Prabhjot Singh: 0:04

But now organizations are really saying okay, well in addition to addressing the technical debt, I should also understand the process. Right? What what are the workflows that are really created model next, or causing my people to be unproductive?

Tom Raftery: 0:25

Good morning, good afternoon or good evening wherever you are in the world. This is the digital supply chain podcast, the number one podcast focusing on the digitization of supply chain. And I'm your host, global vice president of SAP. Tom Raftery. Hi, everyone, welcome to the digital supply chain podcast. My name is Tom Raftery, with SAP and with me on the podcast today I have my special guest Prabhjot. Prabhjot, would you like to introduce yourself?

Prabhjot Singh: 0:54

Absolutely. Thanks for having me. Okay. My name is Prabhjot Singh I'm the CEO of Pyze. We help organizations achieve operational excellence by continually optimizing business operations and critical processes like supply chain and logistics.

Tom Raftery: 1:13

Okay, and can you give me a little bit about your background?

Prabhjot Singh: 1:16

Yeah, absolutely. So I've been an enterprise software for most of my life. 20 plus years, primarily in startups, helping scale operations. And then I started my career at a tiny company that just raised a Series A called wily technology that invented application performance monitoring. And you know, we provided in the early days of the Internet, when you try to book a car or sign up for a credit card and site would go Capote, we essentially help people figure out why. Alright, so we worked with WebSphere, WebLogic, NetWeaver, right. In those days, you know, join Wildling as your employee number 12. To help scale that business, eventually, it was acquired by computer Associates, and then I help scale it, you know, the big company. Alright, so I kind of had that experience as well. And pause in many ways is the business equivalent of what Wiley did, right? Instead of looking at Java servlets and EJ bees to densify. You know, where the memory leak is. We look at business processes, to identify where the waste is, right, where you can improve by, you know, hotspots, remove bottlenecks, improve your workflow orchestration. That's really what we

Tom Raftery: 2:43

did. Okay, so are you more a consultancy than a software solution?

Prabhjot Singh: 2:50

That we are an AI platform? So in in olden days, if someone wanted to do a process improvement exercise, you'd hire a consultancy, right to do it. Today, right, the technology is at a point where given that there's a digital footprint for almost every action, right? Whether it's sort of logistics, manufacturing, supply chain operations, right, we understand where handoffs happen, what the different touch points are for a specific process. So pause is a AI enabled platform that ingest data from multiple different data streams. So think about your process activity, think about user interactions with an application to execute specific tasks. And then we marry that with other data sets like employee system of records, your Active Directory LDAP, people will bring in HR system data, pay down information, to sort of understand okay picks your top 30 minutes to do a particular task, what does it cost the company, even CRM data from customers, and then we put that in a Vitamix and blend that together and we produce beautiful visual diagrams of the end to end business process. You can touch and feel interact with that process, you can hit play and see how transactions flow right through one workflow versus another and then drill down into specific bottlenecks and identify root cause wire pump potential exists.

Tom Raftery: 4:33

Okay, very good. And you, you talked about continuous process improvement for people who, you know, might be unaware or might not have a deep familiarity. Can you talk to us about what that is what that means for you and for organizations that you deal with?

Prabhjot Singh: 4:49

Absolutely. So, you know, in the in the kind of old context of where you're hiring, you're hiring a consulting company to come in and help with the process. We're actually saves me a lot of companies sort of do this once a year, once every six months, once every couple of years, right as part of their strategic planning or overall business review. It's a very expensive, and yeah, manual process, right? Because you've got, typically you do qualitative interviews, you look at a few transactions, and a spot check transactions, right? we've sort of seen the metaphor of someone with a stopwatch standing at the factory floor to see how long it takes, right. But certainly things are more advanced in that way. But it's not that much, far off, right? The problem about approaches, it's looking backwards, right. So you have a snapshot at a point in time. And given how fast things move today, the output becomes stale almost as soon as it's that that project is completed, right? Because you've got so many disruptions in supply chain happening today, right? There's dependencies on downstream suppliers for distribution and transport. So if you want to achieve logistical precision of operations, you really need to have a continuous understanding of what's happening within your supply chain. So what we enable is really hooking up data pipelines into our platform, where we're looking at every single transaction that's flowing through the the system, right, so if you were looking at an order management process, for instance, right, we're looking at 100% of all orders that flow through that system, we understand what the dependency is of those orders on, on upstream suppliers, we understand what the dependency as you know, whether we're drop shipping, or, you know, we've got sort of a trivial process to to actually do delivery, we've got visibility into all of those different components. So as you start to see a slowdown, let's say, in, in the downstream side of things, we can flag that right, we can flag that, hey, here's a problem, before it actually becomes a problem. And that's the continuous improvement piece of it is this the analysis is not static, it's ongoing, it's, it's complete, because you're looking at all transactions. And we have the ability to identify anomalies that are occurring in near real time. So that you can take action to actually prevent them from impacting the business down the line.

