What's New in UAC 8.1: Smarter AI, Stronger Security, and Enhanced Automation

Hello, everyone, and welcome to the third session of StoneBranch UserVerse online twenty twenty six. We're excited to have you join us and continue for our four part virtual series. Today, we're diving into the latest product innovations in our session, what's new in UAC 8 dot one, smarter AI, stronger security, and enhanced automation. We'll be taking a closer look at the newest capabilities in Universal Automation Center eight dot one, including enhancements in AI, security, and automation designed to help your organization work smarter and more efficiently. My name is Lauren Tanzini from Stone Branch, and I will be your moderator today. Before we get started, let's cover a few things. This event is designed to be interactive between you and the speaker. If you have any questions, please add them into the q and a tab, and we will answer them at the end of the session. Additionally, on the right hand side, there are handouts that are relevant to today's session. With all that said, it's my pleasure to introduce our Stone Brand speakers for today's session. We have Gwen Clay, chief product officer David Ainsworth, principal product manager Niels Boor, vice president of solution management and Len Siemens, solution engineer. Let's go ahead and get started. Gwen, take it away. Thank you, Lauren. So let's move forward here, and we're gonna be talking about what's new with eight dot one. We just released this back in October, and it's it's really an advancement of where we are with UAC and especially with Robie AI 2.0. Robie AI is our AI service that we introduced back in eight dot one... Eight eight dot zero in the spring. We're also gonna talk about what's coming soon with the universal portal two o. You know, we're always continuing to expand the ecosystem, you know, all the all the, tendrils that kind of reach out from UAC into the greater ecosystem. Now the AI ecosystem with Len and Nils. And then, of course, we'll be giving an update on our universal date... Data mover gateway, three dot four. So let's, move forward. And when we look at Ruby eight dot eight dot zero, this was what we released in the spring. Right? So the first version of Ruby AI, we really focused on three key areas. Right? So creating faster with the build, operating smarter with run, and connect openly with integrate. Right? And we made advancements in all three of these areas. And you'll kind of see here when we we were talking about Robie AI around the build, you'll see build workflows with AI, and you had in brackets coming soon. Just a little foreshadowing, we've crossed this line out. And, we'll be turning this over to David Ainsworth to tell us more about it. Okay. So... Yeah. Hello there, everybody. I'm David, and I'm going to walk you through the highlights of the, eight dot one release. I won't cover everything that we added in eight dot one, and I fully encourage you to take a look at the release notes in our documentation site. But firstly, let's talk about the hot topic, which is Robie AI. You know, Gwen already spoke a little bit about Roby dot ai and the eight dot zero functionality. In addition to that functionality such as the Roby dot ai assistant and the analyze now, Roby dot ai has become the default way that a lot of users access the documentation. So we provided the new documentation site with eight dot zero, and we are constantly improving that content to not only improve the user experience of the documentation itself, but also to achieve more useful and more accurate results from Robie AI searches. So let's see what we did in in eight dot one. We added an interactive status update for Robie AI. So when you're interacting in the chat window, you will now see an indicator which is showing the current state such as thinking and executing a specific tool, so you'll have a much better idea of what Roby dot ai is doing at any point in time. You'll see this in action in the demo in a couple of minutes, and I'll remember to point that out to you. We also completed work on the more recent chat model Azure OpenAI GPT five dot one. Remember to configure that chat model deployment to GPT five dot one. You know, it really improves the results that you get from the chat, but also it keeps us current in a... In what we can all accept as a rapidly moving space. In eight dot one, we also added the AI workflow builder that Gwen mentioned, and the AI task builder. You know, the workflow builder allows you to create and extend workflows through the conversational interface. Roby conducts a structured interview, gathers detail it needs, assembles a complete workflow plan, and presents it for your review. And we have human in the loop at every step, so nothing is written to the controller until you approve at the end. And the task builder has a sub agent for each task type, assembles the full task definition, and again presents it for your approval before writing anything to the controller. You can see on the right hand side here the tasks that are enabled by default for, you know, timer, manual, Windows, Linux, Unix, IBM I, and SAP. There are... These are simply the most common and most tested AI sub agents that we have for creating, the task definitions, but you can also enable other experimental task creation sub agents that already exist. And you can see how to do that in the documentation. It will just... This workflow builder will make the creation of workflows and adding tasks into the workflows and creating complex dependencies extremely quick and easy. So let's walk through a demo of this brand new functionality. So we can see that I'm I'm in my, my system for the, UAC. And let's start Robie AI and get started. I'll just pull the chat window up here. So the UI always had drag and drop functionality to perform, you know, the the workflow building actions. When we want to add a large volume of tasks into a workflow, it now becomes much easier and less time consuming. I'm going to walk through a couple of examples of the workflow builder and the task builder. The first example that I'll show you assumes that I know exactly what I want to build and how I want to build it. So I can simply declare everything, upfront in the prompt. So here I have create a workflow called P EOD master 1,000 containing all existing tasks prefixed EOD to run one after the other in an