Jack Maher, Pavel Azaletskiy: AI, Digital and Value Streams Enable Data Driven Organizations

This talk discusses the integration of AI and digital twin technologies into value stream management to enhance data-driven organizations. The speakers emphasize the importance of value streams as essential frameworks for measuring performance and adapting organizational resources to generate higher business value. They explain how modern tools and practices, such as machine learning and generative AI, can optimize operations while reducing low-value tasks through automation. Furthermore, they delve into the concept of digital twins, which are digital representations of physical assets and processes that allow for predictive maintenance and enhanced decision-making. Overall, the session sheds light on how these emerging technologies can transform value stream methodologies, enabling organizations to make informed adjustments and achieve their operational goals.

[Music] welcome everybody to this session on AI digital and value streams enable datadriven organizations and with us today there are two speakers for this session Pavel aesi and also jackar and uh they are both they have been uh doing working together for years to assist Enterprises and organizations in delivering business value to clients as Global 500 technology consultants and business process advisors so Jack is passionate about

both value and values as business drivers and Pavel leads his Cutting Edge organization vs Optima providing digital twin capabilities that enable complex business process modeling and simulation so with that we welcome pel and Jack to the stage and uh the stage is all yours great thank you very much thank you for joining us today pav and I are delighted to be here with you and glad you

decided to join us yeah we really appreciate the opportunity to present yeah looking forward so bavo and I are here to talk about a number of things we start off with value streams one of the things that pav and I found that we shared very early on when we uh started working together was that we were very focused on value and while that is a very commonly

used term today sometimes we may lose track of of what that is I like to think of value streams as the basis and value streams are a lot like budgets or goals if they're not written down and you're not measuring your performance against them they're not of very much use and they're all about taking the resources that we have available to us from our organizations and taking

that and and making something more of it of increasing the value of what we get by putting it together now if we WI went home tonight um and and many of us may have started to plant our Gardens for the spring and summer if you went home and turned on the hose to water your garden and money started to flow out of that garden hose would you

rush to the spigot and turn it off absolutely not right you'd be getting a bucket hopefully a great big bucket you would want to do everything that you could to make that money flow more smoothly more quickly and you get more of it um as opposed to losing or wasting it would we want to increase that all of that absolutely of course we would so let's think

about that because that's exactly what's happening in our organizations we are taking these resources and we create money with that now we also create other kinds of value and today's value stream Maps look quite a bit different than the ones that you might have seen in the past um in particular there are many tools that we can and should be using for Value stream maps in some

cases they may look and perhaps should look like the typical vream maps that we've seen in the past that help us Trace what activities are taking place in what areas what kind of value is being created what kind of information are we sharing with those folks that are involved in the process even more so there are many specific tools today that are geared towards specific value streams

or areas in which we have significant interest or investment and these tools help us make sure that we're really doing the things that we want our organization to be doing and we're reaching ing the goals uh and aspirations of our organization again much like budgets and other things that we do to manage control or otherwise understand our activities and results we need to use data right very

few people would argue today that there's no need to measure what we do and there's no way to measure the things that we care about some things are just the same when we think about things like the Dora uh measurements those are obviously reflected in here and pretty simple to get in today's world but there are many other things that are important to us and we should

think about them in different perspectives so we typically talk a lot about our objectives and key u results and we also talk even more about our key performance indicators those are things that are typically pretty easy to count and don't have a lot of value in and of themselves but either drive or represent other kinds of value now as we mentioned there are many different tools and

we should be looking at various tools to look at what can we learn from the marketplace what do we see is actually happening in the way we engage our uh customers our clients and and our constituents you know how are we ensuring that what we believe is happening what we want to happen what we designed our systems to make happen how do we know if it really

is happening and not just from our perspective but more importantly what does it look like from our constituencies perspective right are they happy do they have what they need to be successful are we putting them in the right place and are we checking the right things to make sure that we're all accomplishing those things that our organizations need from us now there's a lot of different ways

to do value stream mapping and if you search online you will find hundreds if not thousands of ideas and approaches one thing that many approaches today agree upon is that it takes some fairly significant work most folks believe it takes at least a couple of days to do a decent value stream map I suggest that you use whatever resources you have available to you including tools but

you should also consider other options and opportunities and Shameless plug Valu streampro docomo available to you the second part of this bringing things a little bit more into what we might think of from a contemporary perspective is the rise of artificial intelligence obviously folks have been paying attention in this space and you're probably one of them have seen Artificial Intelligence coming for quite some time and we

already use it in many different ways machine learning is probably the most pervasive right now it's been around forever it really grew out of what we called Big Data and it pretty mature in its uh capabilities at this point generative AI is probably the newest and the thing that folks are talking about the most there is significant opportunities for us to leverage both machine learning and generative