Tom Raftery: 7:40

What kind of anomalies typically, would you be seeing or even on typically, what, what kind of examples? Can you talk about?

Prabhjot Singh: 7:48

Yeah, that's a great question. And I love talking about sort of, you know, the engagement models that we've seen. So, you know, for, you know, we're discussing order management, so let's sort of take, you know, order or management, for instance, for any specific process, you know, we, we typically focus on what the business goals are, right. So if we're, you're dealing with a manufacturing company, you typically will want to look at things like speed of fulfillment, right from when the order gets placed to when when it gets delivered, we want to look at the cost of, of actually processing that order, right, so the human cost of processing that order. And then, you know, manufacturing, quality is important, right, we want to understand sort of quality measurements at each step of that process. So that if there's issues in quality, we catch them early on versus waiting, sort of, you know, the quality assurance phase, at the very end. Define that something's bad. And now, you know, we've wasted a lot of resources, energy, and in putting that part, through a process that we can actually avoid it right and terminated earlier. And in terms of sort of what we typically see, right, so in every organization, there's a sense of how work gets done. Right? When you are, when you're talking to managers, you know, sort of the senior level or even management, there's a sense of how things gets done. And typically, it's almost incomplete. Right? People have an understanding or vantage point from their view of the world. You know, how the supply chain works, and, you know, their view of the world might be accurate from their, you know, their, their perspective. But if you zoom out a little bit, you know, everyone's got some understanding of kind of how the process works. And that end to end process, no one really understands or there'll be very few people that have context little piece that together. So, you know, people often have Visio diagrams that are antiquated and not reflective of the real world. And, you know, what's surprising is how often we see workflows, where you have sort of a ping pong effect going on, where things are going from, you know, one department to another department coming back to the department, that's something very typical that we see. Right, and the way the PI's platform works is we'll analyze a process. And you know, my my friend, whoever you bet, Tech Mahindra Inc, he says, you know, you guys do an MRI of the process, right? We essentially give you that MRI, and then, you know, okay, well, which can be to operate on for, for what purpose, so to speak, right. And in doing that MRI, we actually generate all the different workflow variants that exists for that process execution, right. So let's say there's 200, different ways that an order can be processed, and, you know, might be 2000. So, for each of those variants, we understand how many steps are involved, what the cost is, of each barrier to the business, like, how much time is spent, in terms of execution? How much time on passed, right? Like, how long do people actually spend doing specific tasks to push that process from one step to another? Right, so we do all that analysis? And then we'll, we'll identify, Okay, which workflows are efficient, which works for inefficient? You know, or you can approach it from the transactional angle and say, Okay, well show me the top 10 percentile of long running production run, which orders took the longest. And, and now let's understand the workloads that enable those orders, right. And, and then you've got kind of a roadmap of okay, well, oh, there's a bottleneck that we can address. And then, you know, that bottleneck might exist, because the workflow itself is architecturally incorrect. It might exist because we have people doing manual tasks that maybe we can implement RPA or some other automation to address. Or it might exist, because, you know, we we hired 50 new people last month, and the system is really hard to use. And the senior people figured it out. And you people struggling with the timing issue, right, Psalter unless you understand kind of the current state of things and the what the issue is, it's very difficult to take action, right? Like you could operate on the wrong kidney. Right? If you don't have that MRI, and then once you make changes for because what we're doing is enabling people really to make the right decisions. So as you make changes, now, you've got a visibility into what that change enabled, right? Did we make things better? Do we make things worse? So that's, that's the typical engagement that we have. But the the idea is, you have to enable continuous improvement by being able to look at, okay, how are we doing week over week, day over day, month or month? And having kind of visibility really at a tiered cascaded fashion? Right, so at each site at each plant, at each region, you can sort of provide visibility at a level of granularity that makes sense for for the people that are looking at it. And then you'll be able to drill all the way down to a single transaction. But to understand, well, okay, this this, a customer had an issue. And, you know, we're trying to sort of remediate that issue, but let's understand why that issue exactly,

Tom Raftery: 14:14

occurred. Okay. Are there any outcomes? You can speak to?

