Hello, everyone, and welcome to today's webinar on UAC eight dot o product update, the future of automation. Discover what's new in Stonebranch Universal Automation Center eight dot o with a focus on Roby dot ai, AI powered intelligence that reimagines IT automation for intelligent orchestration. I'm Lauren Tanzini from Stonebranch, and I'll be your moderator today. Before we begin, let's cover a few things. This event is designed to be interactive between you and the speaker. So if you have any questions, please add them into the q and a tab on your dashboard, and we will answer them at the end of the session. If we miss you if we miss your question, we will follow-up with you afterwards. Additionally, in the right hand corner of your screen, you'll find a handouts tab with lots of resources related to the session today. With all that said, it's my pleasure to introduce our Stonebridge speakers for today's session. Gwen Clay, chief product officer, Robert Olson, principal product manager, and David Ainsworth, principal product manager. Let's go ahead and get started. Gwen, take it away. Well, thank you very much. So we're gonna we're so I think as Lauren mentioned, we are gonna be talking about, one of the, most significant, releases in UAC history with eight dot zero and Robie AI. So if we move forward and start looking at Robie, Robie AI is a combination of several years work into our research into how AI can enrich the lives of our users and and create better outcomes for automating and responding to, the challenges of of automation with AI. There's been a if you kinda look at the evolutionary history of UAC, there's probably hasn't been a more significant jump in the past ten years of what we've done with Robie AI. So why do we call it Robie, first of all? Well, Robie is named after our cute little mascot. You'll see right in the center there of that shield. And he's been a constant companion on the Stone Branch website for years. So we started this project about two years ago, and we actually did do some demonstrations of this on in the user verse at the end of twenty twenty four. However, we've really gone back and done a lot of research, and there's also been some very significant changes into some of the core technologies used to build the that that make up the building blocks of AI for enterprise applications. New standards have come out like, the MCP standard and also new tools that we can use for building AI solutions as well. So when we started looking at building Robie AI, we first of all took a look at making sure we had a future proof and a very modular design. With a lot of the AI standards evolving quickly, we wanted to make sure that we have a platform that ensured that it was going to be, able to be adopted, well into the future. And, also, we also felt that we owed it to our customers to make the design very transparent. So, you know, we're not only providing this as simply like an on premise solution as a SaaS solution, we're providing it as an on premise solution. So we really wanted to have that design transparency so that it was very visible what was behind the features. Right? And then also how they can extend it themselves in the future. The next principle we really looked at was we're making the customers in control. Right? So the AI is optional. It's an optional service that is installed alongside your universal automation center. We're not gonna put a good we're not gonna force any customers to use it or even activate it for that or install it for that matter. It's also private and secure, so no customer data is shared. Every customer has their own private instance. And it's also flexible in terms of how you can, choose which ALM you can use and then also, how you leverage the AI capabilities outside of UAC. And then, of course, what's really important is the human in the loop is, you know, we repeatedly repeatedly also requested that everything we do, that there's a manual step before moving forward. Also, what came up was really doing things that were beyond chat. Chat has some great functionality, and we have built in some fantastic functionality with the Roby chat capabilities. But it's only really one of several core capabilities that we're going to introduce you today to today. And, also, AI workflows and integrations is a key capability moving forward. Right? So this is really where you'd be able to have steps within your workflow where you'd be able to connect to the world's leading, AI services and to be able to connect to them and then, you know, interact with the AI and retrieve information which you would use to enrich your workflow. So say, for instance, if you're mat if you're mixing, commission data with sentiment analysis, Right? That would be a really great great example there as well. And then as we move forward and look at it, was in terms of the private and security aspects. Right? So Robie AI has a one to one relationship with the UAC and the AI server. Again, we're not doing any model training on the customer data. The customer data is yours. The LLM is yours. So if there's any training being done, it's by your LLM on your data. Right? We're providing the flexibility for being online or offline so you could use one of the leading SaaS providers, such as Microsoft Azure OpenAI or Amazon Bedrock or Google Vertex, right, or Claude. Or you could also use your own self hosted model. Again, we don't collect data, and you're in full deployment full full deployment control of how you want to consume Rovi. And then if we start taking a look at how we've designed Rovi, this diagram, you know, isn't going to be that unique to people if they read a lot of AI architectures, and that's because we really wanted to build upon standards. So we have