Name: Build enterprise-ready AI apps and agents with Workday Extend Professional
Uploaded: 2026-04-09T18:30:09.049Z
Duration: 56 min 53 s
Description: Build enterprise-ready AI apps and agents with Workday Extend Professional

Transcript for "Build enterprise-ready AI apps and agents with Workday Extend Professional": Hello, everyone, and welcome to today's webinar that's all about how you can build enterprise ready AI apps and agents with Workday Extend Professional. Today's event is part of our Looking Forward with Workday webinar series designed to give you insights into how you can do more with the Workday platform. My name is Dhara Patel, and I am on the Workday build product marketing team. And I'll be your host for today's session. We will be discussing a couple future road map items, so please note that our product statement does apply. And here are just a couple of housekeeping items. We are recording the session, so you'll receive a copy of it via email within twenty four to forty eight hours. Workday uses the Goldcast platform for this series, which operates best when you're using the Chrome browser. If you happen to experience any audio issues when speakers switch, please make sure to right click on your browser tab and select unmute site. And overall, we have a very jam packed session today, so there won't be live q and a at the end with the speakers. But please feel free to drop any questions via the q and a button because we have a couple of experts on standby here to get you answers live. There's also a short survey at the end where we'd really appreciate your feedback. Our speaker lineup today is full of technical pros. Joining me today from the Workday side is Don Stacy, who's a senior principal enterprise architect. And we're also joined by Clint Harwood and Chris Humphrey, who are both directors of Workday Accent at one of our partners, The Groove. And this is a very quick look at our agenda. I'll kick things off with an overview of Extend Professional before handing the mic to Don, who'll show you some demos of our tools in action. Then Clint and Chris will share their real world perspective on common use cases, the benefits of going pro, and practical tips for building with AI. So let's start with an overview of Workday Build, our AI developer platform for people, money, and agents. Think of Build as everything you need to develop custom innovations natively on Workday's platform. It's made up of three key pillars. First is a best in class suite of developer tools and a unified developer experience that streamlines how you build apps, integrations, agents, and connect data. Second is our robust partner ecosystem resources and programs that help developers upskill, earn credentials, and connect with peers. Today's session is all about that first pillar of tools, specifically building custom apps and agents with Extend. Now, we know you have a lot of options on the market to build custom experiences, but we believe Extend is your best toolset to build enterprise grade people in finance apps and agents for three main reasons. First, with Extend, you're building directly from Workday platform components, including data, business processes, reporting, and UI, where apps and agents have the deep context to be truly effective. We'll be diving a lot more into the importance of data in this session. Second, extend apps and agents are built to interoperate. OpenAPI's seamless integrations, over 3,000 connectors with MCP and agent to agent orchestration ensure you can connect apps and agents across your entire tech landscape. Lastly, when you build on the Workday platform, security, resiliency, and trust are built in. Your custom apps and agents inherit Workday's version safe and always on auditability. Our end to end agent stack with agent system of record empowers digital workforce governance at scale. Now today, there are over 3,000 extend apps in production, and customers are using Accent to consolidate systems, respond to urgent business needs, and reimagine new ways of operating with unique employee programs. Here are just a couple of my favorite examples. Shelter Insurance replaced a twenty five year old legacy system and transformed a complex advertising program with a single Extend App. The app integrates with Workday financial management and automates invoice processing, payment tracking, and corrections handling. It centralizes everything from regional campaigns across 14 states to the actual maintenance of 4,000 office signs, all while providing leadership with real time visibility into ad spend and ROI. On Running built an app that cut contingent worker onboarding time from hours to under a minute. It allows hiring managers to handle the full process with no IT or HR involvement. They also then automated off boarding to eliminate compliance risk at the end of contracts. Lastly, Ecolab delivers personalized employee training at large scale to accelerate the time to competency for thousands of new hires annually. With Extend and Orchestrate, they built a platform where managers can curate individual learning paths from thousands of activities, and they track progress every step of the way to offer personalized support. If you'd like to learn more about these customer stories or other use cases, please check out the Extend Use Case Gallery on our website. Now, with any version of Extend, you can build custom apps that run natively on the Workday platform. But Extend Professional is our premium offering that not only allows you to build more apps with greater data instances, it's how you can take your innovations to the next level and build with AI. You can think of our tools in three categories. First, Extend Pro supercharges developer productivity. You get access to our AI powered developer copilot that can help you create app pages, WQL statements, code snippets, and even context aware validation scripting for Extend Apps, all from natural language chat. It can also help you search, compare, and get recommendations on the best API to use based on what you're building. In June around DevCon, Developer Copilot will be able to generate entire orchestrations based on prompts and uploaded file schemas along with context aware documentation. All in all, these AI powered DevTools can boost productivity by over 50%. Your devs are spending less time on manual line by line coding. Instead, they're skilling up, building faster, and generating more value. The second category is boosting outcomes