Prabhjot Singh: 14:51

Yeah, absolutely. So we oftentimes get engaged in the context of digital transformation, where companies are looking to either move away from pen and paper, right or Excel sheets to, to, you know, modernize their, their systems in the cloud, right? And we've seen a lot of that obviously, in the last couple of years. With COVID or, you know, there's sort of a just a desire to move off of on prem to the cloud and people instead of just taking their current processes, right, and shifting them to the cloud. Want to Look at how can we improve this business process? Right. And this is relatively new in the industry that people are, are starting to focus on the business process. You know, typically, when we, when people talk about digital transformation, or cloud migration, it was sort of a lift and shift, right? I've got, I'm addressing technical debt, and I'm taking, and I'm doing it because you know, my systems or go are going end of life or are going to be supported, or it's just too hard to maintain. And I'm going to move them to, to make it easier. But now organizations are really saying, Okay, well, in addition to addressing the technical debt, I should also understand the process that I have, right? What are the workflows that are really created model next, or, you know, causing my people to be unproductive? And in that context, you know, we we get engaged, when people have really, really aggressive goals, right, I want to decrease the time it takes to do a master data set up and SAP by 50%, right, because all of that is being done using, you know, Excel, bingo, right? Where you've got kind of, you know, 10 different people that are updating your Excel sheet. And then it's been bounced around via email, and then eventually, you know, it takes longer to actually get the order entered them does to actually produce the goods, right. So that that's kind of a, you know, pretty, pretty typical use case. You know, we want to increase the speed of order delivery, or we want to reduce the amount of time it takes to process specific transactions. I'll give you example, right? The fortune 500 Bank, that looking at sort of fraud transactions, and they they process something like 2 million fraud convictions a month, right? And now they can't go and say, Hey, son, this last fall productions, right? The those transactions got triggered whenever someone takes up whatever $2,000 from their from ATM or right? issue the check for $10,000, whatever. That's right. There's, there's business rules that trigger that and are coming from various different lines of businesses. And there's a flow involved in terms of you know, how those transactions have to be processed? And has that volume increases, you also can go and say, Hey, instead of 300 people, and you're passing people, it's just not possible. So the question is, how do you deal with that? Right. And this applies to a lot of industries, like we can talk about similar examples in supply chain manufacturing. But then what you have to do is address that type of issue at two levels. One is at the workflow level, right? So understanding what the work process workflows are that exists, and which workflows are inefficient and eliminate those workflows, right. So for instance, you might have transactions that come from a branch, go to the back office for processing, and then go back to the branch, right? And then you have that ping pong effect happening where, whereas you know, maybe the branch can do something about it, maybe they can't, right, it's just sort of like so you can actually eliminate, right, by putting in business rules, just like, you know, backside rules, let's say, if a transaction is under $25, even if it looks like it's fraud, we might not worry about it, right? Because, you know, or a credit card company says if a dispute is under X amount, we're just not going to worry about it because it costs more to then do anything about it. So that you can kind of tweak business rules, you can you can eliminate certain workflows to decrease the overall time being spent on transactions. And then what you can do is also address what happens when you're actually processing the transaction right? When an operator or an agent gets a foreign transaction? What are the steps that they have to go through? Right? Do they have to go spelunking for data in three different systems? Right? Access SharePoint and, you know, whatever else you can, you know, so you can look at okay, well, how do we consolidate systems, how you consolidate data, you know, how do you add an automation to either do business process orchestration better right for that? flightcase or, you know, implement RPA to, to, to really automate some of those repetitive manual tasks that people are doing. And, and, you know, by doing that you can, you know, get upwards of sometimes 60 70% efficiency gains, right, in terms of how long it takes to process a particular technology, right, that, that people are doing in a haphazard way today, potentially, and making it kind of more streamlined, right. So if you operate a boat those levels, right, so think of it as the Eco workflow level, and then the next task execution level. And what we've done at pi is is Bill, this Miss platform, which really does integrated process, and task mining, right, which we call process intelligence, which is underpinned by this AI layer that detects anomalies and identifies issues.

Tom Raftery: 21:30

Interesting, interesting. And where to next? I mean, what's what's next on your kind of roadmap of things to roll out?

