of digging into the different insights that you obtain from the tool itself.

First off, what is Infor Process Mining? Well, Infor Process Mining is a component of the Infor Velocity Suite. So, there are several different components, and the Process Mining piece is just the one piece that we’re going to focus on today. Infor AI Value Plus, catalog, as well as RPA or robotic Process Automation is our other components of that Velocity Suite. But today’s focus, like I said, is Infor Process Mining.

So, what is Infor Process Mining and how is it used? What’s it used for? What’s the purpose of it?

There’s going to be a couple of different components. We’re going to start out by Discover, right? So, when you’re looking at your processes, that you’re, you’re running in your system today. You want to Process Mining, it is going to extract that process information and the data from your Infor enterprise system.

So, whether it’s your ERP, CRM, or whatever it is, the Process Mining will extract that, the relevant and the important information to do its thing.

Next, it’s going to Process Mining tool will analyze that data. So, looking at a detailed view of how those processes are actually performed. In the system as opposed to an ideal, well, this is how we think we should do it. Now, this analyze tool is really kind of built against the standards.

So, as the software is built and as the software is intended to be used, how does your organization match up to those out of the box processes?

And do you have a lot of outliers, those types of things. So, it’s going to perform  an analysis to help you understand how you are using the software, if you’re using it efficiently, effectively, if there are areas of improvement.

And then finally, it gives you the ability to optimize. So it’s important to note that the Process Mining tool itself does not do the optimization, but really the kind of the intent and the life cycle of how you use Process Mining is to give you those insights from the analysis portion that the Process Mining tool does, and then you take that information and you can ask questions.

You can apply filters and look at specific scenarios within the insights tool. And then, you have the ability to really focus in on where those bottlenecks are, right?

So, looking at the process as a whole and understanding, hey, the bottleneck is really up front here in this process, or it may be towards the tail end or somewhere in the middle, right? And it allows you to really kind of take that data that you absorb from the Process Mining analysis.

And understand and look at your process, look at your organization, look at all the different factors that may come into play. You as, as process owners, really know and understand the business and, and know and understand how the software works and how the software has been implemented at your organization.

So, it gives you the insights, and then it’s up to you to kind of take that information and determine, OK, what’s our number one priority? Where do we focus our efforts? To really optimize our solution and make sure we’re using the software the best way that we can.

It is important to note that the Process Mining tool is really kind of intended for those process owners, like I just talked about, right? You know and understand the business processes.

A lot of times this may be your project implementation team.

So, the ones that really kind of have helped to build your CloudSuite environment and know and understand those processes and how they’ve been put together is from an implementation standpoint.

So now you’ve been live and you want to say, OK, well, you want to know and understand, are we using the software efficiently? Are there ways that we can do things better?

Where are those problems and what’s causing us, slowdowns or bottlenecks?

I mentioned this, so next up is the industry process catalog. The industry process catalog is a repository of processes, and these are multi-level business processes. These are complex, but they’re tied to the Infor standard application roles. 

So as you walk through a process in the system, whether it’s on the HR, HCM application or within the financials and supply management application, there are different processes.

We’ll look at a couple of different examples, specifically, but think about how your different users throughout the system are doing different components of a specific process.

And like I said, we’ll take a look at some examples as we kind of walk through here. With Process Mining, these are preconfigured solutions because the Infor software has best practices in mind.

These process catalogs and the Process Mining metrics and the data that you get out of the system are really kind of geared towards how the software is written, right, hence those preconfigured solutions.

That really speeds up your implementation time of, of implementing Process Mining. So you don’t have to worry about, OK, well, let me define my process map. Let me, let me specify how the, the business is going to be running, right?

That’s all. configured within the system and during your implementation, you should be advised to best practice and how the software works. So we’re really lining up your business processes and how you are using the system with those preconfigured solutions.

And then it also contains these, these playbooks, right? So these are templates. It’s supporting documentation, it’s methodology, like I said again, based on how the software really is intended to work.

As we dive into that industry product catalog, right, so these are industry specific. So all of the different verticals that 4 supports, right, are out here with the various products or the various processes that are specific to that particular vertical in this example, if I am presented with this, I may, obviously, you’re, you’re not going to see when you go into the product catalog, you’re not going to see all of these.

You’re only going to see the one that may be relevant to your organization. So, for example, the IPC or industry product catalog for healthcare. So, if we drill into that, you get the specific processes that are relevant, and these are those solutions, right, those pre-built solutions that you have access to that Process Mining will compare up against.

