Up next in 10
Creating Value from Business Processes. Process Mining in Power Bi
5K views
Nov 6, 2023
What is Process Mining and what does it look like? How can I run Process Mining in Power BI and what are the benefits? If you've ever been involved in business process analysis, you know that a tremendous amount of time and money can be spent on interviews, workshops and analysis - without any real impact. With PAFnow, Process Mining for Power BI, you can automatically visualize actual processes in a fraction of the time and cost required by traditional methods. About Speaker: Daniel Hughes I help companies eliminate friction in business processes, saving millions, using technology that can be implemented in as little as 4 weeks! Conference Website: https://www.2020twenty.net/power-platform-virtual-conference/ C# Corner - Community of Software and Data Developers https://www.c-sharpcorner.com #powerplatform #virtualconference #csharpcorner #powerbi
View Video Transcript
0:00
I'll start and then we'll turn it over to Robert
0:05
Again, my name is Daniel Hughes. I'm the Director of Process Solutions for North America here at PATHNow
0:12
PATHNow is all focused on process mining and helping companies get the best value or the most value out of their processes and improve processes and streamline those
0:24
Robert Connolly is one of our data ysts here in North America and here in just a few minutes he will walk us through a demonstration of the tool and how you can use Power BI to look beyond KPIs and into the root causes of KPI issues
0:43
So I will start with an overview of process mining. You know, most people haven't heard of process mining, especially in the Power BI and Microsoft Power Apps community
0:56
Talk about what the use cases are, how it's used, where it fits into the stack
1:03
Then Robert will do a demonstration to show you how it works and what the use cases are
1:08
And then at the end, I actually give you a link to download and use process mining for Power BI for free for no charge
1:17
So how you can get started without even having to license anything extra
1:23
So process mining, what is process mining? Process mining is a methodology that's designed to deliver operational performance
1:33
performance, right? It was started maybe a decade or so ago in Germany. A fellow named Will Vander
1:39
Alst developed process mining at a university in Germany. It's based, what it basically uses the
1:49
data inside of your transactional systems to reverse engineer or recreate your business
1:57
processes. So instead of having to go out and do, say, subject matter expert interviews or ask your
2:05
employees what their processes look like, you can actually get an actual view of every transaction
2:12
that happened in your organization, learn how your business processes are transacted
2:17
and then be able to take action on that. So not just, hey, here's how we want it to look
2:24
But here's exactly how our processes are functioning. All right. Most of the use cases are typical business processes
2:34
Our number one business process is purchase to pay or procure to pay
2:38
Anything from deciding that you want to buy something to issue in a purchase order, getting approvals, sending that purchase order to a vendor
2:48
vendor sending back the goods and then vendor sending an invoice and you pay in that is a
2:55
typical purchase to pay process. Second most common process that we see is order to cash
3:02
So a customer sends you an order, you fulfill that order, and then you collect on the invoice
3:06
you send them. We also could look at something like quote to cash if you wanted to look at
3:12
everything from quoting on the front end. Beyond those two, we do a lot of work in internal audit
3:18
So maybe you want to audit your procurement process to find road buying or maverick buying
3:25
You want to find out, do you have internal controls issues? Process mining will allow you to sample 100% of your processes and then spend your auditor's time focused on the areas that there's issues
3:39
Other processes like plan to deliver, issue to resolution and customer support, customer journey mapping
3:47
Maybe you want to understand as you're moving online for your customer journeys from in person
3:53
Maybe you need to understand how they're interacting with your organization. Process Mining can look at your customer journey and show you what's really happening
4:02
From there, those are the most common. Those are common across industries
4:06
From there, we can dive into really specific edge use cases, maybe mortgage origination
4:12
We have customers that want to look and see how they're able to create mortgages
4:19
Are there salespeople that are causing problems in underwriting, that have underwriting loops
4:26
But maybe one of the most interesting ones that we have is actually at an airline in Canada where we're taking streaming data off sensors of an aircraft and recreating a plane flight
4:40
So the process of actually flying from the time you close the door and that's logged in a sensor to pushing back from the gate, taxing, takeoff, acceleration, altitude to even things like flight attendant call buttons being pushed and flight attendants going and taking care of passengers