Prabhjot Singh: 21:38

Yeah, that's a that's a? That's a good question. I mean, we yeah, we get, we get asked a lot to sort of, hey, can you actually do some of the automation, you know, suggestions that your platform is, is referring to, or suggests, then, I mean, we've got our hands full, but what we're doing because we do it, we do it while we want to do it better. So we're, you know, it's a pretty big ecosystem, and I think you can't be all things to all people. So today, I think the investments that we're making is more on the AI side, to help with root cause. Analysis. Right. And, and, and recommendations, not not just that you have an issue, but being able to automatically sort of look at the segmentations that provide visibility into what the root causes, right, we were less than better anomaly detection, and better outlier analysis, right? Help people really get things to the mean, right? Because there's a lot of organizations that we work with, where, you know, the distribution of standard work is two or three standard deviations. So if you can, and sometimes more, right. So, you know, as you push things more towards the mean, you can achieve that logistical precision, right? Or, or, or at least have a path to get there. So we're, you know, we work we work with, you know, large, large, large enterprises, like US Air Force to the post office, right, in terms of helping them analyze data, right, kind of focus on things that really make sense. And AI is a great enabler. Right? Because like, we don't expect people to look at look at the dashboard every day, but a lot of people do, right? This is like there's, there's many sort of frontline delivery managers, right. product owners have applications that are that utilize pies, and, you know, we eat our dog food, right, we see that people listen, the first thing they're loading up at eight 9am In the morning, to understand how things went yesterday, and they're going through the process map and looking at, okay, how transactions are flowing and where, where there were issues yesterday, and you know, even senior managers and executives are looking at this data, right to see where they have bottlenecks in their in their organization. Because we presented visually, and it's, you know, it's easy to kind of digest, so to speak.

Tom Raftery: 24:27

Sure, sure. Sure. Sure. Sure. We're coming towards the end of the podcast, no project. Is there any question I haven't asked that you wish I had, or any aspect of this that we've not touched on that you think it's important for people to be aware of?

Prabhjot Singh: 24:42

Well, I I love talking about the process. I can, you're off how that that the one area that I think would be interesting to to just touch on would be the low code angle. Okay. Right. So low code, of course, is all the rage today and overseeing low code applications really starting to eat up your high code application development, right? And especially in the context of digital transformation and application modernization, right. So as people go from legacy, right old school mainframe systems, to whether it's the cloud or internal cloud or hybrid or whatever it is, they're really taking hold of low code platforms, right. So everything, you know, Mendax, and our systems and pega, right, that are being used, not just for application development, but also process orchestration. And one of the things that we've done at PI's, over the last few years is built really, really deep integrations with these platforms, that enable us to be process aware, without any work. So you can think of it almost as like low code processing tasks. Right. So like, if you plug pies into a pega application, we understand cases and process definitions and stages, and steps, right. So if you're looking at kind of your logistics supply chain manufacturing process, or any process of using using Tiger, we, we actually automatically understand just by dropping our, our SDK, and that's really, really helpful from a process perspective. Because the days of, you know, exporting database tables, and, you know, munging, log files and doing ETL and then pulling that into a process mining tool. Our are behind us, right? So we typically do that, right? And of course, we have to do that for for legacy systems. But as people migrate to these local platforms, you sort of get this automated process intelligence capability, right? And now we can also do the before and after that approach, right? So this is where you are. You identify the issues. And then as you're migrating to the local environment, you can actually read the wrestling operations.

Tom Raftery: 27:23

Nice. Nice, nice, nice. Okay, Prabhjot. If people want to know more about yourself or about Pyze, or about continuous process improvement, or any of the things we talked about in the podcast today. Where would you have me direct them?

Prabhjot Singh: 27:40

Yeah, our website is a great place to start right? pyze.com you can certainly look me up on Twitter@psinghSF. And like, then I would love to chat with your listeners and see if there's an opportunity for us to, you know, help them with process improvement and their business.

Tom Raftery: 28:02

Tremendous, great, fantastic. I'll put those links in the show notes as well, so people have easy access to them properly. That's been really cool. Thanks a million for coming on the podcast today.

Prabhjot Singh: 28:12

Yeah, my pleasure. Thanks for having me.

Tom Raftery: 28:16

Okay, we've come to the end of the show. Thanks, everyone for listening. If you'd like to know more about digital supply chains, head on over to sa p.com/digital supply chain or, or simply drop me an email to Tom Raftery at sa p.com. If you liked the show, please don't forget to subscribe to it and your podcast application of choice to get new episodes as soon as they're published. Also, please don't forget to rate and review the podcast. It really does help new people to find the show. Thanks. catch you all next time.