The biggest one, a lot of times when we talk about a, a process. Is the source to settle or requisition to, to payables check, right? So that’s a big process that spans multiple departments, multiple roles, many users involved in multiple steps of that process.

And so if I just drill into this, and again, this is part of the process catalog, that’s, that’s part of that the Process Mining solution. So I just picked this source to settle. Process, for example, right?

There’s multiple components of this looking at demand, looking at par locations and counts, requisitions, contracts, sourcing, all of those different components. So, I’m just going to drill a little bit further, another step in, and just focus here on requisition to receipt.

But obviously, you can see the steps out after that. So, whether it’s inventory transactions that get created or subsystem postings, you know, those types of activities.

But we’re going to just drill into the next step, which is requisition to receipt. And you look at this then, and then we see, OK, well, even the requisition to receipt is compiled, comprised of multiple processes and multiple components.

Are those, are those requisitions being manually created? Through a user running, requisition self-service, are they being interfaced in? How are the purchase orders coming in?

Maybe you’re not using requisitions and you’re starting with purchase orders. Are those things coming in through PCs, all of those different components. In this specific one, I’m just going to look at the requisition to approval. So that’s just a very small subset, obviously of this entire process.

So, when I look at that specifically, now we get to really where, where the meat of this is, right, where really the heart of this content is. So in this particular example, I’ve got the two boxes there outlined in green.

Those are what we refer to as events. So, Process Mining is event driven. So, in this particular example, we’re looking at the creation of that requisition as an event and then when that requisition gets approved or rejected, that is another event, right? So, we’re looking at the events in the process as a whole.

Now keep in mind we have just drilled into this specific example here, but Process Mining is, going to look at the entire process from start to finish and all of the different events that are associated with all of the different paths and different components of the process that we looked at on the previous slides.

So, we, when we get that understanding and understand how these processes are, are delivered and how, you know, this shows us how the software was really intended to be used, right? We’ve, we’ve got that understanding. We get to, we get to the point of like, OK, well, this is all event-driven.

How do we go from here? What does Process Mining really give us? So that’s, that’s these questions, right? So, what does Process Mining answer?

What are we trying to get out of the system?

So just as a couple of examples here, how long on average does the process take from start to finish? Or how long on average does it take between event A and event B, right? I talked about those specific events, and the process from start to finish, right? Your process may start with a requisition.

And then probably end up with a payables check going out the door to pay that vendor. Well, how long does it take between those pieces, right? Is it 20 days? Is it 30 days? Is it 60 days, whatever that is? Process mining can tell you that and can answer those questions.

You know, looking at between specific events, right? So we know that or maybe you know, within your organization, hey, our approvals take way too long. Let’s focus specifically on that approval process, and you can drill in to that information.

The beauty of Process Mining and how the tool works is you can apply filters kind of on the fly and do some deeper dive analysis, right?

You start kind of high level, and then you start identifying the components of where those bottlenecks go. And as we get into later in the presentation and look at the insights that are available out of Process Mining that’ll start to make sense.

You know, and then we look at and, and we get that, that kind of sense of really where it can be, where this has becomes extremely valuable.

You know, continuing on the list of questions. Well, how many variants of that process? Do all of your requisitions go out for approval? Do all of your invoices get approved? Do some of your invoices get approved? Are purchase orders being canceled on a high volume? Are they being created an error? You know, what is the iteration? Are things going back?

From requisitions, maybe it’s a requisition and you see a high volume of requisitions being rejected or returned to the initiator because they didn’t fill it out correctly or there was information missing, right?

And so, then you identify that and that becomes a training opportunity, right? And those are some of the outcomes that you might get out of, out of this discovery and out of answering these questions from the Process Mining tool. Many outliers to the process. So if your organization defines the process to be steps A through H, and somewhere in the middle, somebody did a step M, you know, that that’s kind of an outlier.

OK, let’s look to see what is that outlier doing? How has it impacted the process timeline overall? Right? And then, of course, any combination of these questions, right? So how many outliers to this process, but is that faster than what we think our standard processes, right? And you can take a look at those start to finish and how quickly we can, like I said, apply these filters and go through.

So, as an example here, a couple of examples I’ve got showcased here just to kind of one on both the one on the HCM side and one on the financial side. Right on the HCM side, we’ve got recruits on board.

And we think about that process as a, as a whole at a real high level, what are the key events that would happen in that process? So obviously, you have to start with creating a job requisition.