alphabetical order. So I've really specified everything up front. I'll just let that run. It's... We... We're asking it to do a number of different things here. First of all, to identify and list all the tasks with a specific prefix, then to arrange them in the alphabetical order, and then connecting them with the dependencies that we specified, which is one after the other in alphabetical order. Right? So it's going to prompt me all the way through before... And tell me what it's about to do before each step, give me some options to make changes. So I can look... I can review the the task that it identified. I'm going to remove the two workflows from here, and I'm just going to tell it to proceed. So I've made a little a little revision here and a little change. And then it's going to show me how it will build the dependencies according to my instructions. So again, I can review these. And, yes, I'm happy with the way it's working, that it's going going to be in alphabetical order, so I can accept it again. And the final step is am I ready to create the workflow on that basis? And the answer is yes. So let's go and create the workflow. So Robby's doing all the work for me. The workflow is created. The dependencies are added, and I can now view the new workflow. And you can see very quickly and easily it's added all the tasks as as I told it to add them. Right? And also, I'm now in the workflow editor, so I can make any changes or add more criteria directly in the way that I'm used to doing that. Yeah. So that's the first very quick example. And I'm going to close that chat window and just have the... A brand new context because it's, you know, it keeps the context and it it remembers the chat that we've just had. The second example I'm going to show you is... Assumes that I want to be guided through the workflow building process and I want to be assisted at every step. So I'm going to put the minimum of information in the prompt. And here, this is create a workflow called p e o d master 99. Let's run that. And the first thing it's going to ask me is to specify some tasks. So I can either start typing, paste them into the chat, or I can use the find task button that I have here in the chat. So I'm going to start filtering now. This is an advanced filtering capability. I'm going to look for anything with the prefix of e o d. So I've already got a a significant list here, but, you know, there's a lot more filtering criteria that I can use. So there's all these different options that exist, or I can use the advanced filter. And in this case, I'm going to say I don't want any... Anything with the type of workflow to be included. So it's just taken the workflows out of that list, and I can now add them into my selected task list. If I want to, I can go back and and carry on, you know, doing some more filtering work to to add in here, or I can just be happy with the list that I've already added. Right? So now I'm... It's going to present me with a list of the tasks that I added and I asked it to find for me. And at that point, I can make, changes to the list. I can verify the list, or I can I can make some additions to it? So you can see here, we have the, full list of of tasks that I asked it to include, but I'm actually going to add some extra tasks which is on my end of year tasks, and I can just use pattern matching to find the exact exact task that I want. So I'm using a combination of asterisks for multiple characters and question marks for single characters, and it's going to go and find those additional tasks and add them to my selected list. So you can see them down at the bottom of the list here. Okay? So I'm happy with that list of tasks. So next, Roby is going to want to know how the tasks should be ordered. So it gives me some ideas about how I can order them, and and I could simply start typing how I want to join them together, or I might already have an operational spec description, you know, when I was building the application automation spec. So in this case, we can see in my chat window that I just really joined everything as I want it to be with arrows, and I'm just telling it exactly how I want it to look in my workflow. Right? So let's allow it to proceed. And this is just a very simple way to join it together, but there's a lot of flexibility of how, you know, you can say I want this to follow this to follow this, and you can really just do that in natural language. So again, I can review. This is how it's going to build the the workflow with all the dependencies, and I'm happy with that. So let's carry on. And now it's... Am I ready on that basis to create the workflow? So now it's... It goes away, creates all those dependencies that I asked for, including all the tasks that I requested, and confirmed. And now it's going to present me with a view workflow button. So this is a little bit more complex than the, than the first one that I built, but you can see it's just, you know, taken a lot of the legwork out of building my, my workflow here. Right? And now I'm in the, you know, the workflow editor so I can do anything else that I want to do. So that's two examples of how to how to build a workflow. But what if we have an existing workflow and we just want to update it? We can also ask Robie AI to do that for us as well. So I'm going to update this workflow, p e o d master 99, to add a task that is going to run between EOD process 14 and EOD process 16. So first of all, it's going to identify this is the task that you want. Yep. That's the one. I'm happy with that. And then it's going to tell me how it's going to do it, and then it's going to confirm that I'd like to update. So I would like to draw your attention to the actual workflow itself over on the left hand side here where my cursor is because it's in real time added the task into the middle of the workflow and dynamically updated it. So I could, with natural language chat, really be adding to to the workflow or any existing workflow, and I can see the results in real time. Right? So it's a very powerful addition to to that workflow builder and task builder. So what if I I now have a I want to add a task that doesn't exist yet? So let's create a Linux task. In this example, I'm going to do it completely upfront. I'm specifying everything in the prompt. So it's creating a Linux task. I'm giving it the name. I'm telling it which agent to run on. I'm telling