Video: Build at the Speed of Thought: Developer Agent & Workday Data Cloud Live Webinar | Duration: 3552s | Summary: Build at the Speed of Thought: Developer Agent & Workday Data Cloud Live Webinar | Chapters: Welcome and Introduction (5.76s), Workday Build Platform (182.615s), Workday Extend Overview (634.685s), Workday MCP Tools (956.66s), Agent Actions and MCP (1832.515s), Agent Development Demo (2302.82s), Workday Data Cloud (2375.47s), Agent System Management (2754.545s), AI Impact & Benefits (3286.985s), Next Steps & Resources (3373.53s)
Transcript for "Build at the Speed of Thought: Developer Agent & Workday Data Cloud Live Webinar": Okay. Welcome, everybody. Welcome. Welcome. Welcome. I'll go ahead and share my screen. Okay. I'll just give it a second for some more people to join, but I hope Welcome, everybody. Where is everyone dialing in from today? I can see the chat on my right hand side, so it'd be great to get some interaction. Okay. All over. Hello, mister Bradley. Good to see you. Alright. Welcome. Welcome. Welcome. Just give it a cup a minute or so. Go grab a coffee. You've got me for an hour. Or, call a friend. Ask them to join in. Welcome, everybody. Nice to see some familiar, names. Hello? Hello? K. I'm just gonna give it another moment, and then we'll kick off. Hello? Sure. Keep introducing where you're from. It's, it's good. It's great to see the interaction. As you may know, these webinars, they're kind of like, solitary spaces. It's just me and a screen. In this case, a couple screens. But I can't hear any of you. So any feedback, any questions, any chatter, let's get it going in the comments for sure. Nice reaction there as well. It's a new feature on Goldcast. I haven't seen that before. Okay. Let's kick it off. Welcome, everybody, to today's session, which is all about developer agent and Workday data cloud. Now I I'm sure you've seen a webinar or maybe seems that every week up until rising, there is a webinar on these two topics, maybe even together. It's my intention today to try and show you something you haven't seen, go live in the demo as much as possible, and give you a the personal experience of how I've come to, know these tools, in particular developer agent, as it's the latest iteration in how to build on Workday. So my name is Peter. I'm an enterprise architect based in London. I've been with Workday for five and a half years, having joined through the acquisition of Pecan. But I've since, come to specialize in extend. I would also regard myself as an extend developer. Am I worth my salt? I'm not so sure, but it is the primary function of what I do day to day. I specialize in building mock ups and proof of concepts for customers, so I don't live with the headache of user acceptance testing or change management processes or a full software development life cycle. But it is my intention today to enthuse and inspire you, technical or nontechnical, that extend can be a place where you can build your careers as, certainly as I have. I'm joined today backstage by my colleagues, Conan and Carla. They will be monitoring the q and a. So if it's more than just a hello, it'll probably be them responding to you. But the two superstars on my team both come from extend development roles into this presales team. So, so joy of mine had them on the call. I'm sure you've seen this, but I think this is particularly important. Clear what is scheduled to be released very soon, in October and when these early features go generally available. But I do wanna preface that this is early adopter. We are in our early adopter phase right now. We have a lot of customers actually testing these tools live in their environments, and we're getting good feedback from from all of that. So just a bit of housekeeping. Today's webinar is being recorded, so don't worry if you're unable to stay for the whole thing, but I do encourage you to just sit back and relax and and enjoy. But you will have access to this recording on demand should you want to review down the road. There is also the q and a for those using Goldcast for the first time. This is primary means of interaction. Don't worry. I can't hear if you sneeze or cough, but I can see if you make a comment. So please do keep it interactive, and, I'll try to answer live. And if not, Conan and Carla will be supporting, via the chat. So today's agenda. So we're gonna start with a bit of a recap of Workday Build, which really sets the scene for these two products that form Workday Build, which is Xtend and Workday Data Cloud. I'm gonna run you through what it's like to build and extend with all the new innovation that's, been announced and soon to be generally available, as well as what data cloud means for you, and in particular, how data cloud can supplement you, when you start to build and deploy agents, which is my primary, area of focus. I'm admittedly not an integrations expert. I love UX. I love UI, and I love the fact that extend abstracts me from a lot of the headache associated with those processes traditionally, and I can focus on delivering solutions. So this is Workday Build. And to be clear, Workday Build is not a product. It is a platform. It is our means by which you can build on or with Workday. Now Workday build really started from extend. Now you'll see this is broken down into three pillars. On the left hand side, we have our tools. We have our app builder, which was the extend app builder that lives on the Workday developer site. And, we have Orchestrate. Originally, Orchestrate was packaged with Xtend, but was such a powerful tool in its own right that it is a generally available tool without the need of a license. You get some added benefits with Xtend and certainly some new ones with Xtend Pro. But you all will have access to Orchestrate, and I hope you're making the most of that. Then we have our newly announced agent