Mozart Orchestrator: Split & Merge for Parallel Process Automation | April 2026 Product Club

Hello everyone, welcome to product club. I'm Lou, the community manager here at Automation Anywhere and joining me are Nishiantth and Max. We are excited to have you here today with us. I'm going to go ahead and get started. We are excited to dive into Mozart orchestrator with you all today and highlight the product as well as the new split and merge feature. The way things will go is I will share some upcoming events with you and then I'll pass it over to Nisha Kant to walk you through what the product is and what it can do and then he'll show you a demo. After that, for those of you who have been to our product clubs for the past two months, you'll see a familiar environment with Max that we've been using to sort of product stack over the course of the past 3 months. And we are going to now incorporate Mozart Orchestrator into that flow. And then at the end, we'll open things up for some Q&A. Real quick though, this session is being recorded and we may make some forward-looking statements. So, please base all purchasing decisions on current capabilities and we'll call that out clearly as we navigate the conversation. Now, we have a few different ways for you to tap into the resources we provide to you in the community. If you're into in-person meetups, we have some coming up in Phoenix tomorrow. I'm actually catching a flight in three hours to head down there myself. I hear it's pretty hot. I'm in again Colorado, so it will definitely be a bit of weather shock for me. But we have Mumbai happening on the 10th, Toronto on the 16th, and Pune on the 17th. If you consider any of these areas local, you should 100% register. Great information, great networking, food, and swag. So, please register. We we'd love to meet you in person. And as always, if you don't see your area listed for our in-person meetups and you'd love to see one in your city, reach out to us, let us know, and we'll try to work with you, work with other people in the area to get one going. If you're into hearing from other companies about their automation strategies, what worked and what didn't, things like that, you'll want to check out our user groups. So we have one today about accelerating clinical decisions through automated lab screening and then another on the 14th around passing certification. So tips, tricks, how to get through that, the best way to uh grow your level of knowledge around whatever it is you are taking the certification on. So that's a really cool one to register for. Now, I do want to highlight the developer meetup product club mashup that is happening on the 23rd. I forgot what month it was for a second. April. April 23rd. That session is directly tied to what you're going to see today. In that session, you'll be building out parallel automations with split and merge. So, definitely worth taking note and registering for that one to get handson with what you're going to see today. I'm going to drop those links in the chat now. and go ahead copy and paste it to whatever you're using to take notes today and register after the session. Okay, look at this guy right here in the orange lanyard. This is from last year's Imagine and that's our boy Max helping someone out with an Aentic Quest. His hair is definitely looking better in the picture than it does today. That's just my opinion. But the reason I'm bringing this up is because I want to know if any of you guys are going to imagine. Let us know in the chat. Will you be there? Yes or no? We are super pumped about it. And I think I and Max are super pumped about it because we're launching a new certification called the AI automation engineer certification. And it's going to be a really good one with a ton of information. And because it is so new and informationheavy, we are hosting a hands-on support session at Imagine to help you with passing that. Uh, Max, do you want to speak to this at all real quick? What can the people expect? >> Well, I did do, but now you've given me a complex because of my long hair. So, I think I'm going to go get a haircut first. Of course. So, with our new certification that we're going to be launching, this is all around kind of utilizing AI both within the platform and just generally utilizing it as well. We go over all of the elements of how you can properly adopt this both within the platform but then from a monitoring perspective as well to make sure that you are utilizing this correctly. Getting the most out of everything that you need and really optimizing your automations and business processes by the use of the technology and what we have to offer here. So this is going to be handson. We will be able to walk you through all of the hours of content that we've been able to create, tell you exactly what's going to be important for you to look at, what's not going to be important based on whereabouts you're at, and really have it give you the best start with regards to Yeah. going through the content that we have here and trying out the new features. >> Awesome. Thank you. Yeah, I see a good group of folks headed to Imagine. So, we're super excited. If you can't make it, such a bummer. But we will try to get you the information and support you. Just because you can't make it doesn't mean you should be restricted from being able to take this sort of certification. So, we will we'll work on it. I do see some folks are unable to copy the links from the chat. So, if that is the case, you have registered for this session. I'm going to be pulling your um information to send the recap of the session and so I'll include these links there as well. So, not only are they going to hit the chat here, they'll also be hitting your inbox shortly after the session has completed. Okay. So, I do want to go ahead and get into it. And I think real talk, most business processes were never meant to be sequential. But for a long time, developers had no choice but to build them that way. Parallel logic got squeezed into a single line. Workarounds piled up and processes got slower and harder to maintain. So split and merge, which is our focus for today in Mozart orchestrator, changes all of that. And Nishi Kant is going to show you exactly how it impacts flow. But first, I'd love