A job requisition has to be posted. And then as you’re going through that, you want your applicants to apply. How long does it take, you know, to get through all of that process, and you got interviews and the offer, hire and onboarding.

Of course, there’s a lot of steps in between, but these just kind of at a high level, showcase those, those main components. So those are the events that we would focus on.

On the FSM side, on the financial side, we got requisition to check, right? That may start with that, creating that purchase requisition. That requisition needs to be approved. And then it has to go through a purchase order, receiving the goods.

You get that invoice in from the vendor then and how long, you know, until that invoice gets paid, right?

So, as we’re looking at all of the events that are part of that process, what Process Mining is really able to do is tell us how long between the various events. Are these, are these taking, right?

And so, like I said before, it’s really helping you to identify the bottlenecks in your process, in your organization. We take this a step further, and now we’re going to get into some of the process insights that are available.

So, as an example here, I’ve got just a small subset of that whole rectal check process, right? But really kind of show you how the process works and to give you some of that insight, give you some of those insights that become available as you define this process. And again, using the, the pre-built solutions, right?

So, looking at this particular example, we’ve got an event when a purchase order gets created, and then when you create the purchase order, obviously, you have to put those lines on the purchase order.

So those are a couple of different events that are tracked. Optionally, that purchase order might need to be approved. We talked earlier about requisitions. Sometimes purchase orders are approved, especially if you may be interfacing your requisitions in or, or that type of thing.

And you, those necessarily aren’t being approved. But in this example, we’re looking at just the purchase order to goods received. So, you can see there we’ve got that, we’ve got that little break. You’ve got the decision point to say, hey, does this purchase order need approved? And if so, it goes through that purchase order approved event.

And then you get down to the purchase order being printed or, or, or sent out, and maybe that purchase order is getting confirmed from the vendor. And then finally, you get your goods received, right?

The vendor says, OK, I’m satisfied with this purchase order. We’re going to send you, we’re going to ship you the goods, right? So again, we’re leaving out pieces of this just for simplicity’s sake.

But we’ll, we’ll dive into some of those. So again, each one of these boxes along this little, along this workflow are the events. And when those events happen, we’re recording the date and time stamp that event took place, so that then I can then use the insights to track how long is it taking me to get from point A to point B.

At the heart of Process Mining, when, when the data is extracted and, and pulled into the Process Mining utility, Process Mining refers to these as cases, right?

So, think about a specific in our example, a specific case, maybe a specific purchase order line, right? So, understanding how long does it take for that specific purchase of an item to go through that process. Maybe not all lines get approved or need approval, or maybe some lines get canceled or, or are taken off of the purchase order, right?

All these things as you’re looking here, these are the events. And so, when you, when you extract and you, you begin pulling this data in from your, from your ERP system.

It’s, it’s calculating or it’s grabbing the date and time that that event for that particular, data element happened in the system.

So, in the example here, I’ve got a bunch of different PO lines that have been created. I’ve got some attributes about the purchase order line, right? So, I might have the purchase, would have the purchase order number specifically. I would have the line, I have the item.

So not only do we get the summary, but we get the details as well. You can drill into a specific purchase order, and identify those outliers and take a look then at, at your source system or try to understand, OK, well why is this an outlier?

What happened along this particular process? Why is, why is this kind of why is this a pain point for us? And as you, if you, you double click on any one of these cases here, it brings up a little bit more specific details. So, this particular record,

I’ve got the second record selected, and I can see that I’ve actually got 5 events associated with that one purchase order line. So, I can see when the purchase order was approved, and see when it was printed and sent, confirmed, received.

In this case, multiple goods were received. So, it was a partial receipt initially, and then a secondary receipt to receive the remaining goods. So again, this case table identifies the things that we’re tracking and the things that Process Mining is using to track the process as a whole.

So, that is part of the insights. It gives you that ability. And of course, you can see, you can filter here and, and, you know, obviously, you want to start with the high-level summaries, and, but it does give you now because this is available as you, as you build out your, your dashboards and, and like I said, there’s a lot of pre-delivered content, but you do have the ability to modify those if there’s other specific things that you want to drill into.

But next up, let’s start at a high level. So, average cycle time. So overall, you know, at the very, the top part of this view here, I’ve got the average cycle time, but then underneath, I’ve got it broken down by company.

Now, it just so happens that these happen to be similar, right? But there could be outliers. So, is one company performing much better than another company?