5:03
So recreating everything that happens inside of a flight or inside of a business process
5:09
And it can be used in lots and lots of different areas. So typically, there are many opportunities for improvement when you're looking at a business process
5:20
If you're looking at procure to pay, you might want to improve days payable outstanding
5:24
We have customers that have improved their DPO over 10 percent. Or you might want to be more efficient or process invoices faster
5:34
So we have invoice processing time or automation rate that you could look at
5:39
or really common. Well, there's two ways when we look at payables, paid in their invoices
5:47
Some companies want to maximize their discounts that they get from their vendors and how much of
5:52
that they capture. So we can help you identify what bottlenecks in your invoice payment processes
5:57
allow you or cause you to not get your discounts. Some other companies want to ensure that they keep
6:05
their cash and extend their payment terms as long as possible. So process mining can look at your
6:10
payment process again and identify how long can you push a vendor and not pay them before they
6:17
start giving you penalties or putting you on sales blocks. So there's lots of actual real business
6:24
KPIs and metrics that you can improve by understanding your processes, understanding
6:31
really where the bottlenecks are in that process and how you can alleviate those bottlenecks
6:38
to be more efficient. So how does this all work, right? That all sounds well and good, I'm sure, but how does it work
6:46
Where do you need to get the data? And the data is actually really simple at the base level
6:53
We only need three pieces of information to be able to reverse engineer your processes
7:00
First, we need an activity name or a step or an event. What actually happened at this step of the process
7:07
It might be check order. It might be an event called create purchase order or create purchase request
7:15
What is the name of the step? Second, we need a timestamp
7:20
When did that step happen in this transaction? So basically every URP system is going to capture a timestamp of when a user made a change to an activity or an event So we capture that from your ERP systems
7:35
Then the third thing that we need is a unique case ID. What is it? What unique number or attribute can we use to track a case from the start of a process all the way through all the events to the end of a process
7:50
You know, for procure to pay, that might be purchase order number, that might be invoice number or things like that
7:57
So we get those three pieces of information we actually call an audit log
8:03
And we get those audit logs from your ERP or from your transactional system
8:08
We have pre-built data models for things like SAP or Oracle EBS, Salesforce, ServiceNow, Microsoft Dynamics are all really common ERP systems that we work with
8:21
And we go and create an extraction, extract data from your ERP system into your data warehouse or data lake, and then use our process mining algorithms to turn that audit log into a recreation of your process
8:36
So that's all well and good. That'll show you what your process looks like, how it's been executed
8:44
But a lot of companies want to do things like compare processes across plants or compare transaction or cycle times between vendors or on time delivery from vendors
8:55
To do that, we simply add what we call attributes or tagging data to our event log
9:02
So if you want to compare processes between plants, let's tag all our transactions with the plant number or plant ID
9:10
If you want to look at and compare vendors, then we'll tag all of our transaction with what vendor that's associated with
9:17
Or even things like resources. Maybe you want to, in an internal audit, identify where do you have segregation of duties errors, right
9:27
Whereas the same person performing two steps or three steps in a row when it's outside of your internal controls process
9:33
So we tag transactions with or events with with users or with resource IDs
9:42
So all of those by creating the audit log, tagging transactions with attributes allows us to recreate the process and then compare that process across any of the different scenarios that you might want to look at
9:55
So even though process mining has been around for a decade or maybe two, Gartner has never actually released a magic quadrant
10:07
What they have released are two market guides for process mining, one in 2018 and one in 2020
10:15
And PathNow has been included in both of those and recognized from Gartner as having robust process mining embedded in Power BI
10:25
But inside of those market guides, Gartner has identified about 10 different core capabilities you should look at when you're evaluating process mining vendors or comparing vendors
10:37
One is automated discovery. Simply, can you recreate the process from an audit log
10:43
Second is conformance checking. Can you identify what an ideal process execution looks like and then compare any other transaction against that
10:54