builder. So not only does extend allow you to build apps, but you can now also build agents, custom agents, what we call your managed agents directly from the Workday developer site. And then we have data cloud, which sits as the kind of data foundation for all this actionability, that you would expect of a performant Workday agent. The second pillar, which is our marketplace, extend is quite a remark was quite a remarkable release in that a lot of our partner ecosystem, latched on to Extend as a means of creating products, using these tools and being able to ship them out to customers. The marketplace is a publicly facing website. I encourage everyone in this call to open up the marketplace and check out what's already out there, what's built by various partners. These solutions have been built with extend or other Workday tooling, but they are, unique in that they don't actually necessitate the licenses for the underlying products, at least in the case of extend. So you don't need extend to own a marketplace solution. Could be a nice entry point for you as you start to build an appetite for building yourself. And lastly is the community, often the unsung hero of Workday build. But for me, it's the most important and the, it's the people side of things. Right? We have over 20,000 people in the Workday developer community globally. To put that into perspective, that's about the same number of employees that Workday employs. We have, tutorials, certifications. We have all that facility for you if you wanna engage with it. We have a Spotify podcast. We have a YouTube channel. We have, regular meetups often hosted by partners, and, I hope very much that for those non extend customers on the call that you sue too soon will be part of this community. Every year, we have our developer conference event. Our DevCon event happens in the middle of the year in Las Vegas, featuring a 24. Now if that isn't enough to attract you towards an extend license, I'm not sure what is. But, a whole lot of fun. It is also Workday's second largest customer facing event, which is pretty indicative of where extend and Workday data cloud lies in the overall Workday enterprise AI strategy. So just a bit just a level set. I appreciate not everyone is gonna be familiar with Extend. So what is Extend? Extend is the Workday tooling that allows you to build custom apps and now custom agents that sit on top of the Workday UI. They use Workday's architecture. They inherit Workday's native security. They look and feel like Workday. Let's say you have, a particular button, alright, in Workday, and Workday changes its button's colors to purple. I promise you they won't, but should they, so too will your UIs of your apps, update. So, that is part of the pull towards extend. There are over 200 widgets supported. I believe the most recent to be supported is our AI widget, which allows you to actually embed generative AI in your, work flow of work when using the UI. Much like I'm sure you've seen the automation of job descriptions through AI, that precise widget we've allowed you to use in your extend applications. So a whole host of power in its own right even before the announcements. You know, extend is, an opportunity for you to build custom solutions that tailored to your specific needs, whether they are augmenting or supplementing existing functionality within Workday or creating that new functionality. We have a pool table, and a ping pong table in the office, and I remember one of my early use cases was a tournament booking system, which never went live because they don't allow me to build, on work day on work days instance. But, hey, this the ideation is part of joy as well. And what came out of DevCon was our announcements around, being able to build anywhere and surface everywhere. Now to accelerate, your experience as you build, we have introduced our developer agent. Now before DevCon of this year, we had an introduction into a generative AI assistant that assisted you when you built your extended applications. It was called developer Copilot. Now we used to personally attest that developer Copilot brought the time to production for your use cases down by 50%. This was firsthand evidence from, select few customers. But, personally, for me, I I wasn't experiencing that so much myself. And it's maybe because I don't do the UAT change management regression testing part of things. But I can confidently say that with developer agent, I think this time, at least from anecdotal evidence, is closer to 90%. To give you an idea, over the last five years of building extend apps, I built maybe about a 120 applications, big and small. In the last three months of having access to developer agent, I've built around 30. And, I think there's no greater showcase for AI, accelerating the flow of human work in the space of development. You know, I I think Vibe Coding gets a bad rap, but it's popular for a reason. It makes you feel like you can have a greater impact, abstracting you from the technicality. I saw a LinkedIn post of the Microsoft CEO posting about his weekend vibe coding project. You know, there's something very empowering about, agentic codevelopment, and we are so excited that that's now a reality in the Workday space. And that's that's the key differentiator between pre and post DevCom. You can build the developer agent in your in the Workday developer site, but, uniquely, you'll be able to develop wherever you reside. Users of Claude, users of Versa, codecs, You will be able to access developer agent from within those tools so that as a developer, you can build where you naturally reside. Not only that for the developer experience, but for the end user experience, you can now interact with the agents that you produce with extend anywhere that supports MCP and in time a to a connectivity as well. You'll see some examples here. Our