a separate AI service that sits alongside of UAC. We have the orchestration layer of Roby that we have designed ourselves. We have the AI services, which are the LLMs and the vector database, all of the standard aspects in terms of AI tools, rag memory, also the agentic aspect where you can build it where AI agents can be built. We'll eventually open that up to enabling our partners and customers to build AI agents as well. And then what's very important now is the MCP's client and server. Right? So this would enable you to connect to Robie AI and automate workflows, query statuses, from other AI platforms, whether it's an LLM itself or whether it's an LLM that's embedded in an enterprise platform such as ServiceNow or SAP. Also, in our in our ability to call other AI AI MCP servers through the MCP client as well. And, not only are we introducing the AI the AI capabilities and the LLM task, but also how we develop AI integrations itself. We have a new extension builder that's AI powered that's really taken our time from building extensions down from days to to hours using our new, UAC extension builder. And that we will also be rolling out to partners and customers as well. And we're already demoing it to select customers. And then moving forward, the new LLM task tasks. So now within workflows, you can have LLM tasks embedded within your work flow. So instance for instance here, we have several Azure and an Amazon Bedrock task. Within those task definitions, you have all the connection information you need to connect to the LLM. You're able to select the model that you wish to, to use, and then you're able to set up a system prompt and then, and then also have the chat prompt. And those can either be embedded within the task or they can be embedded in your script library and then reused for multiple tasks as well, which is extremely convenient. And I think this is the point where I hand it over to David, and he's gonna walk you through the the core capabilities of Robie AI. And we have them organized in kind of three ways. Right? We have the, the build, run, and integrate. And, David, over to you. Okay. Thank you very much, Gwen. Hello, everybody. I'm David. I'm going to start today, as Gwen said, with some of those examples of Robie AI in practice. And then later in the session, I will highlight some of the other new features we delivered, in this release, including twenty five and more, customer requested enhancements as well. So these Robie AI use cases are real use cases that are designed to get you thinking about the huge benefits it could bring in your environment. We are all used to having access to immediate assistance when we have a question or a problem. I've noticed myself looking for the place where I can ask a natural language question, and I very rarely think of the documentation as as my first option when I need to find something out. Right? So assistance like Copilot have really dragged AI usage onto the desktop, into the mainstream, and it has become the first logical place that we look for assistance without necessarily thinking of it as consuming AI. So Robie AI can certainly change the way that you interact with UAC. And as Gwen said, we look at this in three contexts. We have the build, the run, and the integrate. I'm going to start with integrate as this is the foundation of our capabilities. And the core of that is the MCP server, as Gwen mentioned. This is the model context protocol. It's it has all the tools and all the capabilities that enable us to interact with UAC, with other tools, and with AI endpoints as well. And as in here as well, we have the AI extension builder which will allow extensions to be built quickly and easily and it will be available to customers in the near future. But I'm not going to focus too much on the integration in this session. It's important to know that our AI is built on the industry standard protocol, so it's ready for now and it's ready for the future. And let's move on to the run context. This is really going to be the cornerstone of how we interact with the product. And I'm going to walk through some examples to introduce you to these capabilities. The most important point is that all of this is optional. Right? You choose when you're ready in your organization to take advantage of Robie AI. You point it to your LLM and your documentation in the vector database and then enable the option in the administration section. And only then will the new Roby AI assistant icon appear in your menu at the top right of the screen. So it's it's important to know, you know, it's your choice. You make that decision when you're ready. So the first example I'm going to use today is to launch a task. And this is really going to demonstrate, you know, interaction with UAC and the human in the loop checks that Robey includes as well. So hopefully, screenshots are not too small for you to, to read. I'm going to help by really reading out the, the different prompts that I'm going to use as well today to make it nice and easy. And as I open the, Robie AI assistant chat window, it is important to know that everything that I do in this chat is in the context of my logged in user, including any permissions that I have. So if I don't have permissions to it, I won't be able to do it in Robie AI either. Right? So I'm going to launch this task. The the prompt I'm putting in is launch task p t b s cleanup daily. And the first thing that Roby will do is ask me, you know, why I want to launch it for audit purposes. So at the top of the screen, what is the reason for launching it? I'm interacting and saying it's an ad hoc request by a specific user. And then Roby identifies that there are variables associated with this task and ask me if I want to continue