it what credentials to use, what script to run, and then in the end, I'm adding a little bit of optional, you know, what would the success criteria be of this particular task. So I can I can submit it and ask it to create that? It will again tell me what it's going to do. You can see the tools that it's actually using and the, it tells you exactly what it's doing at any any one time It's thinking. And now it will ask me, okay. I'm... I want to allow you to create that, and now I can go and view the task. So all of those things that I added in there, the name, the a... The agent, the credentials, the script that I want to run, all the way down to the success criteria that I specified. Okay. So all of that is is created. That's wonderful. And let's go and add that into my existing workflow, and I'm going to add it to run after the final task in the workflow. So I'm going to run it down at the end of the workflow here. So I'll just scroll down so we can see it. Okay. I can just answer number one. I want to add it after the yep. So it's... It... It's understood, and it's told me, what it's going to do and where it's going to add it in the workflow. And now I can click updates, and you'll see that task appear at the end of the of the workflow here. Yeah. So... Yeah. All of that is done real time. We can see it in the... Built into the workflow as we build it. So this new Robie AI functionality, you know, it it doesn't only save a lot of time for experienced users, you know, you could create a workflow and really add, you know, hundreds of existing tasks in there, but it can also make the creation of tasks and workflow definitions accessible for less experienced users, who don't have that deep product knowledge. So it's it's an extremely useful, new feature that we've added into into, eight dot one. Okay. So let's return to the slides. So that was a very quick demo. And now I'm going to run through the things that we've added, in eight dot one in addition to the Robey AI. So first of all, some security improvements. The universal controller now supports a centralized two tier envelope encryption model using key encryption keys and 256 bit AES data encryption keys to protect sensitive data stored in the controller, like credentials, OAuth, and SAML settings, etcetera. So that security level of of the data stored in the database has has significantly improved. In addition to that, we added some user session restrictions. So two additional options we added here, the idle session timeout. This is a stricter timeout than the simple inactivity timeout. It's tracking real user interaction and it's not affected by the dashboard widget updates or activity monitor refresh, etcetera. The maximum session lifetime puts the ceiling on the session lifetime regardless of the user activity. These options were requested by multiple customers to satisfy their compliance requirements, and we're very happy to deliver those in eight dot one. Next, we have some jobs as code improvements. This is a really hot topic at the moment. More customers wanting to interact with product objects in different ways. And Niels is going to talk about this a little bit more in in his session after this. So the first thing is that we now publish an event whenever a supported record is created, updated, or deleted. And you can see those supported record types here, the tasks, triggers, agents, etcetera. All of these are managed by six new system properties, that control that behavior so you can enable them as you need. You can basically switch on these events, as and as and when you're ready to do that. And, obviously, this... All of this event activity can be captured by our jobs as code, capabilities as well. On the right hand side here, we've made some important improvements as to how SIS IDs are handled on import. So now, first of all, the workflow API requests include the SIS ID and the name, in the response, and this is really to enable import to handle the situation where you are importing into another environment where the sysid's are different between the two environments. So on import, if the sysid is provided, it look... It will look up the task with the SIS ID. If the SIS ID is not provided or if it doesn't exist on the target system, it will look up by the name instead. And and therefore, it will it will join everything together that it needs to join on the, on the target. So a significant improvement, a significant add... Addition to our flexibility really about how we import objects. In Windows and Linux tasks, we had some functionality before that was only available with the use of a special extension. It's now been added as part of the core product. So there are four major things here. First of all, you can now specify a new field which is profile file where you can specify a path to a script file on the agent server that can be executed by the agent before the tasks main command. That is really enabling things like environment initialization. So a powerful addition. For standard out and standard error, we added new fields for redirection, which allow the task to write standard out and standard error to files on the agent. We also add, for the script variable support for Windows, Linux, and IBM I actually, we added a brand new checkbox and new functionality, and that means that the script field can accept a variable reference that resolves to a script name at runtime. Again, these are requests that we've had in the system for a little while and we're very glad to be able to deliver those. And last on this page, we added a new field which is the append task variables. It automatically injects task context like ops task name, ops task ID, etcetera, as operating system environment variables into the launch process. So a big step forward, I think, for for the Windows and Linux tasks. Now let's talk about what we did with the zOS, task. We added a saved JCL to override the JCL lib on restarts. So it's a brand new field, use JCL save library. It controls the JCL JCL source that is used when a task is rerun. So when this is enabled, the agent submits the JCL saved from the prior submission, from the JCL save library, basically. And, again, this is an enhancement request that was requested by a number of different customers. For the universal data mover on the right hand side, we've made some significant performance improvements with UDM to allow more