to launch a poll just to see where everyone is at on their journey to help us navigate how deeply we're going to go into Mozart orchestrator um in the beginning and then diving into split and merge. So, I'm going to launch that poll really quickly. Um okay. What best describes your current use of Mozart orchestrator? is that we actively use it in production. We are piloting or evaluating it. We have not started yet but plan to. We do not currently plan to use it or unknown hoping to learn more today. Okay, I'm going to give us about five more seconds. We still have a lot of responses coming in. Now remember the mosa orchestrator this is referring to kind of that functionality that originally was called process composer as well. So if you are used to calling it by that term but you are utilizing the new features like the swim lanes that we've got in there and the split and merge features as well then uh that's also in there as well. Just a uh just a heads up. >> All right I'm sharing those results. Most people unknown hoping to learn more. We have a few that haven't started yet but plan to and then a couple people who are actively using it in production. Now I do want to based on these responses Nishi I would love it if you could just give us a quick overview of what Mozart orchestrator is and why everyone on this call should check it out because they automatically have access to it right? >> Absolutely. Thank you Lou. Hey uh good morning, good afternoon and good evening everyone. This is Nishikant uh from the product management at automation anywhere. Happy to be here. uh so before I get into uh the details Lou what I'll do is I'll just spend uh a few seconds to explain why are we here right so before I even get into mosat orchestrator or get into the specific topic that is fit and merge let me start with the road to agentic process automation and I believe that'll answer your question and all the folks here would be able to know better based on that well so let's look at why process and how AI agents come into the mix and the intent of this slide is to create a much better or a shared language for all the capabilities that's shaping modern automation and the first place to begin is with automation itself. So automations as we know these are discrete executable they do tasks they are usually deterministic and more importantly they don't go away in the agentic world they become tools and next we have processes this is where we move from tasks to end toend workflows coordinating automations people and systems end to end with governance whether we can automate every step of the workflow is a different thing but we are thinking about What does this to end to end process look like? And next we have yeah agents. So agents usually use a reasoning to coordinate the use of tools and these tools could be automations, APIs or even other agents and that is to accomplish the specified goals. And one important point here that I would want to be really very clear about is agents don't replace automations, right? And I would definitely repeat that that they actually only elevate them. They don't replace automations. They in fact use them and their superpower is reasoning. So deciding which tools to which tools to use in what order or based on context and goals. And this is where it all comes together. Agents, automations, processes and humans all orchestrated to reliably run mission critical business processes. And you can see we have some plus signs added in here. And I wanted to do that. Uh so visually we could see that agentic process automation is really all of these things coming together. It's not a totally different thing at all. And now that we are aligned on definitions, right? Uh yeah, now that we are aligned on the definitions, the question becomes how are these things actually different in practice? And I want to start by taking or let's say talking about the scope which is essentially what each of these things handles. As you move right, you are expanding what can be handled that is from individual tasks to workflows to goal-based cognitive work to full end toend business processes. These also all differ in their approach which is all about how work gets done going from deterministic linear processing to structured workflows then adaptive reasoning with EI agents and finally orchestration that blends the deterministic and probabilistic capabilities together. To be clear, this isn't one versus the another thing. I look at this as a cumulative capabilities of all brought together in agent process automation. So you don't ditch automations when you build agents. You in fact elevate them. And finally, I see all I I see all these are different in the way they handle variability. And I think this is one part uh where we have to be very clear. For example, uh as we know automations usually are deterministic in nature. they cannot handle variability. Whereas processes usually are deterministic but with human in the loop as well as bringing all the other uh automations in the mix together. Whereas when we go to agents they actually can handle variability. So due to the reasoning capabilities that they have they can let's say solve lot of ambiguous problems or situations that we have typically in any business processes. And the big takeaway this isn't agentic process automation instead of task bots right I know I'm repeating this but that's that's that's exactly what I want to emphasize it is basically bringing together your task bots with other complimentary asset types so that we can enable greater flexibility greater resiliency and better automated coverage of real end to end business processes. So understanding the differences definitely is is important here and uh let's look at how we can prepare or get into this journey and the best part is this journey starts with your automations what we already have you build them but also being intentional around reusability right so we have to ensure that okay we have a clearly defined inputs outputs with descriptions proper error handling and logging with each automation having a description Otherwise there would have been a collective gap here right and these are the things that usually all of us have been doing uh for ages and typically this can be run on your uh servers or it