We'll talk a little bit later about how you can use the Power Platform to actually take actions on that
11:01
Next is data prep. Data prep is really important in any ytical application, especially it's important when you're looking at transactions or processes that go across systems
11:13
You know, procure to pay in SAP, pretty straightforward, pull data from SAP, recreate your process
11:21
But maybe you're looking at quote to cash and quoting is in Salesforce
11:26
You create your invoices in SAP and then you collect those in a supply chain system
11:33
So being able to data prep across those three systems so that you can track a transaction from quoting all the way through revenue or valid collection is really key component of process mining
11:49
And then as process mining has evolved, it started as a reverse look, a reverse mirror into how processes were executed in the past
11:59
Now that it's evolving into a dashboard, I guess, the ability to look at data in real time is really key to process mining
12:10
And by embedding our tools, our process mining tools in Power BI, putting it in the Azure Data Factory allows us to do real time, real time view into your process and create real time dashboards
12:24
multi-process interaction is one of the other things they look at you know procure to pay in
12:29
order to cash does not um isn't executed in a vacuum right you need to buy from a supplier
12:37
that to fulfill an order to a customer and then that supplier needs to deliver on time so your
12:42
procure to pay in order to cash processes actually interact with with each other in the supply chain
12:48
So being able to combine processes or just different processes into a single view is one of the new key areas
12:58
And customer journeys, right? Everybody has moved online over the last year and a half or so or year, I guess, since the pandemic hit
13:08
And customer journeys have become even more important. And you would think it'd be pretty simple, but customer journeys can be very, very complex
13:15
Robert and I two weeks ago worked with a manufacturer that was looking at their online customer journey and we found 845,000 different variations of their customer journey process
13:31
So being able to handle that volume and also that amount of variation in your process is not as simple as you would think
13:42
So being able to handle all that is one of the key components of process mining
13:49
And then as you move from a rear view to real time to a forward looking view of your processes, being able to have predictive and then prescriptive ytics on your process is really key
14:02
And that's one of the great things about embedding everything we do in Power BI
14:08
So we can take advantage of everything Microsoft's built into the Gartner Magic Quadrant leading BI and ytics tool
14:17
And also the Power Platform gives us the ability to meet the final two requirements in process mining
14:25
Being able to take action on any of the transactions that we see
14:30
Right. Being able to use power apps, power automate to go and tell somebody, hey, you've got a process or transaction that's not conforming to what you'd like it to look like
14:41
I'm going to take action on it or even to automatically take action inside of SAP
14:47
Maybe you have a vendor that's going to deliver late from China because of the backlog at the port in L.A
14:53
But you know you have another supplier in South America that can you know deliver on time because they go into the port of New Orleans So you can automatically take action and route your order from Chinese supplier to the South American supplier and be able to deliver on time to your customer
15:13
And then as you're building out all those automations, maybe with Power Automate
15:17
the ability to record what the users do on their desktops to create your process design documents
15:24
to build your bots is one of the next key features. So a lot of requirements around process mining, but by embedding it in Power BI in the Power
15:35
platform and being able to take the audit logs and rebuild that path now, it can meet all of
15:42
those requirements is probably is one of only two or three vendors out of the dozens that
15:48
do process mining that can do that. And that's why we embedded process mining into business
15:56
intelligence, right? Business intelligence has been around for decades. And there are a handful
16:02
I guess, three leaders now, Power BI, Tableau and Click, that lead and have built out all the
16:09
ytics that you could ever possibly need. So to be able to meet all the requirements that we just looked at
16:15
you need to have really strong ytics and all the other process mining
16:19
vendors in the world had to go out and rebuild everything that Power BI
16:25
Tableau and Click can do today just to be able to do the dashboarding and all
16:32
of that. So what we said is, you know what? Power BI is the market leader way up in the top right of the magic
16:38
quadrant what if we took advantage of everything microsoft's doing and just built process mining
16:45
into power bi and that's what we did that's what we brought to market a couple of years ago and
16:50
we've been very very successful with that and customers that love power platform and love