with those defaults that are already defined or if I want to update them. And at the bottom of that screenshot, I can use natural language to update one of the variables. I can add another variable to the list, and that's all in in natural language. And then Roby will ask me to confirm all the updated variables that it will use to launch the task and offer to launch it for me as well. And once it's done that, it will give me a button to directly view the task instance. So this is a very simple example, but it's all nice and friendly and guiding me and asking me for approval all the way. So next, I'm going to check the status of some tasks as well. You know, so we've got the launch. We've we've got a number of tools that we defined in the MCP that are ready for the eight dot zero launch, and I'll talk about some some of the ones that are coming next. So the next one is to check the status of some tasks. I could obviously ask about the task that I just launched, or I can ask something more generic like I'm doing in this case using just a natural language query. And that is show me all tasks with a prefix of EOD in a problem state for the last twelve hours. So Roby knows what we mean by problem state, and it gives me a list of all the task instances matching the criteria and a button to directly access each one. So this is already useful. It's it's given me all the information that I asked for, but what if I want to use the results of this search for a specific purpose? For instance, I want to reformat this. I want it to be in in CSV format for an Excel import. So I just type that into the Robey chat and the output is reformatted and I have a copy button so that I can just copy and paste the content into Excel. This is not just accessing information from UAC. It's also using the power of the LLM to quickly give me the information in the format that I need. And, you know, it just saves me a lot of time and effort. So just imagine the steps that you go through to achieve the same, you know, same output now. You know, there's a few steps that you would have to achieve and then and then do a bit of reformatting yourself. Okay. So that's that's two good examples of of the conversational chat. The next one, we're going to have a quick look at the root cause analysis capabilities. For this example, I've I have a task that failed in the activity window. I could go the normal route. I could look at the output. I could diagnose the problem myself, or I can now right click and select the new option from the context menu which is analyze now. And the first thing Roby will do is to determine the error and present me with the root cause of the problem. So it's already worked out that it's it needs to be more time zone aware in the in the script and it's picked that up from the, from the output of the job. Right? Next, it will start to make some recommendations. And as it's executing a script in the script library, it will offer to analyze the script in more detail. Then it will present me with a fix for the problem, tell me what it's going to do. And last of all, it will show me the part of the script that it's going to change, and it will tell me what change it's about to make and offer to update the script and rerun the failed task instance for me. Right? So every step of the way, this is interacting with the with the user on the chat and helping to fix the issue and to reach that successful conclusion. In that example, I analyzed an issue with one failed task, but I can analyze at a workflow level and it will find any failed tasks, right down into all the sub workflows and it will give me an analyze button for every one of those failed tasks down through the workflow. It will offer to fix them, it will offer to rerun the failed task, and and another interesting thing, you know, I I'm still interacting here with with Robie AI. So if I need to get somebody else from my organization to look at a problem case, I can ask Robie AI to create an action URL for that workflow, and I can share that URL with the person in my organization and they will be taken straight to the context of that failed workflow. So it could really save a whole heap of time and have everything running successfully again very quickly. Right? So as I said at the start, you know, these are are really quick examples of of, some of the functionality. Certainly not an an exhaustive list of the, of the conversational chat or the root cause analysis functionality. Is really to, you know, peak your interest and and sow some seeds so that you can start thinking about it. So the third context we're going to look at is the build. Right? And this is where we can get help accessing UAC documentation, accessing the LLM knowledge to give me all the help that I could need all from one place. And I'm going to use one of my favorite examples, and I'll read it out just in case it's too small. Show me how to create a workflow called EOD master eighty eight containing three existing tasks, prefixed EOD that run one after the other. So I I really like this example because it gives me step by step instructions on how to build a workflow including all the context that I added to the question, like, you know, the specific tasks that I want to pull out of of UEC, how they're going to run, you know, all of that information that I've given it. It gives me a link to the documentation if I want to dive into some more detail, but it's really given me a step by step approach. And the reason I really like this is because we haven't added the workflow building tool yet to the MCP. That's coming next. It's one of our next priorities, but it's already really giving me helpful instructions. Right? When we add the workflow build tool, a button will appear at the bottom