could it could also be run on cloud services once you have got that so once you have got that the next uh natural step is how do we make these automations intelligent well there are multiple ways that we support one is uh bringing in the AI skills where you can have specific prompts that can either be used for summarization. Uh it could be for let's say text generation and so on. But it is for one specific use case. We also have ability to add Gen AI models. For example, let's say if you already have subscriptions to your Gemini models or you have OpenAI, Azure OpenAI or it could be any other models. If you have those subscriptions, you can already use it as part of your automations. And in case you have not used generative recorder which which makes basically makes your automations resilient already uses some of these generative AI capabilities. And the next is how do we get onto the process first design? Well, think of it from a business point of view. Today uh as I said we are mostly focusing on uh task based or specific tasks that I can automate which are typically let's say uh done multiple times. So it is repeatable and that's why I'm automating them. However, when you just elevate that to a larger business process, then you can actually design that using Mozart orchestrator. So this is where you can bring in multiple automations at one single place or linked together logically as well as you can bring human in the loop. Typically any business process would have let's say uh reviewal approvals or it could also be input by some of the business users or it could be for various other reasons but process lets you bring human in the loop as I said and while you're building these processes there is also or there are multiple ways that you can build it in a much faster way. One is you can use autopilot which will give you a head start where you can use your existing BPM diagrams to easily convert into a scaffold of the process and not just the process even a skeleton of all the steps being converted to automations can be seen and to further fasttrack from there we have co-parent for automators which basically lets you give instructions in form of natural language and that lets you build the end toend automation beat processes or it could be your tasks and APIs and we always have document automation. So within the process you can also have document extraction as a step or you can validate the documents uh based on the extractions that have already been done. Now we have automations already. We make them intelligent by adding AI skills and geni models wherever necessary and then we look at the broader processes. Once we have processes in place next is how we can get into EI agents. Now EA agents as I said they run autonomously. They have the reasoning capability as long as you have clearly defined what the goal is. You can add all the various tools that it can use at its disposal disposal and dynamically make decision on what is the path that it has to take. Now as part of the various steps or the path that it takes it can have human in the loop. It can use AI as skills wherever necessary or it can use API task which can run on your cloud as well as task bots and not just that it can also call other EI agents whenever necessary. Now when all of this comes together is what we call as agentic process automation. When you go through this journey and have AI agents processes intelligent automations you would naturally be in a state where you have agentic process automations. And the best part is I'm sure you would have your uh systems like Salesforce, Service Now and other SAS applications which do provide EI agents. Now the best part is automation anywhere lets you connect to the third party EI agents or it could be any other systems of record either through MCP or A2A protocols. As I said definitely the ROI that we get when we move from task automation to enterprise transformation that is looking at the broader business processes. This is how it looks like. Definitely the intelligence and complexity that you get or solve increases as you move from intelligent automation to processes agents and the agent process automation. But at the same time the ROI or the impact that you make on the business also increases when you move from journey of task automations to agentic process automations. Now with this in place let me just quickly show you what the automation architecture looks like and I know uh there's lot of content here but the things we just discussed they all represented here and then let's look at how we go about it. So we have conversational interface at the top where you can use conversational interfaces to run your automations as well as agents. Now these interfaces could be your Microsoft copilot, your Microsoft teams or it could be even uh embedded u into your uh day-to-day applications that you use like Salesforce workday and so on. And below that we highlight solutions which are pre-built expressions of what we just talked about which is to let you accelerate your journey uh which can also leverage Isera that we had acquired late last year and other pre-built industry solutions so that it can accelerate your agentic journey and in the bigger box towards the bottom really highlights the benefits of automation anywhere in terms of uh what is the underlying technology which makes all of this possible as well as the processor reasoning engine which enables our proprietary approach to most of the AI capabilities that we have. Now let's look at how we can get onto this journey, right? Well uh for that we have uh Mozart orchestrator which uh has been completely reimagined compared to what uh the earlier version of process composer that we had. And just to uh add on to what uh Max had uh mentioned a few minutes back for folks who are seeing this for the first time. Well, Mo was Orchestrator was also available in form of process composer earlier. However, it has been enhanced now. It has more capabilities. It has a new engine that powers lot of advanced orchestration capabilities as well as it also is having a tons of observability that you can actually leverage to look at the operational insights as well as track the ROI. Well, so as I said with Mosot, you can now build and run complex end toend workflows that involves