concurrent file processing. And the volumes of data that customers are moving with UDM are increasing all the time. So this is this is very important to our customers. Next, let's talk about what we did with the web service API. So we've added patch capabilities. It's basically just changing the part that you need to without replacing every every field. We did this for task definitions. We did it for SAP task instances and for user and group record fields. So that's a very good addition to our web service API. We added a brand new role so you can access UC metrics without the ops admin authority. Again, a customer request. And, you can now read a workflow task instance using the web service API and return the full structure in JSON or XML format, including each vertex's task reference, status, position, etcetera. So, again, a very useful addition to our web service API. So as some general UI and UX improvements. So if we look at the left hand side column, so task monitors can now filter match task instances by their business service membership. We added six filter modes, here, so the filter is actually very granular and and extremely useful. We also now added the ability to see the number of retries left, in a in a workflow monitor. So a retry count badge appears in the workflow monitor on the task instance icon when a task is scheduled for automatic retry. If you hover over that badge, you also get a tooltip with retry details that include the next scheduled retry time. So an important addition and something that, you know, we've been requested, by a number of customers as well. And I hope you find that useful. Also, we made those conditional expressions in the workflow editor and monitor more visible. So you can now easily see if a conditional dependency exists in the workflow with a new badge in the top left corner of the workflow vertex. If we look at the middle column here, you can now select multiple values in the composite trigger to target multiple time based conditions. We also represented this in the API as well. With the forecast, you can now override variables in the forecast dialogue, and that means you can preview how a workflow schedule or a critical path or a task list would look under different variable values without having to modify the workflow itself. Really very helpful with modeling of different scenarios. And we also added a next retry time for the task instance list API. So the task instance list and list advanced APIs can now filter results to task instances that are or are not... Oops. Sorry. Are or are not scheduled for automatic retry. It can also report the next scheduled retry time for task instances in a failed status awaiting automatic retry. So last but not least, the right hand column, we added a brand new agent cluster option, which is list, and that supports a new ranked list distribution type. So it routes tasks to the member agents in the user defined priority order, and that allows you to use a a preferred agent approach and it makes the selection a lot more predictable, and that's, again, a customer request. Another one is around managing business services, and we added a brand new role so you can grant permissions to create, update, and delete business services without requiring that broader administrative role. And for the cloud data monitor, we introduced the long attribute and long array type, which allows 64 bit integer values and arrays to be carried in universal events. So a lot of improvements in this solution... In in this in this version, the eight dot one version, and, you know, we we managed to include about 30 customer requests in total in the eight dot one release. And for the last two are the observability, so we've added some more labels to the UC metrics, for observability use cases. And for the certifications, we completed those certifications for Python three dot one four and Red Hat Enterprise Linux 10. So I hope that gives you a good, indication of what we put into eight dot one. It's been a very, busy delivery for the release. We included as many enhancement requests as we could, and it's a, yeah, a good step forward in our in our eight dot one release. So back over to Gwen. Thanks, David. And I hope everyone, is interested to get their hands on the the, AI workflow builder as part of Robie AI 2.0. That's part of UAC eight dot one. So next up, we're gonna talk about universal two point... Portal 2.0. So let's, let's just kinda take a little bit of a backup through memory lane. We introduced universal portal one dot zero back in the spring of twenty twenty five. Later last... Later in the year, we released a one dot one version. Actually, that was this spring, the one dot one version that provided broader international language support, the, eight major... Seven major European languages as well as numerous enhancements. But one enhancement that we got over and over again was, you know, the the citizen automators who are the primary customer for Universal Portal really wanted to be able to show workflows. So let's move to the next slide, and I'm gonna show you a couple of views here. So this one here is a live view of the universal portal workflow that's running within the, within the task board of the universal portal. I've gone into dark mode for this particular view. Don't worry. You'll see the light mode as well. But here, you can see a workflow. You can see... Also see the children task as well. And then also, it will update from stage to stage as well when you when you move through it. You have the same right click... Right click capabilities that you have with UAC. And as you can see in the bottom right hand corner of the canvas, you'll see that there's also very familiar controls to be able to zoom in, zoom out, and then also be able to break that out into another screen. We don't just have the live running workflows if we move to the next screen. We also have the we also have the definitions as well. I think I might have got those mixed up. The prior screen was actually the definitions. This is the live view. My apologies, folks. And here you can see the the visualization of the workflow as it moves from step to step. Also, this works well with responsive design as well. It's one of the reasons that the the boxes are a little bit more square. So if you're trying to get it on an narrower screen like an iPad or on a phone, you can do that as well. However, there's