of the screen here saying, would you like me to build it for you? And then I can just click on the button and start building the workflow. But it it means it's already extremely helpful without a specific tool. And as we add more and more tools over the next releases, that usefulness will just increase. Right? And and I'm I'm getting all the advantages of it right now from the documentation perspective and as the tooling gets added, I will get more and more help as we go. So the next example is really just accessing something that we don't access every day. Right? And that is a documentation help. Give me an example for offsetting the date by one year with a function. And rather than searching through all the available functions that exist in the doc, I can ask that natural language question and it will find the appropriate function and it will show me how to use it. So it's given me the example, it's given me the link to the exact place in the documentation where I need to go, but these examples like functions or using regular expressions, things like that, they're not things that we do every day. So we probably don't have an instant recall of how to do them. And that that's incredibly handy and incredibly useful to give us that information. Another similar example. Again, this is quite small so I will read it out. One of the things we don't do every day is building custom days for our calendar. Right? It's probably something you do once a year or or on an ad hoc request. So this is a prompt of show me how to create custom days via the REST API for all public holidays in England for twenty twenty six and create the JSON for each day. So I really like this example because it's combining several queries. First of all, it's find finding all the public holidays in England. That's really using the LLM. Find out what the rest API call is to insert custom days. That's looking in the documentation. What is the format of the JSON to define a custom day? You know, I'm creating that JSON for me, and it gives me all of that knowledge plus a JSON for each day that it has created that has a copy button that I can just cut and paste and it tells me exactly the API call that I need to make, as well as identifying all the all the individual days that I need to add. So I really like this example because it's using so many different aspects of of Robie's Robie AI's capabilities. So what if I want to add a new task and I want to perform a task that requires a script to be created? So this prompt is a good example of that, and this prompt is create a python script that converts date formats from mm d d y y to d d m m y y. And it this is a, you know, this is a simple prompt, but Robie AI is generating the Python script for me. It's a it's a natural language query from me. The LLM has the capability to build this script and then I can copy and paste that into the script library to run. Right. And the really the really interesting thing is, you know, when I when I go to run this, if I have any problems with it, I can then do a root cause analysis and it will help me to fix it and it will actually fix it for me in this in the script library as well. So, you know, when you start building these kind of scripts and these kind of functions, it can be incredibly useful. So that is a really quick walk through some example use cases. As I said, it's certainly not an exhaustive list. It really just just there to get you thinking about the possibilities. But I also want to talk today a little bit about, you know, what else we did in in eight dot zero and what else we delivered. And I mentioned earlier, we actually delivered about twenty five high priority enhancements from our customers around the world, and I'm going to walk through some of the, some of the highlights. And this is really just to make you aware. It's just to, you know, in case you've missed it from the release notes or from any of our marketing. The first one is for MS Graph support for outbound email connections. You can now choose your authentication method and the clients that you want to specify. This is a really high priority item. Customers were telling us that they're not allowed to use SMTP anymore in their environment, and the inbound support will follow for our email monitor, etcetera. But this at this moment is for outbound use only. Excuse me. So the next one on the list is diff functionality that we added between different versions of objects. And there are basically two ways to use this diff functionality. You can select two versions, from the version list and compare those two versions, or you can select one version and then compare it with the current version that is, active. It will highlight where all the changes are and yeah, it's a really quick way to find out what changed between those versions. This is really a a an enhancement that, you know, has been requested for for the last few years. And it certainly adds adds great functionality and great usability. This next one, so you you probably aware that we added the approval task in seven dot nine. You know, this gives you the ability to, define an approval right in the middle of your your workflow and have multiple approvers and really have that kind of business level, approval feature. We've enhanced it in eight dot zero. You can now now add more conditions on the workflow connector so that a task can be executed when the approval task goes into approve approval required status. So why is this important? Well, it's really useful for integrating with the ticketing or a messaging system to add other ways to manage the approval. And you know, it can it can send off another another action and and start that whole notification or or approval process while you're while you're actually