bots, APIs, AI agents, documents, and most importantly also people. And all of this we can do with speed, security, and scale. It's built for the enterprise, which is bringing together the power of deterministic workflows as well as the agentic intelligence through four key pillars. One, how do we accelerate build time with visual AI assisted canvas? And this supports RPA, APIs, documents, CI agents and humanity loop. And all of this powered by process reasoning engine. It supports longunning parallel processes which can be triggered by API calls. It could be also scheduled or any other real world events. So with AI agents that dynamically reacts to external business conditions. Now you can also connect effortlessly to third party AI genes and tools via MCP and A2A protocols and not just that you can also use thousand plus pre-built integrations or any low code characters for rapid extensibility and we also ensure reliable operations with built-in observability as I said and there are also AI guardrails in case we are using AI skills or AI agents. It contains audit controls and centralized policy management for compliance at scale. And what we have seen is we have more than 90% of year onyear growth in terms of number of customers who have started using Mazot orchestrator to get onto the journey of agentic process automation. Now with that let's get to see what is already available and what is coming soon. So here as I said this is this is already available in your control room. There could be some capabilities which may not be available based on the licenses that you have but the core capabilities for you to create and orchestrate processes is already available in the control room in the cloud version that is 36. Now it does allow you to model your process automations uh through the PPM aligned uh canvas. Uh as I said you'll be able to orchestrate EI agents BS and all the other automation assets in one single pane. You'll be able to have multiple E aents within your workflows and more importantly which is what we're going to talk about is parallel processing through split and merge in version 40 that is coming out soon. We'll also have start triggers such that you can trigger your automations or process automations based on any external events that we have either out of the box or any uh web hooks that you may want to create later this year. You'll also be able to simulate and test your processes end to end. And not just that, you'll also be able to optimize your processes based on the suggestions and recommendations that it would give. Nishi, we have a we have a couple of questions that I think fit in really well right here. The first one is I know that we Oh, where did that question go? Okay. If we have existing processes built using process composer, will it automatically convert to Mozart orchestrator when it upgrades or is that something that they have to transfer over manually? >> Yeah, I think that that's a very uh good question because we have we have heard this lot of time from customers. The good news is you don't have to do anything differently. Your existing processes will work on Mozart orchestrator. It'll get all the benefits that we have in this new version including let's say swim lanes. Uh you can have a new look and feel of the canvas. You'll be able to drag and drop. You'll also be able to see the errors on the screen and it is quite free flowing. So you'll be able to organize the way you want or the way your BPM file looks like. And also the capability that we're going to talk about which is split and merge also will be available as long as you have the license which is in this case enterprise license. So yes short in short Lou customers don't have to do anything they'll get the benefits. >> Okay and then I do have a couple of questions you just touched on licenses. So for Mozart orchestrator as a product that's available to anyone right or is there a specific license you need to be able to access that >> right so Mosot orchestrator is available for all customers who have the uh control room there are only specific features within Mozart orchestrator that is controlled uh through license but otherwise you will be able to create your processes even today you'll be able to organize them uh into different swim lanes uh uh use all the benefits like uh being able to add your BS, APIs, or even other uh humanity loop tasks. >> Awesome. Thank you. And there is one more question, but I think we're going to touch on it a in just a moment. So, uh let's keep going with this and then we I'll highlight that question once it comes up. >> Sure. Absolutely. All right. So we spoke about uh the journey to agentic process automation, why processes are important and how MOSAT orchestrator helps you get onto that journey and what we do understand is as orchestration scales realtime observability becomes very critical not just to track what's running but to ensure that the issues are resolved proactively and performance is continuously optimized. With Mozart orchestrators, you get live monitoring of every automation from current steps in progress to detailed histories of completed flows. And this helps identify issues uh as they happen and also give you full visibility into the process health. And soon we'll also be having a view of the entire process exactly the way you have designed and give heat maps in terms of where exactly the bottlenecks are, where exactly are there slowdowns. uh so you'll be able to optimize the processes based on those additional informations and uh for troubleshooting MOSAT orchestrator will support custom logging. It'll have detailed error tracing as well as process level diagnostics. So it empowers both uh the dev as well as ops teams to act faster, resolve issues quickly and scale reliably. And going forward, uh, we'll also be expanding this further with conversational help, which means you'll be able to just ask your questions in terms of what you're looking for in terms of insights or any kind of information that you're looking from dashboards. You'll be able to just key in your questions in form of your usual natural language and you'll be able to get all the information immediately. With that, let's look at what exactly are we saying or we mean by split and merge. Well, so this is