for end users. With Extend Professional, you can bring AI directly into the app experience in completely new ways. AI widgets lets you add custom prompting capabilities right inside your extend apps. Users get the benefit of Gen AI without ever leaving their workflow. Through our AI gateway, you can tap into Workday's own large language models. The same ones we're using a power applications like Skills Cloud, and embed that intelligence directly into your apps. You can go even further by blending best in breed AI services from AWS, like recognition and Textract for things like image recognition and document processing. To recap here, developers use AI to build faster, and then they add AI functionality directly into the apps so end users benefit too. That brings us to the third category we know many customers are very excited about. We're bringing agent building to ExtendPro later this year. Our early adopter program kicks off in June and it'll be GA in R2. Sonnet Agent Builder will bring together some of our best tools, including the visual building experience from Flowize, pre built connectors from Pipedream, and the full Workday developer platform experience, all rolled into one place so your developers can build and deploy custom agents with confidence. We know there are a lot of agent building tools out there. What makes Workday different is that we bring enterprise grade governance and control to the entire agent lifecycle, including covering versioning, source control, lifecycle management, access control, and rollouts. With agent system of record, you get a single control panel to manage every agent, whether it was built by your team, by Workday, or by a third party. Extend Professional gives you the right tools to develop, test, and deploy these agents throughout the entire enterprise. Combining apps and agents is your key to delivering exponential impact. That's because effective agents need three things to really work in the enterprise, quality data, the right context, and compliance. Let's start with quality data. Workday already holds the system of record data for people and money. This includes your org structures, comp history, rule hierarchies, projects, and so much more. An agent that can't or doesn't have access to this will be working off incorrect and inaccurate data. Agents need to work with your specific data, not approximations or outdated versions of it. Second is context. Agents need rules, so they're not deciding what to do on their own. This means security roles, business process rules, org level eligibility, and approval chains. It's the logic that governs how data actually gets used from both a security and organizational perspective. For example, an agent that knows your comp data but doesn't know your approval hierarchy is a liability you really don't want. Third is compliance. Your employees trust you to handle their payroll benefits and absence correctly. Any agent touching those processes needs the same compliance guardrails your Workday apps already have. Now, this slide brings it all together visually. With Agentic Apps, you build directly on top of Workday applications and custom extend apps that provide the data and business logic foundation that's specific for your business. Our unified data layer, configurable business process, and always on security and compliance underpin everything you build. Custom agents sit on top of all of this, not beside it in some sort of silo. Agents are conversational layer on top of these existing platform investments that need all of your data and context to make meaningful action and execute effectively within enterprise guardrails. Now I'm going to pass the mic to Don who's actually gonna bring a lot of this to life with some live demos. Don, over to you. Alright. Thank you, Dhara. So I'll just give you a moment here to go ahead and start sharing. And if you can confirm when you see my slides. Yep. Sure. Okay. Awesome. Great. So so I'm Don Stacy, enterprise architect here at Workday. Been at Workday about eleven and a half years now in this capacity. So my job really is to talk to customers and prospects about technology. A lot of AI, a lot of extend for the past few years. It's a pleasure to be with you all today. Wanna start with just an example that we're gonna use as we move through this section of the of the presentation around rewards and recognition. This is a very common scenario because customers do it differently. We have a sample app that's out there, allows you to do feedback and the one time payment in one motion. For our customers that leverage projects, you can hook into that infrastructure as well because a lot of people give feedback directly related to projects, but it doesn't have to be project related, certainly can be. We see customers building custom extend apps that are leveraging their unique data. All that data that's sitting inside of Workday already plus any custom data they might want to collect provides the context, the rules, the logic within your organization, right? Which organizations can, provide these kinds of benefits, which ones need these certain approvals, etcetera. All of that can be delivered through these applications, ultimately connecting to third party solutions where necessary to pull data in, push data out. We're gonna see an example of that today. And as we move into this world with Sonnet Agent Builder, we're going to give you the ability to build agents directly on top of those applications. Instead of a form based UI, you may have a conversational UI or in addition to a form based UI, meet your employees where they are. Whatever experience they want, you'll be able to deliver. These agents aren't built off in some other silo somewhere, they are built directly within the context of your application so they understand your logic, they understand your security, your rules, your deployment options. We're gonna have those lawful actions that are gonna be there in terms of awarding nominations, who's allowed to get them, what is the maximum amount, etcetera. All of your rules and logic are at the disposal of this agent and it's running safely and securely within Workday. We are going to deploy it to the agent system of record just like as if it was a partner agent or a Workday delivered agent. As we work through the session here today, we're gonna see some live demo, a little bit of video where necessary for features. And the thing thing we're gonna do is we're gonna start with the building