front door of choice is, you know, openly, it is Asana. But if you use Microsoft Teams, Copilot, Gemini, wherever, you have your own custom front door. I'm dealing with some customers right now who have their own custom front door that wanna connect. It's a possibility for you as well. Takeaway from this slide is the key, in diving in with both feet strategy approach of Workday is we are embracing openness. If you take a look at how our product and engineering teams are aligning, they are building with the direction of openness, which wasn't necessarily a priority, in in years prior. Okay. And just to keep it a little bit lighthearted, but growing up, I all I thought agents were only for celebrities. It seems the corporate world can have agents too. And if this is the team on the call that can make that a reality for your company, well, then all the better. So let's dive in. Before I head into the tools, I just wanna preface. We still have APIs. Right? So Workday is largely built on interoperability. We have around 4,000 APIs of different kinds. They will still exist. We've renamed them to application APIs because they serve a specific purpose. They're great for integrations. They're good for application connectivity. But when it comes to agent communications, as is being made standard by companies like Anthropic, The model context protocol is the UI of choice for agents. And we have and was as we recently announced, come to support MCP connections. We have our own Workday MCP. I'm sure there's a couple out there that have built their own custom Workday MCP. I'd like to know who you are and have a conversation with you afterwards. But, I'm gonna show you why working with the Workday MCP is optimal. And in time eight zero eight. Now eight zero eight is already live for some of our first party agents, like self-service agent, for example, but this too will be a means of communication for your custom agents that you build. Alright. Let's dive in. So let's hop into the Workday developer site. So I'm gonna be pretty hands on with this. I'm gonna just try and show you some live use cases, show you how they were built to give you more than just what the slides would, would would convey. Right? So you can see all my different, use cases here. I'm in a particular environment. You can see I've been pretty active over the last few days. I've got 15 apps in here. But you'll see that I've named them agent. Right? So it's an important distinction for you. For now, an app serves as the container for an agent. It may be in time that you, you know, we have a specific agent file, but, for now, your apps are the house for your different agents that you create. I say that because I just don't want any confusion with the terminology. But let's hop into the tenant. So in the tenant, I have a similar view, which is the app manager. Same, but accessible within the tenant. So I can see which apps are being deployed to the specific tenant. Alongside my apps, I have the agent management hub, otherwise referred to as the agent system of record. Now it's here where I can see the usage of my agents within the Workday tenant. This is not restrictive to the first party agents that we produce, first party being those agents like, self-service agent, deployment agent, payroll agent, and a whole host of others that I invite you to keep abreast with the road map of, but also your custom agents. Right? And when I say your custom agents, these could be may agents made with extend, but they could also be agents made with other tools that communicate with other systems and talk to Workday registered in here as well. Right? And this grants us access to the underlying telemetry to be able to monitor these agents. Right? So you'll see here I have, an agent. You can take an example of the market skills insights agent. I can configure, activate, and, configure their skills. Now something to make notes of is that agents, they have skills. Right? Skills are essentially, text files that have a set of tools attributed to them. Tools being an abstraction from APIs. Right? These are essentially APIs. I'll come to highlight why our MCP tools differ from a traditional application API. But you can think of this is how the agent works. The agent has a digital brain, which you will choose. That digital brain is your LLM of choice, whether that is a chat GPT model or a Claude model, for example, or Gemini. You choose that. You give the agent its definition, and then you append to that agent a set of skills. And within those skills, you have the tools. They can be Workday tools. They can be any other tool. In Workday, in our agent system of record, the key convenience to you is that you get to administer access to individual skills based on security group. And you can do this easily simply by toggling, right, directly within the agent system of record. That's a key level of flexibility and control that, yes, you could replicate outside. You will have difficulty doing so, but, it's not impossible. What using going Workday native grants you is complete ease. Okay. Let's go ahead and take a look at an agent. So now in the top left top right hand corner, you can open up the agentic chat. And in here, I can see my historical chats. Now I have preloaded some chats because, well, this is just a conversation. I don't want you to have to wait for the responses, but this is an example of an internal Talent Finder agent. So this agent allows me to see requisitions open within my organization that I could fill with an internal employee. Something that may seem menial, but through an agent, I'm able to surface this insight readily and in a secure manner. Right? You'll see here a grid. This is an example of what's called ATUI or agent TUI, which was another one of the announcements that was released. The ability to go beyond just plain text, back and forth with an