16:56
power bi really like to have a single ytical platform that does dashboarding and also can can
17:03
and look into their processes. So, and we take advantage of all of that, right
17:10
So inside of Power BI, we can discover your process, yze your process, visualize your process
17:16
Inside of Power Apps and Power Automate, we can share and take action and alerts
17:22
and create bots around that. And then by embedding it in Azure and Office 365
17:31
take advantage of the Azure data factory. Don't have to build our own machine learning
17:35
We can use the Microsoft Cognitive Services. So everything that Microsoft does in the Power Platform
17:43
and in the Azure data and ytics system, we can take advantage of because we've
17:49
embedded process mining inside of Power BI. So that's overview of process mining
17:57
Hopefully that helps, gives you an idea. I think this demo from Robert will make it really clear on what that looks like and what the use case is for process mining
18:10
Thanks, Daniel. I will share my screen now. Okay, looks like we got the infinite loop going
18:24
All right. So thanks, Daniel, again. I'm going to be showing our PathNow
18:32
demo here on the PathNow on Power BI. So to start out with, like Daniel said
18:40
we have an event log that every process mining journey begins with. In this case
18:46
we are using a purchase to pay event log. So we start with create purchase order
18:51
go through some approval levels, post our goods receipt, and then do some invoice paying and
18:57
posting. And there's maybe some other steps, but that's kind of the general process and what you
19:02
might see in a purchase to pay process. So a term that I want to real quick go over with process
19:08
mining is a variant. Daniel covered this a little bit, but I want to give you a little visual
19:13
One variant is one way that your process can happen. So if I click our top variant over here
19:19
One way our process can happen is create purchase order, go through three levels of approval
19:25
and then we have our post-cause receipt and post-invoice and pay-invoice. Another way that our process can happen is that we can actually skip some approval levels
19:33
and go right to post-cause receipt. So basically, this is just, I just wanted to do like a clear example of what a variant
19:40
actually is, so that if I'm talking about that, and if you're wondering about all these
19:43
different pieces of the puzzle that you're seeing right here, they're all different variants
19:48
being shown in one graphic here. So right in the middle of the page, we have the PathNow
19:55
process explorer. This is our main bread and butter, the visual that shows your process
20:01
straight to you. On PathNow, we have these reports that we offer. This is a template report where you
20:08
can load any event login, and as long as you've got your data, then we can load it right into
20:14
this report and we start with the discovery page. The discovery page is the page where we start our
20:21
process mining journey because we get to look at the overall process. We get to see all the
20:26
different ways it's happening, our most common ways, and then the different steps that may be
20:31
happening. And here you can see we have a self-loop indicator, meaning that sometimes our process goes
20:37
from approval level one right back to approval level one. And then you'll also notice that we
20:42
have these red marks here. This is showing us that sometimes we have a long average lead time between
20:49
two steps. So as you can see, when it's bright red, that's meaning it's longer. And when it's gray
20:55
it's a more regular, shorter amount of time. So this is just a quick way to identify maybe some
21:01
bottlenecks like right off the bat when you take a look at your process. Down here, we have a slider
21:07
And in this slider, we use this to display our different variants. So if I bring this out
21:11
It's going to show more variants all at once, all the different ways your process goes through, more edges
21:17
And if I go all the way to the right, it'll be your most uncommon variants
21:21
You'll get some pretty wild processes there. Most of them probably only happening one time
21:27
So, yeah, the reason that we integrated in Power BI, and Daniel covered this a little bit
21:32
is that we love the way that the Power Apps and all of those integrate and how Power BI works within itself
21:38
everything is so customizable and everything interacts with everything else that it makes
21:43
process mining that much more powerful your insights and your actions are easily linked
21:48
so let me keep going through this this report a little bit next i'm going to touch on benchmarking
21:55
page so on the benchmarking page this is just a quick page where what we try to do here is
22:02
take your process. So now you've gone on Discovery, you've looked at your process a little bit
22:08
We use this page to do a look between all of your, sorry
22:18
your processes and you can compare them based off of different attributes So here I can click the shorter cases and over here I click the longer cases So greater than 66 days and 0 to 11 days And we compare how those processes and a few KPIs change