a new capability that we are introducing as part of uh version 39 patch and it is getting rolled out to our cloud customers as I speak here. Now it is designed to make automations of complex and longunning workflows far more efficient. In many real world business processes, multiple tasks or approvals need to happen in parallel. Whether it's processing loan documents, validating transactions, or coordinating multiple department workflows. Until now, this had to be modeled sequentially, which added time, complexity, and made maintenance quite harder. With split and merge, you can now model and execute parallel branches within a process. And that will allow any non-dependent tasks to run simultaneously and also converge at one single merge point. So this will result in faster execution. It'll improve scalability and also a process design that's much more closer to how you usually design in any external tools that are based on BBM. What's also important is that many customers have been asking for more advanced orchestration capabilities that can handle real world business processes and split and merge actually strengthens that position as an enterprisegrade agentic orchestration platform. And to walk you through how exactly this works and how you can create parallel branches and also get them to merge at a single point. Let's get into a live demo. So, I'm just going to switch my screens. Okay. As you're doing that, I'm going to launch another poll because I think this will help navigate the the demo and show you where to dive into things uh in more detail. So, we do have this question. What is the biggest challenge limiting your process automation adoption or expansion today? Is it identifying the right use cases, building skills and confidence to design process automations, getting started with the product and creating the first process automation, or demonstrating business value and ROI? I'm going to leave this up for about five more seconds. Uh >> so while everyone is uh responding to the poll, Luke, can you just confirm if you're able to see the control room where I have a employee onboarding process? >> Yep, we got it. >> All right, >> I'm going to end the poll and I'm going to share those results out. So the biggest one is getting started with the product and creating the first process automation, which I think today's session is definitely going to support with that. demonstrating business value and ROI, building skills and confidence to design process automations. We have a ton of resources in our Pathfinder ecosystem to support with that. So, we will be I'll be mindful of sort of throwing that out as we go through this as we go through this demo and then identifying the right use cases. >> All righty. So, let's get on to the demo. So this is Mozart orchestrator. As you see uh the entire editor or the canvas has been enhanced. It's quite uh visually improved. Uh you'll be able to drag and drop all your nodes. You can move from one uh lane to the other lane. You can add your swim lanes either vertically or horizontally based on what actually fits your process. as well as you have all the automation assets starting from AI agent APIs boards as well as human in the loop for approvals or even document validation and not just that if you have a really large workflows and that has been split into multiple processes well you can have one process call in the other process and we do understand that usually the processes will have multiple conditional logics for that we have if and else and more importantly split and merge which is what we'll look into in detail and if there are situations where let's say you want to jump uh from a particular step to the other step if certain conditions don't meet well we also have go to now as I said split and merge lets you actually have multiple or I would say series of steps to be run in parallel to do that you just have to drag and drop split and merge onto the canvas like what I do here in this case let me add it here which basically makes it as nested, split and merge. As you see here, you can have multiple split and merge branches not just uh which are running continuously or parally. You can have nested as well and you can have as many branches although we have certified for let's say 10 branches because that's typically what we see most of our uh customers have but you can have more than 10 branches as well. Now when you define branches there are multiple things for example yeah definitely you can give a name basically based on what exactly this particular branch is going to do in our example which is employee onboarding process we have three different branches and I'm sure most of you would agree that typically employee onboarding process can have multiple steps running in parallel for example as part of your onboarding you would have your finance team who is working on the payroll systems you'll have your HR team which is getting your profile already on workday or any other similar portals and you'll also have IT team working parallelly for getting the access to your IT assets right and that's exactly how we have designed it here now when you design these different branches you will be able to decide whether you want to run these branches conditionally so maybe only based on certain conditions a particular branch has to be run or if you want a branch to be run always run you can select this option now if you have most of your branches conditional you'll have to have at least one branch that has to be a fallback so that the flow doesn't get uh let's say stuck at a point instead there's at least one branch that it can execute and then logically get concluded so in this case we have three branches one is payroll as I said uh workday and then email access and this is exactly how you would create conditions and this is something that uh you have been doing as part of your process conditions or even within the bot so it's exactly the same here Now within a branch you can have a series of steps. So you can have a API task or any other automation asset that you see here on the left p. And what you'll see uh here and which is also quite different from uh most of the other orchestration tools is every branch will have a branch exit and all of these branches would logically add at one