AI in our domains one that's specific to all of us is the operations of our organizations artificial intelligence for it operations is probably among the most mature and the one that may be the most available to us and and something that we should really focus on now again when we think about and talk about AI machine learning is the one that we see most commonly today and

generative AI is on its way we're going to look at several different tools today that include both of these tools as part of their skill set including tools like from BS Optima what do we mean by that again we start off with things that we already have Investments we've already made tools systems and processes that we already have in place coming from the Big Data space so

we don't have to really worry too much about where we're going to find that or if we have the capabilities those are likely present to some degree in our organization already the other thing that is super important for us and something that has probably not in my opinion gotten enough attention yet is how we can leverage our tools data and and processes to reduce the relatively low

value work that we frequently have people doing those things that we could automate and liberate those folks to doing higher value work to increasing our capabilities as an organization without having to add more people because we're smarter now machine learning has been a big part of this and it's increasingly being available to us through tools especially in the in the devops space the uh Headway that we've

made with building out software uh pipelines has really uh put us in in a great position to manage this information and share it across our organizations and as a result we can actually move into actual value stream management now not just looking not just watching not just counting but at actually being able to adjust um and make changes that will improve um our creation and delivery of

value and last but not least is the rise of resilience engineering or maybe chaos engineering as you've heard of it in the past things like the tools that come to us from the folks at Netflix in the Simeon Army that and that approach as we design our processes will enable us to better leverage the things that we already have um in our move forward last but not

least is is getting to observability we don't really have very many observability not true observability Solutions so those of you that are in a vendor uh role I would suggest that you put this on your list because observability means I can see what I need to know without having to know that I need to know know it right it needs to jump out at us and we

don't have that yet today but we're getting there last but not least is digital Twins and I'm gon to shut up for a moment and let Pavo take over from here yeah I believe this digital thank you Jack uh for the great overview and digital T you can think of as a as a natural continuation of this journey of advancing technology that helps us to to Really

Embrace efficiency so what Digital P is about uh digital twin is the digital representation of physical or U physical assets or processes or systems that comprise of assets and and processes initially it started back in I don't know 70s if not 60s when people started building like complicated systems like rockets and they need to test them out before uh they make certain changes so then start building

certain models mathematical models uh how it works like first mission Napo create a physical digital twin if you will like there's a uh reference or data tracing from actual rocket all the way to the physical representation uh and they can test what is happening they can see what is happening with the rocket and can verify various various hypothesis moving forward starting from 2000s uh people realized that

they can leverage digital twin in um civil um civil settings for instance jet engines uh like coffee machine if you will so when you have ability to see what is happening with the uh device based on the telemeter data and you can predict what will happen in future given the huge amount of historical data you can make better decisions you can uh have a predictive maintenance you

can reduce the cost of serving like it's much more cheaper to fix before you break something than fix after you break so preventing maintenance became a big uh use case for twin and um moving forward we can think about digital TW of the processes of the streams where they can bridge the gap between the just looking at the metrics or looking on snapshot of the uh what

is happening of the value stream structure to what is actually causing the issues and efficiencies or what do you have to change to make certain desired outcome so in this case you need to have a causality in your model this is where digital twin can help just to give you an example let's jump to the next slide what uh yeah you might think like oh digital TN

is something like we never seen before but no it exists you can think about uh training for the pilot of the pilots you have a digital environment where you like really train the pilots like in different situations it's much more cheaper than break the plane and uh yeah injure the pilot so it's a safe environment you can do whatever risk decisions and verify the outcome like going

back to to Domino um I know it was introduced Jack correct me if I remember but maybe like at the beginning of 2000s right so you already can see this is actually exact example of value stream you place the order you see what is happening you understand when it will be delivered it's like observability not not observ it's transparency of what is happening with the order and