22:37
So here we have an average lead time of 111 days and 8 hours
22:42
which is drastically different than our average lead time of 5 days and 14 hours
22:47
Now keep in mind we selected on purpose our longer lead times over here
22:51
But as you can see, we selected greater than 66 days and we have 111 days as our average, meaning that when we have these long cases, they get really long
23:01
So we can zoom in. And as I said before, right off the bat, you can see a few different places where you're having really long lead times
23:09
If you skip all these approval levels and you go from purchase order right to a goods receipt, you know, it takes on average 30 days
23:16
Maybe that's not even a conforming part of your process. Right there, you may have found a step saying
23:21
hey, who is skipping approval levels? We need to get on that because that's not even a part of our
23:25
process that we like. Over here, you can see your fastest variance. So over here, we see
23:33
okay, the longest amount of time we're getting in these is three days between post invoice and pay
23:38
invoice, which is really not much of a significant issue to us. This is a good sign that we're
23:44
we're settling our invoices quickly. What you can also see is your variation. So
23:51
we have 239 cases covered by our short lead times, and we have 17 variants. This is maybe
23:58
one every 10 cases you get a new variant, whereas over here, you have 101 cases and 34 variants
24:06
You've doubled your variants and then more than halved your amount of cases, showing that when
24:11
you have these long lead times, these are not your common variants. These are your really uncommon
24:15
ones, which can maybe lead you to a bit of an audit perspective where you're saying, hey, we need to
24:21
take control of our process and we need to make sure we're really standardizing it. Because when
24:26
we standardize it, we get a faster lead time. It's more predictable. We know what's going to happen
24:32
So we have another set of pages right in here from variants, loops, automation, lead times
24:40
and process steps. This is our ysis section. They all look somewhat similar. So I'm just going
24:45
to cover one of them. And I want to show you a key visual that I haven't shown you yet. That is
24:49
another path now part of process mining is this root cause yzer. So right here in the center
24:55
of the page, we have this root cause yzer, and this reacts to your data as everything in
25:00
Power BI does. And what it does is attempts to show you some correlation between your selection
25:06
in data and what's maybe involved in that data, you know, which resources are involved
25:12
in your long cases, which process steps. So in this case, we see that when we start with purchase requisition rather than purchase
25:19
order, we are getting really long lead times because up here, what I've selected is our
25:24
cases with a greater than 66 average, which is similar to what I just did with benchmarking
25:29
But now we've got a little bit of a drill down going through so we can see what may
25:34
be causing this thing. So yeah, that's the root cause yzer. You can throw anything into here. And again, the magic of the Power Apps and Power BI in specific, everything interacts with everything so seamlessly that when you throw your resource, your plant, your different managers and materials that are involved in these, you can see how all of those correlate to your durations, your things like that, your high cost invoices
26:03
now it doesn't always mean that those things cause those things but it can be it can show a
26:09
relation at least so for example if i saw that okay when we ship from taiwan it takes this amount
26:15
of days well it's like well let me think about it i live in the u.s and the taiwan's farther away
26:20
than mexico is it makes sense that it might take longer so is there something else here that's going
26:24
on um but yeah so that's what we do here with the root cause yzer um we like how it works with
26:31
all the data and it's really good to use for that. So I'm going to skip over to compliance ysis
26:44
So in compliance ysis we have two new path now visuals. We have this event filter here and
26:50
then we have our case viewer. So originally I showed you the process explorer which shows
26:57
you your variance, it shows how they all how the process can happen, how all the different steps are
27:02
and it shows them side by side. Here we're just seeing a case view, we're seeing one case by one
27:09
case and seeing exactly what those cases did laid out exactly as we as the steps went. So here we
27:15
see instead of some kind of indicator that we have a self loop, we see post good receipt and then we
27:20
see it again. And this up here, the event filter, I'll blow it up real quick. This is a really great
27:27
part of our visual sweep. So what you can do here is actually build your process and see how often
27:33
this process happens or the different attributes involved in it. So let's say I want to find where
27:39
we don't have approval PO level one. I can do that. I filtered my page and now we see only 141 of my