the same challenges that you will have when you have a large number of objects on a very small device. There's only... You know, we... We've tried to optimize it as much as possible, but, you know, there's still, of course, going to be limitations. One of the goals, of course, with a lot of the citizen automators is if they're running something like a, like an SAP process or if they're running a data pipeline, sometimes those don't tend to be the largest workflows to begin with. So here we here we are. We we will be releasing this this fall. Don't ask me for... And pin down an exact date. We're working on that, but it's about ready to get handed over to... For testing, and we're pretty excited about delivering this functionality into the universal portal. And, of course, if there's any more questions on that, feel free to, ping me through your, your, your Stone branch representative, and we can set up a time to talk about this more as well. We're... We always love to get your feedback on people who are interested in being early adopters. Okay. So I think it's now time to move to the next slide, and it's time to hand this over to Nils and Lens. And they're gonna go through what's new with extensions and integrations. And, it's it's kind of remarkable how we we continue to speed up and build into more and more, platforms, especially around the world of AI. So take it away, you two. So I'll hand over these slides to you, Niels. Here we go. Good. Thanks a lot, Gwen. Thanks a lot, David. Yeah. I'd like to show you some of the integrations we've been working on. I will start with Oracle and SAP and then move to data pipelines, monitoring, and AI. And I've also switched to Universal Automation Center a few times so that you can see the tools in in practice. So let's start with Oracle Fusion. We have added integration for ESS, Oracle Integration Cloud, and Oracle Analytics Publisher, also known as BIP by Oracle people. And HCM flows, the dedicated Oracle objects task are already in testing. When I mean testing, it's... They are... Basically, we already executed them successfully, and they're they're ready to be tried out. We also started with the implementation of a new Oracle o d... ODI integration, so that's something which we have just started. With ESS, you get ready to use processes that follow the Oracle ESS interface. We cover applications such as payables, general ledger, HCM, and accounts receivable. For Oracle OIC or Oracle Cloud Integration, it's called, we support now asynchronous execution with parent and child. The parent task only shows success when all the child tasks have finished successfully. Also, we included paging now to support more than 500 trials. So we had this... We experienced one of the customers that more than 500 trials, and then you need to use paging. But this is something which we have also implemented now. Let me quickly switch to the universal controller and show you the ESS task. So what you see here is the ESS task. So, basically, you can con... You can connect to ESS with the standard authentication mechanism, o out or basic. I guess most of the customers would use o out, then you can have the different grant types, client credentials, or passwords forward to. And the and the interesting thing is now that the process that you want to schedule, you can select here under job import. So you press this dynamic choice button, and then you select the process that you want to schedule. For example, I don't know how to post journals, credit accounting, or you can use import costs as you see on the screen. Altogether, we have now 125 processes implemented, but this number will grow, and customers can also add their own processes, which... If they are missing in this list. Then you have here the parameters for the processes, and the interesting thing is that the parameters should have the same name as the related parameters in the ESS UI. So when you're in the ESS UI and you know how to configure that, you import cost task, then you will also be able to configure it here in in the screen. If a parameter is not mentioned here, you can just edit here by clicking here and then adding the parameters. So this is the concept of ESS task. So the advantage compared to, let's say, other... Let's say, the most generic task is that we are not just passing the rest API call. We also preconfiguring the processes. Let me switch back to the presentation. Now let's move on to SAP. This year, we focused on SAP cloud environment, including RISE private and public cloud. SAP Business Data Cloud or VDC uses a zero data copy approach. VDC Connect led platforms such as Snowflakes connect with shared HANA Cloud data lake this way. You'll see supports now out of the box automation for all VDC Connect applications mentioned at the ZAP file 2,006, which I attended, including MS fabrics, Databricks, and Snowflakes. Also at the DSAG, they did not mention more of these applications, so we support all the DC Connect applications. To support b c zero data copy concept, we have edited now a new task for the SAP cloud data lake transfer task. It supports both the data lake relational engine and also the upload and download files. So you can, for example, check the status of a data product if it's ready or not. To support... You can also run SAP application jobs based on templates, including SAP IBP or s four HANA application jobs, for example, on Rise public cloud. And we have now... This all put into one single task. So in the past, we had two tasks. Now we have just one task to run all the SAP application jobs, and we even get the log files for the s four application jobs. We have also a new integration for SAP build process automation to start and monitor flows, and you can use now our certified ServiceNow app, which is available in the ServiceNow store to schedule your SAP jobs. Okay. For data pipelines, you can run Microsoft fabrics... Pipelines from USC. And important thing, what's new is that you can use now managed identity for most of these products, like MS fabric in this case, so you don't need any password for the connection anymore. You can also run now NNN workflows as part of your larger business processes, so you can bring these steps into the same UFC workflow as your other tasks. For example, you can run an ERP job and a data pipeline as part of the same process, including, for example, an NNN process. Monitoring. For