single point called as merge. Now if you use any other tools you would have to create a merge point and add additional logic. Whereas in automation anywhere every split and all the branches will automatically have a merge point. That way you don't have to write any additional logic. And what it also means is this particular point will wait for all the branches to get completed until it moves forward. So you don't have to write a logic to look at all the branches see if they have got completed and only then move forward. that's not required. Automatically the system will wait for all the branches to get completed. So in this case uh I have added multiple steps in each of the branches and in future what we'll also have is give you additional capabilities to write logic as part of branch exit. Maybe you'll look for some conditions and only then you would want to exit the branch logically or you may want to have some post-processing logic. For example, let's say you have various automations, bots, APIs that have run in the branch and maybe you want to collect that information. So we'll also have postprocessing nodes as part of these branch exits in the future. Now this is what we have. Now let's see this in action. Until now if you had to design you would have designed let's say maybe payroll to begin with. Once all the payroll steps have been completed then you would have workday steps and then you would have maybe email and access or any other branches that you may have. Whereas in this case you'll see all of this running in parallel. So let me just undo some changes and let me run. All right. A very simple form here. So I'm just going to give uh employee ID, give a name, say whether it's a contract or not. Select the location. Let me select save and then click on submit. What you'll see is definitely the next step gets executed. Once this step gets executed, we expect to see all the three branches running in parallel. So let's wait for a second. All right, here you go. You see payroll, workday as well as the IT branch. All the three running in parallel. And what you'll see is only after all the three branches have been executed, you will see the next step or the step that is right after the merge. So let me just complete some of the steps as part of these branches. So you have one which is coming from the uh HR side, the other one which is coming from the IT side. Let me just go ahead and submit this. Now until I have submitted this, you will not see the rest of the flow moving forward. So let me also submit this. And now that all the branches have been completed, you will see the rest of the flow continuing and moving forward. I hope uh it was clear where you could see all the branches running in parallel and then once all the branches were completed the flow moving forward. All right. Apart from that uh we do have quite other rich capabilities. Let's say if you have a quite a large workflow you definitely need uh a mini map to be able to navigate to different sections of your process. That way it'll be easily uh you'll be able to easily move to different corners of your flow. Uh we also have mini map which can fit your entire flow in one single uh pane that you see or you can use your control uh plus mouse or you can use these icons that you see here to zoom in and zoom out. And not just that you'll also be able to search for any text or any data that you see on the canvas as well as on the property panels. For example, if I search for something, it'll be able to search for say birthday. It's able to search across the canvas as well as anything on the right property pane. You'll also be able to use the same advanced search that you have used uh in your task boards as well as you'll be able to let's say search and replace any of the text or the properties that you have added in a process. Now you also have ability to build it at a much faster pace using copilot for automators. In case you have not used this, this is available as part of the automated AI uh package where you get to build or let's say update your processes or even task bots giving instructions in form of natural language. All right, with that we have covered uh the demo of split and merge. I hope you you have been able to fully understand how this is uh useful as well as are able to use this as part of your business processes or process automations that you have created. With that, let me hand it over to Max. >> Yeah, I think before we do that, I do have a couple of questions in the chat. So, BJesh asks, "If the process doesn't require a human in the loop or AI capability, is there benefit to using this instead of a masterbot?" >> Absolutely. Uh I'm in fact I was expecting this question. Uh definitely there are benefits. Uh uh so whenever you have master and let's say child bots right typically you have designed it not looking from a process perspective specifically business process right you just have a series of steps that you have uh stitch together using masterbot and childbot but when you're looking from uh not just let's say cutting down the hours but looking from efficiency of your business processes then typically you would want to create them as a process automation and not just uh child and master bots. Now apart from that uh what will also help is when you want to let's say uh transition to agentic process automation you don't have to again change your existing bot to process and then to agentic process automation it gives you natural uh transition path to agentic process automation >> great thank you and then if all three processes that you were just showing are running on the same server wouldn't they be impacting common resources like excel or word and if so how is that managed >> all Right. Very good question again. Uh so this actually takes me to one important point uh that I want to share. Let's see if you have task bots as part of your branches. Uh then definitely they will need bot runners. So which means if you have let's say mapped a process to a device pool, you have to ensure that it at least has let's say as many devices as the number of branches because only then it'll be able to run in parallel. However, if your branches do not have task bots, but they only have APIs, forms, and AI agents, then you don't need to worry. It'll