agent, but the ability to create AT UI pages that support a subset of the extend UI widgets. Give you an idea of what that is now. I think there are about ten eighty UI supported widgets. Things like it might be particularly useful if you want the input from a user to be of specific format. Like, date time formats can be, cumbersome. But you'll see here I can, identify a particular requisition and proceed, identify this employee, and fill that position all through a natural language. What more If I choose to go out to market, if I want to hire someone externally, I can use my market skills insights agent to infer what are those trending skills in other external, requisition repositories. Maybe that could be LinkedIn or, indeed, can I connect to those systems and surface the trending skills of the roles that I'm hiring that maybe I missed out on as a recruiter? Well, you can do that with an agent. I've, comedically named Eagle. I have a colleague, whose last name is Eagle, I think. And, I only highlight that, because this is a key differentiator of an agent as opposed to an app. There's the opportunity to inject some personality. Right? One of the early thinking that I had was, you know, if you have if your company has a Twitter or an app, you could encode the personality of your digital persona and, instruct your agent to behave as such. But, anyway, perhaps a bit experimentive. But, yeah, you can see here, I can now identify three top trending skills from LinkedIn, from Indeed, and I can append them to that job requisition should I want to hire outside. Now let's look at project management. So this is this project this is a project staffing agent where unlike a requisition, I'm actually filling a project based on employees' skills, education, certifications, various datasets that reside within Workday, right, across various data objects and data sources. I can pull those together with the Workday MCP into a singular agent that gives me oversight over all those different areas. I can ask the agent to list me my different projects. I can select the project, and it can identify a top candidate. And it can explain to me why. Point is is that you have control over how this agent, responds, what it uses in its reasoning, what, how it gets access to the data, the underlying data, whether it's reading or writing back into Workday, that all happens through the Workday MCP. So let's go ahead and take a look at this app under the hood. So now I'm in the developer site. I'm in this specific project resourcing agent. And if we take a look here at the agent definition, I'm in our visual mode right now. But what you'll notice is that if I switch into the code mode, not a whole lot has changed. Right? And I think this is something I want you to appreciate is that composing an agent is like short essay writing. Right? It's mostly text, which opens it up to, nontechnical users who are trying to think of and implement use cases. In my experience, when you come to, troubleshoot an agent, it can be a little bit con confusing because how do you troubleshoot this? Right? Like, these are just a set of instructions. But to assist you in all that, we have the Workday developer agent that lives side by side as you build. I can take a look at any one of these skills. Again, skills are appended to the agent, and within each skill, we have a set of tools. Now these tools, they are representative of the Workday MCP. Now how do I know which tool to use? Well, it could be as simple as saying, I need a skill to, find direct reports location and letting developer agent do the thinking for you. Our developer agent is trained on not only our Workday MCP, but also our app catalog, all of our documentation. It is the expert on all that there is to do with extend. And you'll see there, I've forgotten the key first step, which is to connect to my tenant. Let's go ahead and do this and try again. Now without letting this complete, what I hope it will show you is that how it's thinking. Now you'll see that it may be starting to take, trigger off sub agents. It's reading through material. So developer agent itself is actually a collection of smaller specialized sub agents. But let's take a look at this skill under the hood whilst it works. I can switch now into code mode, and you'll see how I've defined those different Workday MCP tools. Now this raises and okay. There you go. So it's now proposed its changes. You'll see that every change that it proposes, you will see the individual lines that is deleted and as well as any new material that is generated, and you always have to explicitly approve. What more is that you can check, you can toggle precisely which amendments you want to, include, but that doesn't mean proposed by developer agent. So let's hop out of this. Let's take a look at these tools. So you'll see here an example of a Workday MCP tool. Again, you don't need to worry about these IDs. Developer agent does it for you. And we have an example of another kind of tool, which is an agent action. So let's take a look at the work to MCP Explorer. You'll notice that in our early adopter period right now, we have a list of around 426 tools. So you may be asking, well, where are the other 3,500 APIs? Right? Well, it's our intention that in a month's time, we will reach the thousands. K? So this list is continually growing. A bit of insight as to why we've had to start from scratch is that these tools, they are not just APIs, which is the common misconception. We've defined these as agent ready tools, which are a completely new construct to Workday, wherein we have rewired and refactored existing APIs so that they are more appropriate for agent consumption. That means their the data inputs and outputs are faster to read. They perform better at scale. They are respectful of the workday security of, human in the loop, all of which means