monitoring, we have now dashboards preconfigured for Dynatrace, Grafana, and Datadog. So all the three major product for observability. And we have certified integrations already for Dynatrace and Datadog where we also partner. This helps you keep an eye on your automation related application. Let me just show you the Dynatrace dashboard, the default one which we have. So this is close to the Dyna... Example Dynatrace dashboard, which you can import in in a less than a minute, then you have it. And, of course, you can then adjust it to your lease. You can monitor many many controllers. So if you... In this case, I have just one controller that I'm monitoring, but you can also monitor several controllers within one DynaTrace dashboard. It shows you the different database connections, OMS connection, the the tasks which are the most used. And it also monitors the AI services. I mean, you've just seen from David, our workflow builder, so you can see how many tokens have been used and also how many MCP server connections you have. At the moment, I have just connected to one MCP server. Of course, if you wanna do observability, you're also interested in trace and locks, so you can just switch to the locks in this case, and then choose the lock source. In this case, I choose a lock source, for example, the universal controller lock. I can also use container locks and so on, then just run the query. And then you would see here the different lock files from the universal controller just inside your observability dashboard. Okay. That was just a quick quick view on observability. Kubernetes... When we talk about Kubernetes, we have seen more and more customers now run Kubernetes jobs through USC. When we first started, it was mainly customers using OpenShift. But now this has... We are seeing a shift. So people are using also a lot other platforms like Azure Kubernetes services, EKS, and even native Kubernetes. And if you wanna deploy your agent and controller with Helm, you can just ask just ask us to write us a mail, and then we can also provide you the Helm charts to deploy your controller and to request agents. Jobs as code. So you already saw that we have this new event concept. So we are planning to release a new jobs as code task soon. When you change a configuration, you will see it creates now a so called record change universal event. This event then can trigger an update in your local or remote Git repository, and, basically, without doing anything, your Git repository is always in sync, fully automatic. Let me show you a very quick demo on this so that you understand the concept. So what I have here, I have here this new HANA data lake task, which I just talked about. And let's say I give it a new name, call it maybe b three, version three, and I say upload data from my home directory to the HANA data lake and upload, for example, these reports to it. And then I save it. So now I made a change on the name, and I made a change also on the configuration. What you see here, immediately, two times our jobs as code task has been started because it was triggered by the universal event. How is this triggered? We configured just a simple webhook, and this webhook says, each time a ULC record change happens, start the jobs as code task to sync the records with your Git Git repository. If I look at the task, what what it does, it basically goes to this develop... Git repository branch to this repository path, and there it saves the changes that you... That he that he has discovered. Actually, there were two records generated. Why two records? Let me look at the first record. The first record says, there was a new task, HANA Data Lake Cloud version three. That's the task that we just changed with the name and these things. That's one. And there was another event generated because this task is part of a workflow. So it also update... Also created an event saying, ah, the workflow was also updated because in the workflow, this task has a new name. So so what I wanted to show you here is that it also automatically also updates every related events. So when I now look into my GitHub and then press control r, press r, you see there is the new... And and then I go to the integration. So they see there's new SAP HANA data lake integration, and you see here... Maybe, like, bigger. Sorry. You see that's really... And you see version three. Yeah. So it would... So in summary, you don't have to do anything. It will completely automatically update your GitHub when you make any changes on the controller. And one additional thing which I'm not... We cannot show you in this demo because it takes too long. Instead of directly pushing it to GitHub, you can also first submit everything to a local to a local repository and then define... And then decide when you want to push it to the remote repository. And, of course, the other way around works also. If you do a change here, then you can import the change automatically to the universal controller. Okay. Let's finish with AI. So we are adding more ways to use AI and UC workflows with our a... AI LLM and Kong AI gateway task. You can make an I... AI action one step in a larger process. You can also start and manage AI pipelines in other platforms, including Google AI agent pipelines and many more. For SAP JUUL, the plan is to let you start an agent from a USC workflow. That will depend on the JUUL two dot zero agent to agent gateway, which will be officially available, I guess, by end of this year. However, the task we have already we have already a prototype of the task. We're just waiting for SAP to give us access to the agent to agent gateway, and then you can run your SAP dual agent from a workflow in Universal Controller. And now finally, and if the integration you need isn't available yet, we are working on that too. The AI integration build up prototype is designed to help you build an integration without writing any code. Let me hand over to Len, which will show you the latest update on the AI integration builder so that you can build your integrations without any coding. Len, over to you. Thank you, Niels. So lastly, in our integration spotlight, we want to get focus on the new AI integration builder. For everyone who missed it, the AI integration that allows UAC users to create new universal extensions directly inside of UAC. You're only using a requirement file, and there's no coding