still run in parallel. It's not going to clog the server, or it's not going to really impact the performance. Even with task bots, let's say you have a device pool which has three devices, but you have four branches. Well, the three branches would run in parallel. The fourth one would wait for one of the branches to get executed, and then it'll continue. So it will be efficiently be able to manage uh the resources and the devices. >> Okay. And then can Mozar orchestrator be triggered via API? The example they've given Raful, we have a dashboard in Oracle Apex. We do batch upload of applicants trigger orchestrator through API. Is there a way to run task bots through orchestrator paralle parallel sorry in parallel on a single machine and then follow-up question what about the PII how is that being safeguarded >> all right so process automation that you build using mozz orchestrator uh it'll follow the same guidelines for governance that we have for taskbot so be pi data or any other confidential data is still going to follow the same guidelines so there's nothing new that you would have to Yes, if you are going to use EI agents, EI skill, we do have EI governance through which you can decide if you want to mask any uh specific type of data and it'll also give you the toxicity rating. So, you'll be able to uh govern that. I missed the first part of the question. >> Uh the first can it be triggered via API? >> Yes. So, you will be able to trigger through APIs or through even web hooks. actually that is coming as part of version 40 later this month where you will be able to use out of the box triggers that we have web triggers or you can use our universal listeners to create a web hook based trigger >> right and then to build on that we are in July hosting a session around that here in product club and then in June we are hosting a session on um AI governance so Rahul I hope to see you at both of those because they will support you in this journey that you just posted about. All right, Max, take it away, bud. >> Perfect. Thank you very much. Let me share my screen now. Of course, it is just a second logged me out of the control room as well. Beautiful. Let me just quickly log on back in. Okay, can you all see my screen? Can you give me a nod? Beautiful. >> Yep. >> So, just touching it then on the split and merge process. Now, one of the poll questions that came out with the answer being that getting started with the product and creating the first process automation was kind of that blocker in play. I just want to touch on this slightly. Now, when it comes to using Mosar orchestrator, it's also a case of shifting your mindset about how you're starting to focus at looking at managing these processes. Now with the split and merge, it's also about efficiency. Yes, you can do all these tasks one by one. But do you really want to do that? Now people that come to these sessions, you know that I like my analogies, and I have another one here. Now, I know I should be doing something music related, but I've got food on my mind, so I'm putting myself in the kitchen for this one. Think of it like you are a chef. If you get a large meal that comes in, so let's say you're getting steak, vegetables, and some chips as well on the side. As a head chef, you're not going to cook all of these one by one. Yes, you can keep it all warm. Yes, everything stays, but it's not the most efficient way. You would manage this by dictating different groups to do different parts of the cooking. You'd have the one chef do the proteins, another chef would sort out the chips, and another chef would sort out the veg. These would then all come together at the end to be a perfectly timed meal. And it would be a lot more efficient to be able to do it this way. And that's exactly what this split and merge functionality enables. Now, we did have one question come in where it's a case of, well, why wouldn't we just use a masterbot? Why kind of go down the route of using process composer? And this is why. the flexibility and the efficiency that you can achieve with your process times, with your executions, with even just managing the builds itself is a lot easier. You're no longer having one giant automation file that you've got not just actions added to, but different bots, different data passing through. Whereas, if you've got it like this, you're forcing your builds to be much smaller and a lot more modular, which means other developers can easily pick it up if they need to when it's not their logic and not spend as much time understanding the build itself. Now, with going over spit and merge, Nisha did an absolutely perfect job where he has covered everything and all the features that's in this. So, this is pretty much a recap of kind of that way of thinking around it. Remember, utilizing the split and merge functionality and making sure that you build out a process rather than a master taskbot isn't just about this is the only way to do it. This is about being the most efficient way to do it. And that is the difference. There's multiple ways on how to achieve the same result, but which way is the most efficient? Which way is the most reliable? Which way is the most scalable? And coming down a route like this is going to tick all of them boxes for you. Just to quickly recap, when it comes to actually implementing split and merge, all you're doing is dragging in the actions from the left hand side in the elements pane, the split goes straight in here. You add the description and then when you go down to the actual merging, this is the way that it goes. You've got your branch type in a case of always run. Do you only want to run based on a condition? And with the fallback in place, just in case any of them can't be executed. Now as well just touching on the mindset of all of this again the process that I have built here is to manage employee on boarding. So with this what I'll be doing is I'd be going through and adding the new employee into the organization chart and then I would need to send them an email. Now this could all be done in one single automation. Add them in and then send the email. But why do it all in one automation when I can run both of these at the same time and have it done even quicker? Even if I'm only saving a few seconds if that's going