more compliant and, ultimately, it's another aspect of all this that I won't delve into, is lower token consumption. To give you insight on this, we had a Workday MC pre EA program earlier this year, and it wasn't particularly successful because we thought it was sufficient to simply wrap Workday public APIs in our MCP. We also tried similar with our own first party agents, like self-service agent, and they did not operate, as expected. And so for anyone building a custom Workday MCP out there, trying to maintain custom logic to chain different Workday APIs together, that that is, technical debt that you do not have to own if you come to have the Workday MCP. The Workday MCP also tries to aim for two second response times versus an API, which will typically time out after thirty seconds. So that also enhances the conversational flow of an agent. But those remaining APIs, we still allow you to wrap them. So now I'm gonna hover into orchestrate, and this is gonna show you the orchestration builder, which, like I said, this is a workflow builder. You can, build integrations with orchestrates orchestrate. You'll notice that these are prescriptive steps that must be followed. You could describe them as deterministic. This is perhaps, particularly important with those more sensitive use cases around payroll or or hiring. When the agent identifies what to do, perhaps you want to prescribe precisely what it should do, step after step through an orchestration. What is new to all this is the ability to expose an orchestration via the MCP as what we're calling an agent action. Right? So this is huge. You could argue that this is a workaround for all those 3,500 APIs. I invite you not to think in in that way. Imagine chaining together multiple APIs, not only Workday APIs, but third party APIs or implementing custom logic directly within here. Right? This is hugely, hugely powerful. What more, all these agent actions can also be accessed by your third party agents as well, extend or none. And to add further, oomph to this announcement, you'll see here a bit of a simple agent action that I've created. You know, it's only one step, but it's an integral first step because it calls on Google Calendar. Now Google Calendar, you could have, connected into their APIs before, but a key announcement here is our set of pipe dream MCP connections. Effectively, an MCP gateway into thousands of different third party providers, MCP tools. Most of the Google suite is supported in here. Most of the Microsoft suite is supported in here. So you can create these connections to other MCPs directly from within an agent action, from within the orchestration builder. So hugely powerful when it comes to connectivity within Workday and also outside of Workday. So, let's go ahead and take a look at an example of an agent that utilizes one of those agent actions. I'm gonna take a look at this. Well, I believe, actually, it is an application. So let's take a look again at the app manager. And you can see here, I have a particular application, and this was created also with developer agent. You'll you'll start to notice what's been created by developer agent. It will be experimentive experimentational with rich text and add widgets and, emojis. Seems to be a popular choice for UI design for developer agent unless instructed otherwise. But here you can see I'm able to create complex pages. Right? This isn't just a singular widget on the screen. This is conditional. My colleague Conan might be chuckling a little bit because this is his application rewired. I mean, even that, you know, for existing extend customers, I imagine your early experimentation with developer agent is troubleshooting something that you've already built in extend and taking it that next step further. So I can evaluate Steve Morgan's succession plan. I can compare different employees who are assigned to their particular succession plan. I can compare them side by side, which is, actually a a pretty common use case when, for extend, it's creating custom talent cards. But one thing that's pretty neat here is export to PowerPoint. Now if you've ever wanted to well, okay. Well, there you go. It failed. But, you you imagine what's going on behind here is that I have my succession planning agent, and I can take a look at the underlying agent action. And you'll see here connection into Google Slides. Right? So the opportunity for you to use agent actions in this way creatively is another huge gain in flexibility. What more is that the agents that you create, they're also available in third party platforms. So here, you'll see an example of me speaking to an agent, my interview companion agent, directly from within Copilot. And if you'd like, you can then publish this to Teams, which is again, we wanna meet your employees wherever they reside in their flow of work, and we wanna cause no obstruction of that. So our means of embracing the openness is through tools like the Workday MCP and in time a way as well. And to show you the accessibility of all this, I wanna show you something that, I I built with a, the the daughter of a colleague of mine on their work experience at Workday. Right? No experience with development at all and wanted to build a use case around, a bit like your Spotify, recap directly from within the Workday. And this is quite amazing that in the span of one hour, she was able to build an operational application with an interactive UI, with a variety of different widgets, even some use of emojis here, as well as an accompanying AJ. So it's pretty remarkable, the accessibility that this brings. Speaking candidly, the time to test, as in the time from first bill first building anything to the testing of your solution, has been drastically reduced. Right? This, I think, makes us in this era now of rapid ideation and rapid experimentation. It's, it's not the most