waiting on that approval in the box. The other thing we added was an operational memo so that you can record your reasons why you approved or rejected, and obviously that's for audit purposes for the approval task as well. Again, something that was requested by by customers on on seven dot nine. This is a new option that we added, because date functions have always been resolved based on the default, server time zone. So what whatever the default time zone on the server, that was being used for the date function resolution. You can now set a preference, to use the, the time zone of the of the task itself instead. So that means, you know, if you do if you are expanding and becoming more more of a global organization and running work workflows for different customers across the globe, you know, this is really important because it can it can represent their own geography. The next one is that we add in a number of web service API enhancements. So this there's a few, different enhancements in here that you can you can obviously read, but, you know, it includes how you can interact with reports using the API and selecting steps for rerun on ZOS. Again, something that's been, been asked for, for for the last couple of, yeah, probably the last year, I think, because it was available in the API in in the UI. And we added an API for interacting with promotion schedule and reschedule capabilities, and we added several enhancements to the task instance, trigger APIs, etcetera as well. Obviously, more information in this swagger doc, about the enhancements that we made here. And this this particular one is about a security enhancements really to make security that bit more granular. So we added the possibility to restrict agent registration with an OMS only if it is a member of the same, business service. So it really allows you to tighten up your, your security and into exactly, the agents that you you recognize and you want to be able to connect to your OMS. The next one is about ZOS specifically. We've added a a number of ZOS specific enhancements. So support for the TLS one dot three sessions, we added some filtering capabilities as well, for the, agent file monitor, by job name specifically. We also added support for the SNC connection, the secure network connection, to SAP from the USAP on ZOS. So this SNC connection existed already for Windows and and Linux for the use app on those platforms, but we added the capability on z o s as well. And last but not least, we added the z o s three dot two certification. That required some tweaks actually due to the default ciphers changing in the system SSL for three dot two of z o s. But, yeah, very important now. Everything is certified and working with eight zero and seven dot nine. On the, IBM I, platform, we also added some enhancements and these again were were driven by by customer requests. So adding job log retrieval capabilities, convert to PDF for email notifications and supporting variables in the job name, which which obviously we didn't we didn't support before. And I think the last one here is the OMS server. You might remember that we announced in seven dot nine, in the what's new, the OMS capability to operate in a in an h a environment without the need for a network file share for the messaging database. So this was delivered between releases, so we wanted to make sure that you're aware of this change. I'm very sorry. So I'm gonna hand over to Lauren, actually, because we have a poll now, and I'll try and get my voice back in the meantime. Apologies for that. Over to Lauren. Alright. Here is our poll. Which AI feature are you most looking forward to using? Analyze now, RCA, Robie AI chat, AI task workflows, or all of the above? Okay. It looks like fifty seven percent are excited for everything. The next highest are RCA and task workflows. Yes. I'll move it forward, hand over to Robert. Thank you, everybody. Thanks, David. Yeah. I'm going to be talking about what's new in UDMG three dot three, which is going to be released here in a couple days, actually. So end of business Thursday. If all goes well go to the next slide, David. First off, thanks for sharing all the what's the new capabilities in UAC eight point o. The AI AI capabilities are so amazing in my opinion, and I can't wait to, we can also incorporate, some of that goodness into the universal data mover gateway, capabilities as well to to further augment our overall kind of AI strategy. But as we as we all are aware, the movement of data itself is is pretty critical to workload automation, and that's why our, file transfer is is a core component of our universal automation, center strategy. Most of you on this call are probably well familiar with our universal data mover, both the protocol itself or just the capabilities for moving files within UAC workflows using the various protocols, like SFTP and FTP, the UDM protocol, or even the, intercloud, data transfer protocol for moving, from and to private clouds. The Universal Data Mover Gateway or UDMG, is a more recent addition that provides a a means of facilitating, file transfers between your organization and your external partners. So more, leaning towards a b to b exchange of data that you would then receive and integrate into your your back end workflows. If we can go to the next slide. This one kind of helps illustrate a little bit how how that movement of data can occur, what's called the life cycle of data between business to business partners, where the UDMG component of our UAC portfolio, is a separate application, and it typically resides closer to the edge of your network. And it is, sits there and and receives connections from your external partners, which could be, you know, vendors, clients, and others that