involved from you. The users have a vision of a new integration, and they start a new integration builder that fulfills this version by using its agent skills. If we head over to the next slide, we also made an enhancement already to this integration builder in order to make it more accessible to all our users. At the beginning, we only were able to use Claude and Claude codes with its Claude skills in order to generate integrations, but we have extended this with a translation layer on top of the agent skills. Users can now use any LLM they want and configure it inside of the task, meaning you can use Cloud, OpenAI, your local Olama instance, or an AI gateway like Kong AI, for example, or any open source gateway that is able to access to tool calling and has sufficient context. I would like I would like to take over and then show you a finished example inside of the controller. So inside of the controller, I would just quickly show you already generated integration, but how the process would go through. For example, I wanted to create a new GitHub's action workflow integration, which launches GitHub actions automatically. For this, I have created a requirements YAML or requirements MD rather that lists the structure, the dependencies, and some basic, like, authentication and that, for example, in the list workflows and run workflows, I want to have a dynamic choice field so I can list the workflows inside the task itself. I also have given it specific test instructions because I have a a GitHub repository already with a workflow, and it can... The AI integration builder can run tests on it already. So I give the instructions for this in the in the test instructions. I also give the git credentials that is needed for it, which I can query directly from the controller so you don't need to type it in if it already exists. And lastly, I also configured my LLM provider to use GPT five dot six Solve as it's one of the fastest models for coding. Once I have launched the task, I can open the instance of the already finished task. And if we head down, we can see that it has already finished within thirteen minutes. The complete integration has been generated. It's... Also, if I head over to the output on the logs tab and I scroll through the STD out, I can see... If I make this a bit bigger, a a table of the token usage needed. For example, this whole integration with testing and documentation has taken me 1,600,000 tokens, which for GPT $5.06 prices would be around $7.8 even as an estimate, for example. It has already up... Loaded the integration, and it has already tested it based on my testing attempts. If I head over to activity tab, you... We can also see the five tests that did. It changed. For example, we did a test for the dynamic choice field as we wanted to have a dynamic choice field. If I open up the task instance and scroll down to the GitHub action details, we see everything that I structured out has already been generated, and we already have the dynamic choice field, which also works if I test it out of the button. We can also see in the output, which has been nicely formatted. And even inside of the GitHub's... GitHub integration task itself, If we go to the dynamic choice instance and go to the instance timer, we also see that it does custom output fields already. For example, the run ID of the of the workflow that I did and also which status is that. This just shows you how easy and fast it is to generate integrations by yourself using the AI integration builder. And I I want to remind everybody that the... We are doing an early adopters program for the AI integration builder. So if you're interested, especially now that we can support any LLM, I would highly recommend to... That you check it out. And with this, I will now head it back... Give it back to Gwen. Okay. Thanks, Lynn, and thanks, Nils, as well. As you can see, you know, we continue to be busy, which is why neither of them look like they've really got much of a suntan over the summer as well. So, you know, I think this also continues to show that we we continue to pick up momentum, doing things to the product that everyone can find value with, and then also, you know, continuing to push forward into the world of AI. This extension builder thing is purely awesome. You know, five years ago, this would have almost been considered like science fiction that you'd be able to do that, or would have required teams of people years to... Years and years to develop. So, you know, I think the work that we're seeing is great. We are going to be running a controlled early access program where we'll start off with a few people and then widen and widen it, from there as well until we get to a point where we feel really comfortable in releasing this. But this is really going to help you expand connecting UAC into new platforms, generating new ideas of where you can take your automation, you know, even further within your organization. So we're gonna wrap up today talking about UDMG three dot four. So for those of you not familiar with UDMG, you know, we've talked about it on the last few what's new sessions. This is universal data mover gateway. This is our b to b managed file transfer solution that you would use for your trading partners to exchange files with. It's a very modern solution. It's integrated within UAC, and it has a lot of capabilities that, that you don't see in a lot of the, legacy or traditional, managed file transfer solutions that are currently on the market. So let's move forward and kinda go through some of the feature highlights. So in this release, which, was released just a few days after UAC eight dot one, as we're trying to bring the cadences closer together, we've touched on pretty much every major area of, of of, what you would require for a a managed file transfer solution, whether it's in the administration section, the partner management, you know, security, of course, deployment, and transfers. So we... You know, I think we can just pick a few things out of here. I'll have some screenshots as well. You know, adding the client CLI as well as the API, we've added that. You know, we've added container support for deployment, which is a big which is a big one that was asked for by many customers. The Postgres support, of course, as well for the, for the secure proxy. We have a very, very solid secure proxy that as, you know, forward