to be executing every day, every hour, that time adds up over the course of the year. And this is also a metric that you want to start monitoring now that you've got the ability to be able to build automations and processes with this level of efficiency. So I will quickly just execute this but that goes over the demo of what we've got here just to show that execution happening. So straight away for mine because it's a very basic example compared to both of these processes execute straight away. Both of these are in progress right now. Now I have the taskbot in here that does my desktop automation but you can see my API task has already kicked off. The API task is going to be go well the API task here that is importing employees into my organization. If I refresh this organization chart, we'll see how that's gone from 65 employees which was just here to 70 employees. So the five have been added and the desktop automation that's kicked off at the exact same time is doing my email sending. So this has already saved all of the time needed in order to be able to do this because well it's finished. So just touching back and just to quickly recap, remember when it comes to utilizing this functionality, it's not all about redesigning everything you've got. It's about how do you make it more efficient, save more time, which means you can in theory just run more automations with the time that you have saved there. >> Amazing. Thank you guys. If any of this got your wheels turning about capabilities within your own environment, I'm going to launch this poll. Check yes. This is just letting us know you're interested in follow-up. We can set up a tailored session to go deeper or we can just chat through specific use cases, whatever is most useful for you. This is just saying you're interested in that follow-up about this topic. And our goal is to support you in your automation journey and help you get the tools and resources you need to find success. So, just go ahead, let us know. Saying no is completely fine. You guys could be working on this and this is just a touchup, but I do want to offer this opportunity for those of you who are interested in sort of a deeper dive or use case evaluation, anything like that. I'm going to leave that open for a few more seconds. Uh Nishi, we did have some additional questions come through. Um, BJesh asks, "Does this require cloud execution or do we have to use multiple bot runners retrint?" Did we answer this one? >> Yeah, we did. We did. Yes. So, but I'll just quickly repeat. Yeah. So, if you have task bots as part of the branches, yes, we need uh device pool which has multiple devices. But even if you don't have as many devices, as I said, it'll efficiently cue the branches uh as required. If there are no task boards you you are just using let's say uh APIs APIs definitely required cloud but then customers don't have to do anything they automatically run on the cloud that we provision and uh if there are human steps uh nothing else is required. >> Okay. Are you considering adding any ETL functions to Mozart? >> Yes. Yes. We will have ability to have ETL also post-processing. So any additional data massaging or information that we would want to update will be possible. >> That's the road map. >> And then is this available in the community edition? >> Mosot orchestrator itself is available in community edition. However, uh split in merge that we spoke about uh is available only for customers who have enterprise license and I know enterprise license itself could be part of multiple packages. So based on which uh package or the license that the customers have they will have ability to use split and merge. >> Okay. And then in merge can we compare status of all parallel steps or have to m or do we have to maintain individual variables to track status of parallel executions? As I said uh once we uh give let's say post-processing nodes or additional conditions that we can check in branch exit you will be able to compare the branches. However, even today there is a workar around let's say you can add additional step it could be a task or an API uh at the let's say either after merge or it could be uh just before the branch exit where you can compare all the branches and take a decision. >> Great. any >> now with there are two raised hands in the attendees. I just want to check if if you meant to do that because you have a question, please put the question into the chat or the uh the Q&A. Yeah, I would totally click allow to talk, but in my experience, the majority of the hand raisers are accidental. So, I don't I don't want to put you on the spot if you don't want to be. That's my worst nightmare. >> Oh, yeah. >> Okay, we did have another question come through. Uh Jay asks, "We use WLMQs. If we have items from Q's running on botr runners as at the same time a process is running a bot step, how would the botrunner prioritize running the process bot?" All right. It would definitely take the priority that we have set for the boards. I'm sure most of you would know that we have priority and based on the priority the WLM cues would take the task boards. So be it from process or the individual boards. It'll run through that. But I'll just clarify u we don't have WLM capability for processes. uh but I'm guessing the question is in terms of if a same runner is used uh for a WLM as well as a B that's running from process then yes it'll still take based on the priority right okay I am just going to share this one more time I don't see any additional questions coming in so thank you all for carving out time for us today. We hope that you are walking away with something you can use. Just a reminder that you can get hands-on with this tool at our developer meetup product club mashup session taking place April 23rd. Uh so please be sure to register for that. And again, if you want to keep the conversation going, you know where to find us in the community. You can email us at community@automationany anywhere.com. And yeah, have a great day everyone. We appreciate you. And don't forget to check out the events page in the Pathfinder community to see what's coming up next. And we will see you there. >> Yeah. >> Have a great day. >> Thank you everybody. >> Thank you everyone. >> Thank you. >> Thanks everyone. Bye.