encouraging sometimes when you hear of an extend app being ideated and then deployed to production and there being, eighteen months between those two periods. I strongly think that those are gone. Right? I I I would very much encourage everyone to you know, we have these extend test drives as well on a regular basis if you'd like to come and try out these tools for yourself. It's simply a matter of asking, within the developer agent what you want to do, creating, delivering, and then trusting that it's being done with confidence. Right? So something I wanna show you as well is not only in the, surfacing of your agents wherever they may reside, your workers, but also in the flow of work. So you'll see here an example of a Miro board. Now I know some people like to work visually. Perhaps you use Miro boards or other visual tools to define your use cases. Well, you can now use, use visual information, encode it within what you're seeing here is cursor, which is like a tool like Claude. And this is connected to my Miro instance via MCP. It's also connected to developer agent via MCP. And now I can have my workflow take a visual output and produce, a mock up quite readily. I haven't timed this, so I'm not sure how long it will take. But let's try something else. Let's see if this works. So let's try this. Okay? I'm gonna try this. I need an agent that allows employees to create and manage meeting rooms. Let's give that a go. And whether or not that works, it's a representation, right, of the openness that we are providing. This goes beyond the Workday command line interface, which you might be familiar with, any developers on the call. This goes into actually giving you the extend expertise that developer agent is trained on, to to build. Okay. Well, let's give that some time to work. And while that's firing away, let's talk a little bit about data cloud. So where I think Workday data cloud comes into the picture is that it is really setting the data foundation for your agents to perform, in a performant way. Historically, it has not been the easiest process to make the most out of the rich semantic information and data that your Workday tenant holds. Right? You may be exporting that information into your own data lake solution, into various downstream tools. Like, if you use Power BI or Tableau for your analytics, data cloud serves, you may be exporting your Workday data outside. Now this creates quite a lot of friction. One, you have to duplicate the data. And secondly, once you've duplicated it, you have to transform it and then load it into your receiving tool. Now that ETL process is particularly cumbersome depending on on which systems you wanna marry together. But not only that is that process of creating that pipeline complicated, but also, yes, something that you will have to maintain in the long run. So this is how historically, how it's been done is Workday sits amongst the various tools that you use within your organization. It is just one part of, a many piece puzzle. Maybe maybe feeding data into Workday or bringing it out into Snowflake or Databricks or, again, into those downstream systems like Power BI or Tableau. These traditionally require you to duplicate data from its source into, the receiving tool. But what Workday Data Cloud grants you is the ability to do so via zero copy, to be able to query the data where it naturally resides without having to duplicate it, without having to maintain, different data retention rules across different systems. I think of the example when an employee asked to delete all the data that you have stored or processed for them. Now you can do so meticulously through searching audit trails of where their data has, been entered and been processed. But with zero copy, you delete once from the source, and you can be sure that there's no risk of that their data being retained for any longer than it should, ultimately keeping you, in greater compliance with whatever data processing regulation that, is relevant to you. For customers on the call, I expect it will be GDPR. Now Workday data cloud is a composition of three different things. The first is Prism. Now I I mentioned Prism first because you may have heard of Prism before. That is our data ingestion and management tool. Prism plays an important part of Workday data cloud because it is the means by which you ingest third party data into, Workday. That still is true. So, therefore, Prism is now becoming part of Workday data cloud. Then we have the Workday data lake. Now this is where your Workday data is materialized so that it can be accessed via zero copy from third parties. For any techies on the call, the protocol behind that is Apache Iceberg. And without caring too much for the protocol, what this should signify to you is the potential to connect with any provider that in theory supports connection via Apache Iceberg. Now I I I would say to get excited by the potential, but the current reality is that we support this connection, this particular Workday data lake connection into the five trusted partners that you see on the right hand side there. The most recent to be announced in this mix was AWS. Now the Workday data lake, it is designed for higher volume, data. It has a relative latency of maybe about an hour to up to a day, which is configurable. Now where you need more real time data but lower volume data, we have the Workday live data query. Now this is, again, it's it's lower data volumes, but this, in my opinion, is precisely what you're gonna be using when you come to use tools like Power BI or Tableau to run your analytics. Now I think LiveDataQuery is, incredibly enticing as well because you can connect to it, in time via MCP, which makes it, hugely interoperable wherever whatever AI tool that you use. And in the context of agents, you can imagine an agent, traditionally having to call