are that exist outside of your organization who need to exchange or move files, ingress or egress into your organizational boundary. So UDMG sits external. It is accompanied by a a companion product called secure proxy, and we're not gonna cover that in detail on this call. But the secure proxy enables a a secure movement of data traversal through your DMZ in a secure fashion so that files can be ingested into the UDMG application from partners. Once UDMG receives files, that's where the integration with UAC occurs in through either a published event, or UAC can be also configured to pull for files that are received onto the EMG platform. And then within UAC, we perform all of our post processing on the files, including, if desired, we can move files back into the EMG for that same partner, alternative partner to pick up. So as the green line show here, the partner is, doing a pull of a modified modified data that has been has gone through entire life cycle of being uploaded from that partner, being ingested into the organization, gone through transformations of of various sorts that are applicable to your business, and then, finally, delivered back to, the partner in in modified form. Now this may look different for your organization, but this is the the type of life cycle that UDMG, facilitates. So that's just kind of a quick overview of UDMG for, for those of you who may not be familiar with it. It operates along all the standard, secure protocols that you would expect, for an external facing, secure file transfer server, like SFTP and FTPS, EDI protocols, like a s two, even HTTPS inbound, through a web client. And then it also facilitates, all of your partner, provisioning and authentication and authorization through whatever source you deem necessary, whether it's, local or or federated out to LDAP or a single sign on source. So let's talk through kind of some of the new features that are gonna be that were added to, the latest version of the product, if we could advance to the next slide. But first, here's the three areas of focus for this latest version. For each major release, we kind of focus on kind of thematic different thematic elements. For this most recent release, we we continue to prioritize automation, and that is the the integration with UAC in terms of one of our our kind of our our cornerstones of our of our of our orchestrated NFT strategy. We also continue to add some innovative type of capabilities that are somewhat nonstandard or or really competitive differentiators. And then, of course, we focused on core NFT. Right? Some of those capabilities that our current EDMG user base has asked for or things that we realized were, you know, commoditized or or necessary for standard NFT operations. Next slide. First and foremost, we added on the on the deeper integration with UAC, we now support multiple different triggers. Prior to this version, you could only trigger a UAC workflow when files were received. We now have ways of triggering when files are in a stage, so kind of a pre pre received status. Also, on air when when there's an error in the transfer or when files are retrieved or pulled down from from UDMG by a partner depending on the orientation of the setup since UDMG can be both a client and a server. So in all in all conditions, we can now trigger different workflows or publish events into UAC so that UAC's workflows can be tailored according to, the the, that particular, file transfer, type. We've also added a new, form of a new action called a run command. So prior to publishing a workflow into into UAC or after publishing a workflow into UAC, we can run this command locally in the system, which can be running terminal or any application on the UDMG system. And that's useful for doing sort of pre or post publish event type of activities. So as an example, after publishing an event to UAC to to launch a workflow, we might want to rename or maybe delete the file that was there, or maybe we need to, decrypt it locally before, publishing event. All that can be done now, through these, these run commands or run command tasks. And furthermore, we've added the capability of stacking, both these, run commands as well as the publish event, type of tasks. So you can do more than one. So if you wanna publish a bit, unset of commands, publish a number of events, all of that can be done from the context of the UDMG application. From a more innovation perspective, we've added new what's called endpoints within UDMT, which which delineate, either a source or a a destination, type. In this case, we've added on top of our SFTP and a s two and local file system, endpoints. On top of the traditional, file transfer endpoints, we've added this new cloud endpoint type. And what this does is it kind of turns the UDMG into more of a, kind of from of a point solution to a, kind of a a a true gateway where UDMG can receive files and then stream them to the back a back end cloud destination, whether Google, AWS, or Azure, and and several more to come soon, directly without ever having to touch the UDMG file system or or the file system that it points to itself. From the partner's perspective, it looks as if they are navigating a a file a local file system, but behind the scenes scenes, it's it's actually a cloud destination. So that is a a net new capability. It accelerates the movement of data into the cloud without having to kind of do it, you know, two step type solution, and it's completely transparent to to the partners. A new business services tag capability has been added as well, which provides a means for sort of more more layered access control over kind of your your your RBAC around the different UDMG objects. So this is something that has been requested by some of our