And it really depends on how your organization is, right? So out of the box, you’ll probably get an overall, but if you need to break that down further, by company or maybe it’s specific departments, or, or something else like that, right?

You can, you can break that down. You can kind of slice and data, slice and dice the data until it meets your organization to really get out the information that you need. Looking specifically at this, right? I’ve gotten this broken down by two different companies.

I’ve set a goal of 30 days, right? You can set visual alerts. For example, I believe I have this set where it’s anything over, over 20 days. It flags it as yellow. Maybe you want that flag to 20 days is yellow, 25 days is red, right? Because it’s near, it’s too close to our goal.

So, it gives you some visual flexibility here. There are a lot of things that you can do to configure, even a simple metric like this, right?

If you’ll notice on each of these widgets or each of these little components on this dashboard, right, you’ve got a filter icon. So, I can dynamically filter this data on the fly.

I can build these dashboards with pre-built filters in them. So, for example, the Company 10 and Company 20 is already there, and is already defined as a filter. But it does give me the ability to add additional filters, right?

So, cases with a specific flow, am I looking for certain, like point A to point B, or am I including or excluding certain events? Or do I want to look at certain attributes?

In this case, I’ve got these two metrics at the bottom split out by an attribute, which is company. And so that’s what allows me to kind of prebuild those filters. Or I’ll kind of drill in on the fly to get a little bit more depth and understanding.

Another, view of that average cycle time. Well, let’s break it down. Let’s break that cycle down by month and see what affects our cycle time.

And you can see on this chart that there was a spike in January. So what happened in January, right? You may know, for example, that, hey, one of your buyers retired at the end of the year, and now suddenly your remaining buyers have an additional workload, right? And so now because of that, they’re not processing and getting through things as quickly as they used to be.

So, this would be good data to show, to say, hey, we really need to replace that person.

There’s a lot of other factors there too, but we can, we can drill in and again apply another filter. In this case, I have applied a filter to say a specific process flow. So, when I’m talking about a process flow filter, that’s, that’s, that’s a refining our search to say, hey, in this case, show me when show me when PO line created, and remember those are our events, is directly followed by an approval.

For when a process flow or when a purchase order goes through an approval process, we can see that we have a spike.

Well, maybe the organization, maybe your organization implemented a new approval process, right? And so maybe that’s why we have a spike.

I believe it’s the month of February here in this particular example. And then maybe there’s a drop off, but you can see beyond that, there’s still a little bit of a higher trend as far as our average cycle time goes within that process.

This might help you to understand, OK, well, suddenly we implemented this process, and now things are taking much longer to go through the process, right? So you might want to reevaluate your approval process. Do we have too many approval levels? Do we add a lot of complexity to it where we didn’t have before, as an example.

So that’s just an average cycle time charts that you can get out of. Obviously, there’s a lot more in-depth that you can go with custom filters and things like that. There are bar charts, pie charts, all of the different charting options really that, that you can, you can dream up are available.

Next up we’ll look at the Process Explorer. So, what is, so by default that’s Process Explorer will show us our most common path of the process. So, if you look at this example, what we’ve got here, our most common path, purchase orders lines are not going out for approval.

You can see it’s being created. And then going out, being printed or sent out, it’s getting confirmation, and then we’re receiving goods. That is our most common path in our process.

But what happens when you expand out the number of events or the number of connections, then this is where this starts to get a little bit more interesting, right?

Now I can see if I expand my number of events out. I can see that a large portion of my purchase orders are now going out for approval. Now, it’s not the majority of my purchase orders, but there are still a significant number, right? If, depending on what your organization says requires a purchase order to have an approval, right, you may want to start looking into those purchase orders. And why are those purchase orders now?

If that’s kind of not our normal process for a purchase order line to go through an approval, approval process, why are those purchase orders suddenly being approved? Why do we have so many, right?

And drilling in, you can see the image on the right. You can see we’ve got some POs that are being canceled. So why do we have such a high number of POs being canceled, right? Are they being created in error?

We do not have all the information before that PO got sent out. And so, we can, we can kind of drill in and remember with the case table, we can now filter out and say, OK, show me purchase orders where the event, where a canceled event occurs, right? And so, you can drill into those and specifically get a little bit more information. So that’s the Process Explorer.

It allows you to really kind of expand out the, the number of events that may, that may exist in, in your process. Looking at another view is the variants, right? So, I said our most common variant was those going through without approvals, but now we’ve got a whole bunch of different variants that may not follow the straightforward process, right?