27:46
cases don't have the approval level one and 15 variants don't have it. So now and then I can dive
27:53
into each one of those steps, and now I can see all these red giving us a self-loop indicator
27:59
saying, okay, when we don't have these things, we're seeing a lot of self-loops
28:08
So the way this works up here is you define an initial step, so we have create purchase order
28:13
and then you have a following step. If I do this, create purchase order, and I click and make this
28:19
line solid. It's saying, I want to see only the cases where create purchase orders not directly
28:24
followed by approved PO level one. So now we're seeing, okay, we still see approved PO level one
28:29
we just don't see where it's directly followed. So this is really powerful because it allows you
28:35
to build maybe the process you want to see and see how often that process is happening. Maybe it
28:41
lets you build a process you don't want to see and you can see how long that's happening and
28:45
everything in between. You get to see all those little steps and all those little variants as you
28:51
define them. Finally, on the conformance check page, I'll show a few more things that you can do with
28:58
the Process Explorer. And then I'm going to do a small little demonstration of how PathNow interacts
29:05
with other PowerApps. So what I've done here is I've clicked this conformance checking button here
29:10
I'll click off real quick. This button right up here is your conformance checking button
29:15
And what this does is it sets all of your process to be read like this
29:21
So what this is telling you is that none of your process is marked as conforming right now
29:25
So what I can do is use this thumbs up and say, okay, now everything that I have on is conforming
29:31
If I bring this out, okay, now we see the steps that don't conform to what I just marked as conforming
29:36
So this conformance check is really nice to actively build a 2B model
29:42
So if I have all these steps that I'm like, okay, yeah, I want this step involved in my process, I can individually, I can also go on here and right-click and mark that as conforming or as not conforming
29:54
This allows us to do almost what Visio does, where we can kind of build that process
29:59
and then see how the rest of our process pertains to that, what we've built
30:04
So when I really zoom out here, we can just see all these different process steps that aren't conforming
30:11
We can see all the different ways that our process skips out on what we want to do
30:15
And in this case, you can also see that self loops are also not what we want
30:20
So one other piece of the process explorer that I didn't show yet is that say you're just not interested in something
30:28
Let's say I go down to these low common variants that only happen once
30:35
I can click on one of those and I can say, okay, you know what
30:38
I don't really care about when we have pay invoice to post this receipt
30:44
It just means we had a little hiccup in our process. It doesn't usually, it's only a day longer
30:49
So let's instead mark that as conforming. And we can just let that go and we don't have to worry as much about it
30:56
On the flip side, we also have something like this, where create purchase order goods received, maybe I'm not interested in these cases, I can select exclude
31:10
And now I won't see any of those cases that have that edge. All of these different things are geared towards making it so that you can drill down into your data as far as you want to go
31:23
We want to make it as easy as possible to go farther and farther into your data to see what's causing what, to see how your process is actually accomplished, and how you want it to be accomplished
31:35
Finally, every time you apply something on the Explorer specifically, we have this thing called Breadcrumb Filters, where basically we have a list of the filters
31:47
So I'll add one more just to show. I'll exclude this self loop on ApprovePLA Level 1
31:51
and now it's just showing me all the different filters that I've done. So if I want, I can get rid of one that I had way before
31:57
because maybe I want to show those cases again or I can just get rid of all of them at once if I click this trash icon
32:05
Again, this is just geared toward ease of use and making sure that everyone can use this in the simplest way possible
32:13
With the conformance checking we also have a thing called your 2B model So what this does is shows you all of your steps that are marked as conformed On the flip side I can click this and see only our cases that are conforming
32:35
There we go. But I can also do this where I click this red one, and I see only the cases
32:41
that aren't conforming. So if you wanted to find that 2b model and then say, okay
32:45
I know my 2B model, anything that conforms to it, I don't need to see
32:49
You can just click this right here and see only the ones that don't follow your process that you've just defined
32:55
Okay, so I've gone through a lot here. This is kind of a little crash course on the path now in our visuals and how process mining is accomplished in there