customers for more granular control over the different capabilities of the product from an administration administrative perspective. So if you have different business units or maybe functional areas within the organization and you want them to make be able to make changes to one end point or what's called the data pipeline within UDMG, but then you have a different business unit that you don't want touching that, same pipeline, you can control that through the use of these business services tags. So sort of a layer on top of the, the already existing, RBAC mechanism, which is based on, permissions and roles. Okay. Next, next slide. On the universal, secure proxy, which is the companion product to UDMG, we've added a new, a new mechanism or a new mode for brokering those transactions securely through your DMZ. Prior to this, we had what's called session break, and now we have a direct pass through mode, which has pros and cons. And we have plenty of documentation on that, including, I believe, a new blog article that will be released here pretty soon. So if you wanna read up more on both the benefits of the secure proxy and the differences between the different modes of operation, you can peruse that. We've also added some built in health checks for when proxy is set up in a high availability configuration on an active active cluster. So that way, the cluster sorry. The load balancer knows if one of the nodes one of the one of the routes to secure proxy is unhealthy. And we've added more database support different database supports. In addition to the already existing Oracle support and in the built in SQLite, we've added Microsoft SQL Server. Let's see. Next slide. I'm not gonna cover all of these, but there's just been a lot of other little small enhancements as well. Couple highlights, I think, that stand out to me on the protocols perspective. We've added wildcard file transfers for remote get operations, remote get and remote put operations, which before we're limited to single file at a time. We can now do pretty sophisticated filtering around that. So when you're automating, pulling files from partners or pushing to partners, you can use, you you you can you can selectively do do so through, some sophisticated, wildcard approach. From an auditing and a logging perspective, we've continued to work on that. It's kind of the, more rudimentary aspects of managed file transfer, but it's it's still, very necessary. When troubleshooting, we now can, segregate the different log types, when when you need to be able to debug your logs you're not looking at, you know, database inserts when you're wanting to look at, file transfer logs. So we've we've allowed, you to kinda segregate those apart. Security always is a theme is an underlying theme in everything we do. We've continued to enhance our security controls, exposing more of them to the UI itself, which were previously in the configuration file. And I've added new security policies around things like our our password complexity rules. For example, you know, limiting password reuse so that you can't, you know, you know, limit to the last x number of of passwords during password change, limiting the number of repeating characters in a password, things of that nature. And last but not least, we've we've continued to focus on core architecture. So we've under certain conditions for certain types of file transfers under certain protocols, we've improved performance. We've continued to eke out performance gains. It's always something that we're looking into. Core performance capabilities for for transacting at immense scale. Right? We are interested in understand that our clients need to move millions of files a day sometimes, and that is a key focus for us. We have dedicated performance benchmarking and enhancements that are continually being made into that into that area. So that's kind of what's what's new for three three point three in the release announcement that goes out depending if you're signed up for that. You should see kind of this this list with probably in more specifics as well as in our online documentation. You can we're able to get access to the detailed release notes. You'll be able to see, you know, various screenshots of these of these various features. So if you wanna get any more detail, feel free to take a take a look there. And I think that's all I have for UDMG. K. Sorry. Just give me a moment. Okay. We have lots of questions, so let's get started. Will Roby let you know the user does not have permissions and what is needed, or will it just fail? And the my voice is back now. Sorry about that. It it will it will not allow you to do anything that that you don't have permissions for, but it will just yeah. It will just say that it you don't have permissions to do it at the point that you try. It it won't actually try to help you unless you ask what permissions I need to grant, but I guess if you if you do ask Roby, it will, it will go to the documentation and make some suggestions and recommendations. Okay. Great. Who can access Roby dot ai? Only administrators or everyone has access to UAC? Yep. Everyone has access once it is enabled. It is per instance setting. That's that's the way it is right now. So the Robey icon will appear. And, yeah, if you have have the permissions and the authority to do it, you you can you can ask questions about tasks, etcetera. Yeah. K. If Robie AI is behind a firewall, does that not allow access to online resources? Will it work with the documentation since I believe the documentation is only online accessible and not embedded in a local installation? The documentation is you basically, you, as the customer, provide, you know, two things. You provide the the LLM that that the Robie AI