And you can, as you select different variants, and you can layer these variants here it tells you how many instances of each variant and what path that it follows. In this case, I’ve got a couple of different variants selected.

So, it’s showing me multiple paths, but typically, when you select just one variant, it’s going to show you that path, right? So that gives you a little bit of insight. In this case, I can see variant one.

Now that’s, that’s not your most common. That’s the first option, your most common path.

Again, remember, um, from my previous slide that went straight through without approvals. But now variant one includes approval, right? So, you can kind of get a sense of what the different variants to the path are, the deviations for the path.

It’s a very powerful insight into seeing how the data is making its way through the system.

Next up we have conformance. Remember this image here, and I know it’s a little bit small, but this image here was that original process that we looked at with our, our events. So this is what we said is our standard process that we’ve got. As you select your variants, then you can see what path those variants might have taken. Now, in this example, I actually have an outlier here.

And it’s, it’s interesting as you go through the application and you select your process, and you select your variant, and it draws this dynamically for you to show you the process and it highlights it in red to say, hey, this step is not part of the standard process.

And looking at it, you see, OK, well, I know it’s hard to read here, but this is actually purchase order canceled. That is going to be a thing that really is part of the process. We don’t want it to happen often, but still, it is going to happen.

So, we don’t necessarily want it to show as an outlier. So, what can we do? Well, the tool allows you to say, hey, I’m going to create a new process diagram based on this variant.

And when you do, it will immediately redraw or recreate that process. And so now you can see here at the, the tail end of this process, there’s a branch to determine if the PO is canceled or not, right?

And so now I can use this as my standard process and reanalyze the data to find the outliers. Now, we still may not want purchase orders to be canceled on a regular basis, and we can at least drill into that and see, OK, why so many big purchase orders are being canceled. But it does allow you to kind of dynamically create that process on the fly and reapply the analysis against, against that process.

So just kind of going through. And now I can rerun the analysis and look to see, OK, well, now how many outliers do I have and, and kind of do it dynamically, on a real-time basis.

Just some summary and final thoughts as we finish up here. Remember, so these are preconfigured. These are industry-specific solutions. It accelerates your implementation.

It’s important to note that these models can be extended. So, if you’re using custom steps or you’ve got custom events in your, in your process that, that are unique to your organization, or you want to pull in additional attributes, those can be done that can be extended to really help you get the insights that you need.

The dashboards come pre-built, pre-configured as part of those solutions, but they can be dynamic. You can adjust, you can modify, you can completely copy and replace elements on those, apply additional filters and that type of thing.

And that gets us to the next point quickly and easily apply those filters on those process insights. So, to be able to look at specific examples and to drill in and to really gain an understanding of, of those processes and what’s actually happening.

And to find those exact bottlenecks, right? Help, helps you differentiate between the one-offs and the trends, right? OK, do we have these outliers? You know, we looked at the variant explorer, right? Look at the different variants.

OK, we’ve got a whole bunch of variants, but they only happen once or twice. So they’re probably not affecting our overall timing. So, but you could look at those and really kind of make that determination.

But then the beauty of, of this is you build these, to meet your, you, you tailor these, right? You don’t have to build them because they are our solutions, pre-built solutions, but you tailor, tailor them to meet your organization’s specific needs.

And then you rerun the analysis every 3,6,9 months, right? You do an initial analysis, and you say, hey, our approval process, we implemented that and it’s caused some problems. It’s added several days or, or even a week or whatever time it added to our overall process.

And let’s focus on that. Let’s look at our approval process. And let’s make some changes, right?

So then you implement a change, and then you let the system kind of generate the data for the next 3 to 6 months, and then you reload the data into the Process Mining against with, with your process change that you put into place, some sort of change that you implemented, Whether you simplified your approval process or you changed the selection criteria of what POs need approval or that type of thing.

You’re kind of configuring those, the initial time to kind of help you troubleshoot, and then you reload that data with, with the next 6-9 months’ worth of data and see, OK, we’ve implemented this change.

What’s different between the last time we ran this analysis and now? Did it make a difference? Did it make a positive impact in our process? Right? So being able to set up and reuse those insights on a regular basis to monitor your overall performance and usage of the system.

And with that, I’ll wrap up. Thank you so much for your time today and attending. And as always, if you want additional information or have questions, please feel free to reach out to us and contact us at questions@rpic.com.