proxy, reverse proxy, and integrates also with ICAP, ICAP s, providers as well. If we move forward, I'll show you just a couple of quick screenshots. One of the main features we added was the configuration, versioning, and differences. Right? So this is just kind of a shot. One of the challenging things with, with any managed file transfer solution is if something fails, you know, and it looks like it's failed on your end, you know, it's pretty... You know, one of the first things to do when you're going through the diagnosis process is, you know, is what changed. Right? So this will help you really highlight, you know, what's changed, between different versions of, of UDMG. Now let's move forward to another screen, and, this is a server health status monitoring. So this is really designed to, at a glance, you know, tell you, you know, how your... You know, where you are in your license. Right? Your endpoints, your certificates, your SSOs. Right? And to make sure that every single, main component of UDMG is healthy. And then finally, one more, screen from there is, is the forced password reset, which gives you a really nice visual of, of getting a password reset. This is not only for UDMG, but also for its partners through the web web transfer client as well. Now we haven't talked about much around AI and UDMG yet, but one of the reasons we built Robey as a separate service was to be able to connect UDMG UDMG also to it. And I'm not gonna spoil the fun of our next what's new, but you'll also see that that we'll be adding in some AI capabilities into UDMG as well in the near, near future. Okay. So I think that's all I have in terms of, UDMG for you. Of course, if you're always interested in finding out more about UDMG, you know, just, just reach out to your, Stone Branch contact, and we'll get the right people to talk to you and, talk about your landscape and how we can help, support any transition from... To a, to UDMG as well. Okay. Let's move forward here. And just kinda one last thing, before we get to question and answers, we have one more user recession. So we had two great sessions last week. We had, this, session, you know, automation in the age of AI, right, which was kind of our keynote. We also talked about the hybrid orchestration control plane. We had this session today, which was our third session, And we have a session coming up on, on Thursday, which is entirely samples and examples, right, use cases, seeing how we use AI in production. So you're gonna get... You're gonna see much more hands on, you know, very few slides, lots of lots of, you know, live demos of of AI in action as well. So I'd really encourage you if you're if you're you're... You really wanna find out more about this, you know, how we're leveraging AI within UAC and you want to see some examples, you know, to, to dial in, either live or through the recording and, and and, get up to speed that way. Okay. So I think it's now time for some questions and answers. Lauren. Alright. So let's go ahead with the first question. Will the audit entries include the IDs of the person using Rovy to build the workflows? I think that's a perfect David question. It it is. Yes. Absolutely. It uses the... It it includes the ID of the person using Roby, so everything's done within the context of the user that's logged in. It also contains that it, you know, came from a web service call because that's what Roby uses to update, and it also contains a reference to the fact it was AI, Roby AI that that made the change and even the version of Robie AI that did it. Right? So, yeah, everything that you need in there. Alright. Can you remove an existing task from a workflow as well? Okay. Another one for me. Not yet is the answer. It... It's just a matter of, you know, what we prioritize and what what we... What order we, we add these tools in. You can you can add an existing task. You can't remove one yet, but, you know, you could, just with natural language, ask Roby to give you a a link to open the particular workflow in the workflow editor if you don't know how to do that through the UI. Right? So it's pretty... It will still help you to get to the point where you can, you know, delete that task from the workflow. Great. Will there be an option to migrate or switch existing enhanced Linux Windows task to a default standard out or standard error profile setting in UAC environment? This is a great question, actually. The, we wanted to add the functionality first and then think about how we're how we're going to get you from one state to the other. I'm I'm pleased to hear you have some enthusiasm to adopt the the new functionality in the in the standard task, but we'll certainly be looking into that, yeah, over the next months. Once Roby created plan, can we ask to change the plan or expand interactively before final create? So let me think about that question. Once Roby created the plan, and so we're talking about a workflow here, I think, with the plan. So, yeah. I I mean, absolutely. In the conversational chat, you can you can be adding and and removing tasks definitions from the list before you actually say, you know, create. It's it's a complete, you know, conversational chat going on, and you're you're right in the context so it knows what you've done before. So... Yeah. I I believe we can do already what what you're asking to do here. But yet, always right at the end is your commitment to creating it. But first of all, shows you everything that it's going to do before it builds it. Okay. Well, that's all the time we have for questions. Thank you, Gwen, David, Nils, and Lynn for walking us through all the exciting new features and enhancements in uac8.one. And a big thank you to everyone for joining us today. We hope you enjoyed the third session of StoneBranch UserVerse online 2026. Join us on Thursday, October 22 for our final session from prompt to production, hands on AI workflow training with Stonebranch, where we'll demonstrate how AI workflow builder, AI integration builder, and roby dot ai help accelerate automation development, and we hope to see you there. You'll also receive an email in the next day or so with a link to today's recording where you can rewatch or share with your colleagues. Thank you again for being a part of UserVerse online twenty twenty six. We appreciate you joining today, and we look forward to seeing you on Thursday. Bye bye. Bye.