and the AI components will connect to, and you provide the vector database, and we we then provide the documentation in the format that you can insert it in the vector database. So it's very much local to your instance. It's not accessing the online documentation. So you should you should definitely have access to it. K. What other languages apart from Python are supported? Well, we have Klingon. Right? That's a bit of a joke. You can actually ask Groovy in different languages questions, and it will eventually and it will also translate for you. That's largely a function of the LLM and not Rovy itself. But you can use any scripting language. We use Python as it tends to be the de facto programming language for scripting for a lot of automation tools. But, you know, I've I have used Ruby to write JavaScript, for instance, or just Bash Shell or SQL queries. K. How is the approval message delivered, and how to specify who can approve? So that's a good question. So this is all defined in the in the approval task. It's a standard notification that can be, obviously, email or or, you know, standard product functionality. But, yeah, you you can you can list a a number of different addresses or or people that can that can approve, and then you can decide if it's all need to approve or if any need to need to approve. So it's it's, you know, you can very clearly define how the approval task works. Yeah. Definitely encourage you to have a look at look at the documentation or when you have Roby up and running, ask Roby about about it. Okay. Can we integrate our existing AI agent or custom runbook based assistant with Roby dot ai to leverage its knowledge based and context aware capabilities? You won't say that one, Greg? I should Yeah. I think you you should be able to today through, the MCP server. Right? So you should be able to connect an MCP client, like, if you wanted to connect, like, an LLM or your the application that your AI agent sits within using the m c an MCP client to Robey AI. You should be able to do that today. As we move more in-depthly into the agentic phase of AI, you will be able to, call our AI agents directly as well through the appropriate security naturally as well. You'd still have to have the right credentials to do so or security tokens to do that. I hope that answers it. You can also do AI tool calling directly from the MCP server. That's how you do it today. K. Is there a step by step documentation for moving high availability OMS server off of the file share resource? Yes. There there is a quite detailed documentation on it, Hal. They as I was going to say before before my voice disappeared, apologies again, the the default behavior is the same now as it always was. You so you have to change the settings if you want to use the the replication capability. But each OMS server has its own OMS database now, and you set up that replication to replicate between them. But, yeah, that's fully documented of of how you set that up. But the default behavior is is the same as the old behavior. Okay. How does UDMG confirm that a file to be sent is not empty? Because the file might be on the server ready to be sent, but if it's empty, the task will execute with a false success. Yeah. That's a good question. That's why we have a we have a two step process. We have a staging directory for the file to be completely received in before it's moved into the final destination, and then we trigger the publish event. So we don't we don't trigger the publish event until we've received the entire file and moved it into kind of a its target its target destination folder, hence the two step process. Okay. Can you elaborate more on the OMS database option using Neurraft, and what is Neurraft? Sorry if I pronounced that wrong. Yeah. No. It's Neurraft. It's it's industry standard replication. It's it's a it's a tool that exists out there. It's pretty well pretty well documented in the seven dot nine documentation, And there's a lot of a lot of information out out about it if you want to Google that. But, yeah, I I can't give you I can't give you a lot of detailed information on this session, but, yeah, it's certainly available in in our documentation or or online. K. Can m p u t and m g e t delete the files post transfer on the transferred basis to prevent a comms break restart creating duplicate transfers? Robert. Not current not currently. You can there is the the, like I mentioned, the new task which you can call independently. So you can do your schedule transfer, then call the task to do delete kinda manually, if you will. But it's not baked into to that operation itself. So there's not, like, a built in sync capability. However, we do have that that all of the the kind of a synchronized both in bidirectional or single direction with delete on on you know, after the files have been moved, all that can be done currently in UAC. So if we need a more sophisticated, kind of input and get, we can we can we can currently offload that operation to UAC. That's it's not, completely possible within in u, within UDMG today in an isolated fashion. So that great question. Alright. We have a few more questions, but we've run out of time, so we will have to follow-up with those after the session. Thank you, Gwen, David, and Robert for the deep dive into what's new in UAC eight dot o. With that, I thank you for joining. Later today, you will receive an email that contains a link to today's session, and you can rewatch or share with colleagues. If you're interested in getting a live demo of Roby dot ai, please join us in person for the final StoneBranch UserVerse event in Florida on May fourth and fifth. Thank you all again for joining, and we will see you at the next session. Thank you. Bye, everyone. Bye.