upon many different MCP tools to read or establish the context layer that they need to perform. Rather than that, you can assemble the data via data cloud to give the agents a rich contextual layer that may not only be Workday data, but may be the data that resides across your current stack. Now Workday data cloud isn't a push for you to move your data into Workday. There is equal utility, arguably more, to move your Workday data out. The key affordance to you is that this is all happening through zero copy, which, at the end of the day is still data. I I I liken this to, my my guitar amplifier at home. I've had it for ten years. I bought it in a different country, and I never bought, replaceable power socket. So all this time for ten years, I've been using an adapter. Right? And that's what, in essence, you've been doing before. With zero copy, you plug straight in. So at the end of the day, the output's still electricity. The output is still data, but how you got there matters for a whole lot, and it is certainly the path of least resistance that you will benefit from in the long run. Okay. And let's let's take a look. Let's see let's see if, if my if any of those apps were created. Okay. I can see that project resource and agent has been created from Miro. I can see my meetings room agent that was created from my, phone using voice to prompt. It looks empty right now, but perhaps we'll see an update soon. I'm sure this is supported with Siri and Alexa, if any of you wanna try it. In fact, the workflow behind this is arguably more convoluted than the building of the app or agent itself. But let's take let's go back into cursor because something I wanna show you is disconnection into live, data, the live data query. So you can see here that I am conversing directly through my AI tool of choice, which in this case is Cursor. I can simply ask how many employees are in sales in Asia who've been here for more than ten years or less than thirty years? Now anyone who's had experience with SQL, you'll notice that this is just an SQL. You could do this with SQL. Right? Which is also something you can do with Workday data cloud. But the ability to turn this singular prompt into a complete Workday workforce analytics canvas or dashboard within a matter of minutes, Right? Whilst respecting the security that the end user has access to in Workday to allow for self-service reporting with the assurance that no end user is accessing any more than they'd be able to from within the Workday tenant. That's the level of flexibility and control that Workday data cloud offers you. And one little piece you can see, I I I have a whole lot of slides that I did not use, but one thing I wanna show you as well is back in the agent management hub, final announcement is in the agent registry, you can see your various agents. Now I can take a look at this particular interview Copilot agent. I can see the skills. I can see the analytics behind it and the monitoring. It's not particularly fleshed out here, but there's actually also a page within Workday where you can see the explicit inputs and outputs of an agent if you're given the appropriate security level. If you want that level of auditability, you'll be able to see precisely which actions that the agent has taken on behalf of a user. That's particularly important important when you're thinking about business process auditability, which much of the tools, seek to interact with. But what you will also see here in time is a new tab called agent passport or agent stamps. Now what this does is it will continually attest your agent against various industry standards. Right? It will continuously, vulnerability scan them for these various, resistance stamps to align with OOS, NIST, NIST ISO standards, for example. You'll see here a vast variety of tests that every agent needs to comply with. You can set the rules of what meets your certain grade of of compliance when you build your custom agents and register them in agent system of record, but this allows you to, again, have that level of confidence. You could say it marries the the trust component to the operational component that agent system of record grants you. I think there's also a row one thing on the road map, and I encourage you not to get so attached to road maps. But one thing on the bottom right hand side is you'll see interoperability with other agent system of records. Right? I think this is an important consideration is that your agents don't all just live in Workday, and we understand that. Our our goal is really to embrace this openness and to give you the tools to do so. If we're not there now, you can be sure that that's where we're heading, and it's just a matter of timing. So that is the ASR Road map. Come on now. Let's see if there's any any impressive thing going on here. Nope. I often think, though, that with all this, you know, where does it where do you get the time back? Right? Like, if I can make a mock up this quick, what do you okay. This this one clearly has something going on. Okay. So here I have my project staffing agent. That's a work in progress, and you can see it's, actually assembled the the app as well as, all the tool identifications needed to operate each of the skills. So preheat, you can see it's even updating as we go. This is actually another feature released is the merge conflict, which allows you to apply changes, and see the version changes between, your previous version and your current version if someone's working on it live. So particularly useful. Let's apply this, and let's see what changes it's made. Okay. I see what it's doing. So it's built the agent first, and now it's actually trying to build the app as well. So, again, this is rapid ideation, rapid experimentation. Right? So I I often get asked by a customer. You know? Traditionally, extend has been sold on use