predictability about what's going to be next at the same time at the back end for the Domino this is very important they can streamline the value stream they can reduce the waste as much as they can and be efficient and dominant players in the market meta go ahead and advance that to their like uh real people you can build Ulus uh for the uh for the game

settings and for the Enterprise settings if you have a manufacturing or facility or you can plug it in your Oculus and look what is happening you can learn how to behave in the different situation so it's already here and I believe what we what we want to do is to embrace and start using it at um other settings like the W settings how we can Ure that

we have productive engineering environment that helps us to deliver more value rather than I know fix buxs and infrastructure where we need to allocate our efforts to streamline the vops pipeline as a whole not just local optimization doesn't move the needle uh of the system performance you know pav it strikes me as well that this this is a great example of a distinction between our processes and

value streams because our processes are usually from our own perspective but each of these is from that user or consumer's perspective absolutely yeah uh there is another very important thing that uh digital twin can help and Advance the classical value stream mapping methodology right because value stream mapping give you great opportunity to learn what is happening and the whole team can GA it together and understand okay

this is our steps that and how we produce the value those are kind of limitations or issues that we observe because of lead time delays waste Etc so we can understand that but this is Snapshot in the picture at the time the next question is like how it going to behave in the different circumstan what if we're going to make certain changes how the dynamic will be

um updated or what if Dem man going to grow what implications we will have to our value stream for instance if we need to deliver not I know 10 features per release but 20 features what if we're going to like scale our engineering team like what would be the throughput how many people and which role do we need to have so this understanding is really important and

that can bring for Value streams uh sorry from from digital TN of value streams when you run simulation and you can see the dynamic and you can understand okay those are decisions I need to make to handle this situation this is environment and um this is super powerful uh aspect of digital twin so we can understand Dynamic you can understand variations of the let's say demand fluctuation

your performance of the team you can look at the how it how it entails on the overall productivity overall throughput and so on you can look at lead time structure and how it deviate from the what you see at the moment uh and what you anticipated to actually what it going to be in in in the after the change is implemented that whatever change you want to

improve in in your Implement to improve your value stream you can look at the seasonality maybe it's not very relevant for the um uh devops World however maybe if you work in retail T might really uh affect you because like as far as I know like there is no releases whatsoever at the last quarter of the year because the last quarter of the Year retail really should

like work robustly and earn money and focus and nothing no changes in production there is a very high risk of out ages that will affect the revenue so this is anality in software in some Industries as well but how it affect the value stream so you can and how affect Dynamic of the value stream you can look at that through the digital twin settings isn't that key

that there's so much Dynamic activity today as opposed to when we were able to do value stream maps with Post-it notes and PowerPoint things were a bit more static I think yeah yeah the speed of change is growing for sure yeah when to use digital Twins and um yeah I believe there are three main um use cases if you will if you really need to understand the

dynamic or you want to improve the value stream and you want to understand uh what the implications would be if you want to understand okay whether do you allocate your efforts to the right reason or to the right issue to really make a system effect of on the performance right because this is it might be the case that you can Implement I don't know let's say trun

based development but it doesn't move the needle right maybe what is really holding you behind is infrastructure because it's not reliable or test automation this is where you need to priorize your efforts so it helps you to make those decisions or you want to kind of calibrate or find Optimal configuration of your uh value stream value stream consists of not just activities but you also have a

capacity in each step people or systems performing the various kind of uh activities as a part of value stream you need to balance them if you have certain pressure at the demand on the value stream you need to make sure that you don't have a bottleneck on the capacity lbel you might have a bottleneck on the other layers uh that's in in time work use but make

sure that you have enough resources available to handle the amount that you plant or experience this is where you can help uh you can get helped with digital TN to find the optimal model um at the moment digital twin uh you can build you as an Enterprise your own capability there are certain platforms but in general I would love to show you a high level architectural diagram

what is comprised of the first part on the right side you have a real world and it's is very complicated but we need to get from The Real World observation building value stream uh map uh looking at the data uh Levering AI find the feature really important for the performance of the teams so we can create we can U elicit those features and start getting the data

so the next part is integration of course so we need to have I either it's iot if you thinking about physical processes in digital I believe all of all we have is in J GitHub cicd tools uh deployment tools and so on we have a reach of like a lot of data in in software development uh world so what we can do is really start working on