33:05
But I also want to show just a small little bit about how you can integrate this with the other power apps and visuals and things like that
33:15
So I'm going to go here. I'm not here quite yet. So I'm going to go to this
33:20
This is an RPA edition that we have and we offer. On this edition, the goal is to find places in your process where you can automate and use that information to then eventually make actions
33:37
So in this case, I'm going to just skip over what I would normally do in a presentation of RPA
33:41
But I'll tell you through the demonstration, we discover that, OK, we have sometimes where we'll post our invoice and then we'll have let me click on this one
33:52
We'll set a payment block, meaning that, OK, we don't have the money right now or we don't want to pay it right now
33:58
Let's set this payment block so that someone knows they need to either go talk to someone else or they need to clear this in some way
34:06
So what we find there is that set payment block is often automated
34:11
What we don't find is that remove payment block and pay invoice are automated
34:15
Let's just use the example that sometimes you set this payment block because you don't have enough money
34:23
So you're setting this to make sure you have enough. What happens when you have enough money and you've maybe set an authorization for this to go through
34:31
What you can do is automate that. In this case, I'm going to show you like a half manual, half automated step
34:37
but what I've done is I've found this case where we have this remove payment block and pay invoice What we can do is this is a Power Apps visual or Power Automate visuals one of the two And what you do over here is you give it some predefined app that
34:54
you've created in all the different customization things that the Power Apps have. And you put it in
35:00
here and we've defined a flow behind the scenes. So that's why I have this part up. It's just a real
35:07
quick example. This isn't even how it actually happens, but when you have a data-driven alert
35:11
in Power BI, you can send an email notification. You can send a Teams notification. You can even
35:17
change things in your common data service. In ServiceNow, you can do your tickets, and I'm sure
35:26
everyone is familiar with all these different things to some extent, but we just wanted to show
35:31
how that integrates into Power BI and how it works with process mining. So I select this PO ID
35:37
I found case 813, and then I'll say, send my own email, and then I'll run that bot
35:46
So this is where the RPA starts to come in. This is how we see what we can do insofar as creating action from all this insight with process ytics and with the process mining
35:58
You can do this, and it will then send me a Teams email, or sorry, a Teams notification and an email
36:04
It takes a second and it'll say, hey, you need approval here
36:08
All right. Now it's happened. And this is what's great about it is, you know, maybe you are the person who can give approval
36:15
You can send it right to yourself and I can give approval. I'll select that
36:19
And what it's going to do is we've defined before. OK, we want to automate this removal and this pay invoice
36:27
So I approved that. And now what it's going to do is get rid of case 813 and say, OK, we automated that. We're all good. We don't need to go forward. It might take just a second, but basically it will do that
36:43
And then there's also notifications on the back end. So what's important about this little piece that I showed here is that process mining and Power BI, as I'm sure you already know, is so great because it integrates with those Power Apps
37:00
There's so many flows and things that you can integrate with all this. Process mining doesn just have to stop with Power BI You can keep it going throughout all of your stack and make those changes in an automated basis You can send your data alerts through an automated basis
37:16
You can also make things like this, this visual right here, just to send data-driven alerts to
37:22
your co-workers. So if I find some case and I'm like, whoa, this is really alarming, you could
37:27
could always send that in a Power Apps visual. So yeah, that's how that goes
37:34
That is the demonstration on all this. I'm going to end it here, and I'm going
37:40
to pass it back to Daniel, where he can show you the free visual. Perfect
37:54
Thanks, Robert. Hopefully, there we go. So that was a demo
38:03
That's how you use process mining in Power BI. That's how you take advantage of Power Apps to automate steps when you find variants in your process
38:15
So we make the Discovery Visual free. You can download it. I've got a link here on the screen
38:23
Bitly equals... on process mining, what are some metrics you should look at, and then later this week
39:04
hopefully I'll get it written tonight, we're finished up tonight, what steps you should go
39:09
through as you look at a proof of concept for process mining. So with that, that's our
39:18
presentation for today. Thanks for joining. Appreciate your time
#Business & Productivity Software
#Business Operations
#Computers & Electronics
#Consulting