cases. Right? We say, come to us with two to three use cases to make the deal go live. That may be important to your procurement team, still is. But I invite you to think of it this way. You know, you don't purchase Claude Code or the enterprise license of ChatGPT because you have use cases in mind. You purchase them because when the use cases arise, you wanna be able to act on them readily. And that honestly is not something that Workday's prior generative AI developer assistant granted you. But with developer agent, that's a reality. So, yes, do assemble those two or three use cases, but have the confidence that, you know, you could mock up your use cases five in a week for an without the dependency on a technical team. Now there is still utility in being able to troubleshoot, right, this with the technical know how of how to. But in fact, I'm not gonna touch this up because it's something's going on the background. But if you injected an error, it was it's as easy as copying and pasting the error into developer agent, and it will actually troubleshoot, for you very effectively. One thing that you'll see here in time as well is a new tab called agent traces, which will actually be able to give you precisely the amount of token consumption per MCP tool call, a layer of visibility that we don't yet have in our EA program at the moment, but is for certain coming live, yeah, very soon. But let's, let's let's close it up. I know we're near the end. If you do have any questions, please do put them in the chat. I am, keeping an a watchful eye over that. But in the meantime, I'm gonna close as well. So what to take away from today? So the first thing is, I think, an overall appreciation that Workday build is agent ready. Right? You can now but we are have come to support the MCP protocol in time eight zero eight. We have our own native Workday MCP for all the reasons why the new construct of agent ready tools are advantageous over the use of traditional application APIs. We've allowed you to build custom agents with the extend tooling, but also to build agents wherever you reside using the work VAMCP. And to assist you in all that, we've given you the most proficient digital assistant that we could, which is developer agent. Alongside that, in the creation of agents, we have the management of agents, which is the agent system of record component, which in time will also have the security attestation component, which is the agent passport. Then we have the ability to build at the speed of thought. Now I must say in the last three months, my flat has never been cleaner because I have all this time to do in between apps, you know, building that I can now dedicate to go listening to an album I haven't listened to. I think we are still in the early stages of relishing in AI, accelerating the flow of human work. So let's enjoy it a little bit. Let's go be human. You know, I'm I'm saying how are you to someone. I haven't said how are you to in a long time because I've got that extra time now with developer agent. And like I said before, I think there's no better showcase for AI accelerating the flow of work than in the development space. We're not talking about minutes saved. I'm talking about weeks, months saved, with a single API call. Right? With a single conversation. And lastly, we have the ability to enforce zero copy with the Workday data cloud. You know, no longer do you have to worry about the duplication of data outside from where it resides and all the technical debt that comes in doing so, managing security in different systems, managing, ETL processes, and so forth, not only for your technical landscape right now, but of particular utility in the agentic space. And with that, some next steps. So firstly, if anyone's come and rising, say hello. I'm, I hope I'll be there. I'm sure others will be too, but there's gonna be an extend boost. There's gonna be an extend hands on labs wherein you can test all this out live if you want and, do your best to break it. And then we also have our extended test drives. These are virtual but held regularly. If you wanna join virtually, you can, contact your account manager, and we'll give you the latest schedule on that. We have some, you I invite you to check out the YouTube. There is a video out there featuring our head of developer relations, Nick Morris, who spoke at DevCon about all these announcements. It's twenty minutes long. I encourage you strongly to watch that. Far more conversational, but, comprehensive of all the updates. We also have the Workday blog. There's a great article out there on building on Workday's guardrails. I invite everyone to look at that article because it prescribes on how to build on Workday in three different paths, native, custom, and, native, completely custom, and custom using Workday MCP. So I invite you to read that article. As well as that, you'll see the LinkedIn community, the Workday developers' LinkedIn page. They are releasing posts, frequently in the run up to rising as we start to engage more and more with these announcements. You'll come to find that these announcements are the biggest ones that Workday has in my opinion, you know. So it's it's important to stay abreast with all this. And lastly, give feedback. I'd appreciate very much. I don't get to do this all that often. And I'm very grateful to, Florin and Johannes for marketing for inviting me to do this, and especially grateful to Conan and Carla, who've been supporting with the q and a. And, as I mentioned, you'll have the recording to hand, and you have you have more questions, ask your account manager, but I'm an open book too. So I'm sure you can figure out how our emails are written. If you have any questions, please do ask. But this has been enjoyable, and go have a great rest of your day. Let's go let's go be human. Take care, everyone. Thank you, everyone.