how we can leverage this data to represent the not just metric I believe this is the first step start collecting metric like Dora or or kpis what just Jack showed you but start building those metrics as a part of the causal model so those metrics are actually consequences or implications on the way how you structure your value stream on the way what activities you put there and

the way what technology and how many people do have those metrics will be uh outcome and uh this is where you need to have integration part and you need to have a on the application layer analytics and simulation to uh verify various scenarios of the changes of the structure of value stream and how it entails and or effect on the kpis and metrics of your um of

your uh company as we talking about process digital TN it might be slightly easier I just provide you overview of architecture that we follow so we slice the um architecture of the processes or value stream digital twins in four layers of course you have data sources you have an integration layer where you need to and there are you don't need to build this integration layer there are

plenty of U Solutions in cloud in asure AWS or whatever cloud provider use but on the next layer you need to have Services right simulation as I mentioned you need to have a causality uh in your value stream you need to have a calibration engine to make sure that your simulation model is really reflect what is happening in reality to make sure okay the scenarios that we

experienc in the past is really what we can observe in simulation model the simulation model can give you much more it can answer not just what happened but what could have happen and what implication it might entail optimization engine to optimize and find Optimal configuration of the value streams analytics of course you can uh think about Tableau powerbi or whatever solution you like even like python notebook

why not you can uh get the data through the simulation model um analyzing this and then last but not least is modeling language so how you represent the value streams I would um you can think of multiple ways some companies like still use bpmn which I believe it's more focus on the processes and very difficult to embrace high level value stream uh or you can use like

real notation um which make things simpler yeah and uh focus on the value and focus on the waste you can like really nail down those two aspects um yeah and application layer this is what we have like visualization of um those kind of services um you can plug it in uh with our platform on API layer we uh you can leverage geni or um natural language interface

through llm that helps you to understand where the bottlenecks are you can really have the model that you can speak to asking okay given this structure given this historical data where my bottle legs are why do I experience that what options do have uh to to improve so it really helps to drive their uh performance Improvement and Innovations in the value streams so we know that we

have thrown a lot of information at you in in the past 25 minutes um in which bille and I have invested many years to get to this point so it seems quite easy and obvious to us but in 25 minutes it's probably not easy or obvious to you we would invite you to leverage many different options available to you first of all you can come to vs

optima.com or value streampro places that we think we probably have the greatest opportunity to grow is around AI Ops or artificial intelligence for it operations and leveraging digital twins as a better way of understanding our organization our processes our environment and all of those things that we care about all of which bring us to the concept of of value stream management five years ago it wasn't even

a category for Gartner now it is something that most organizations have very high on their list of things to do but we would very much recommend that you reach out to us and to others and we would be delighted to help you on your journey please feel free to contact us and let us know how we can help you you have any questions or anything that we

can answer for you today that was uh thank you yeah let me um I'll also check just to see if there are folks asking questions um there are no questions in the Q&A uh neither are there any questions on the chat dialogue too um I can try to ask folks um here's your chance uh you still have some time uh 15 17 minutes uh to ask all

the questions You' like to ask of Jack and paval so yeah feel free to do that I me we'll give folks like a minute or so to uh ask questions any no maybe also is the end you know end of the day maybe folks are yeah maybe tired or maybe we answered all of their questions in advance yeah or perhaps they're just waiting until they have more

information to share with us when they reach out and ask us for help because we want to help folks that's what we're here for yeah feel free to reach out with us later you have our emails we' be happy to have a conversation and um yeah be happy to meet you on LinkedIn as well yeah great yeah wonderful yep so if there are no questions I believe

we can wrap up and um we appreciate this opportunity to present thank you yeah thank you both paval and Jack and uh yeah that's a very interesting too to yeah to look into digital twin and uh yeah and I I've heard of it the term but I just didn't know but now I learned something so that's cool that's actually very true for almost everyone we speak to

that they've heard of value stream maps and value streaming um but it's usually not very formalized it's usually uh like we talk about with budgets and goals a lot of times if they're loose they just don't happen and the digital twins have been around for a very long time they've been very mature in Aerospace in Pharma in petrochemicals but in our front offices not so much and

that's where we think that there's the the biggest opportunity now yeah yeah very cool yeah wonderful great thank you very much for having us have a wonderful day thank you yes thank you yeah have a good day everybody thank take care for