of an application. So let me go ahead and swing over to my browser here. And if I'm in a new space, I'm working with feedback, I'm not familiar with it, I need to figure out what the API calls are. So extend works extensively with APIs. And I can use developer Copilot for APIs. So this is the first of our Copilot examples here. And it's given me some suggestions. And then because I was practicing, I have my query already in here, find an API for Workday feedback. I'll make that query into the Copilot. It's gonna go out and it's looking across REST and Graph and SOAP and figuring out what are the right APIs that can help me achieve this job. It suggested three of them. Looks like we got two SOAP and one REST. That's not all of the potential APIs, but it's the ones it suggested so I've just expanded to see more. I can blow this out into a different page, see all the details. One of the things I like to do when I'm building something new like this is I like to go in and do a preview and compare. So let's look at this particular talent give feedback and let's contrast that with one of the other recommended APIs. Create feedback request on worker. We'll let that pull up. That happens to be a graph API as opposed to a SOAP API. This is what the response structure looks like. This is what the request structure looks like. These are the tools that help me build my application faster. So again, that is developer Copilot looking at, at APIs themselves. But let's go to the next level and let's drop into our application here. So we have the shell of an application already created for our demonstration today around recognition and feedback. And let's talk about Copilot in this context. The first thing is something that kinda gets missed sometimes but is really powerful and that is the ability to just have a conversation with the documentation. So I asked you know what is two way reporting earlier? How do I filter a list? It gives me all of those different ways I can use list filtering all the way down to the sources and I can ultimately get the documentation but boy am I not sad when I can just find out this way without digging through docs, I can just conversationally tell it what I'm trying to achieve and get insights. Likewise, I can ask the the system to build components for me. I can build scripts, there's a prompt library here with lots of examples, and we are going to ask it to build a page. So I'm just gonna copy this one, drop it in here, telling it what we want. I need a page, it's gonna be an edit page, I'm telling it some of the text I want, I want a list of values for the gift card amount, a gift card email message, then I'll ask the system to go ahead and generate that. So we're calling out to that Gemini model provided by Workday, you don't have to hook anything up to make this happen. It's using the context in this case of the prompt but it can also use context of my application. If I told it to just generate a page for a particular application it can look at the business objects and go ahead and figure out what those fields need to be and build the page. So it created all that code for me. We're off to a really great start. Let's go ahead and see this thing live. What does that page look like? So there is my feedback. I told it a rich text field. I told it a drop down with three values. It did all that for me. This would have taken me, even with our great drag and drop technology, several minutes to create something like this and Copilot just did it for me. When I'm happy with that, I'll go ahead and add that to my application. And I now have in my application that new give employee recognition page. So very easy to leverage that and we're gonna enhance that page here in a little bit. Before we do that, I wanna take you to another element of Copilot and I need to show you a video on this one because it is not generally available yet. We are targeting GA for June in the DevCon time frame so you'll be able to leverage this. And this is orchestrations. Orchestration is our new integration technology, drag and drop, very powerful, doing the things that studio used to do for us, and we are going to allow you to use Copilot to create an orchestration. This is hugely game changing as we move forward. So let me go ahead and start this and then I'm gonna pause it here and show you something that's a little hard to see. I know it's small on the screen there. Oops. There we go. So we're gonna pause that, zoom in a little bit. So we are asking it to create an orchestration that receives recognition award details, looks up the worker, sends feedback, issues a gift card through a third party called Tango, and creates a one time pin in Workday. It's a pretty heavy load. Even with drag and drop inside of Orchestrate, it would take me a while to set this up. I need a JSON response and I'm gonna give it the open API spec for Tango and let it just use that. For those of you who've played with, you know, whether it's a cloud code or a cursor, you're used to this kind of experience. We're bringing this directly into Workday. Here's the YAML file for Tango. Here's a sample input of what I'm going to provide to this orchestration. And then just a button that says go. And we're gonna go ahead and generate the full shell of that orchestration. Of course, we're gonna tweak it a little bit, but it's a really great start. But it doesn't end there. I can accept that. I can reject it. I can ask it to try again. I'm now gonna say, hey, I wanna log step everywhere. Again, I would have spent a few minutes setting these up. It just did it for me. And then the thing we all have to do at the end that we don't wanna do and that is produce documentation. Copilot's there to help us as well. So Copilot can go ahead with a request given here. Go ahead and document that orchestration for me. And what we're gonna get is a rich set of documentation. What's the purpose, the inputs, what's the flow, what are the outputs, what integrations are we calling, how are we dealing with authentication. This is a huge time savings. It will be great as we move forward with this and it becomes generally available, we'll be able to give you quantifiable metrics around the time savings achieved, but we absolutely expect it to be huge just because of how easy it is to create those integrations. Alright. Let's go ahead and pop back into our our present oh, sorry. I'm I'm already there. I wanna talk to you now about infusing AI into the app. We're gonna use AI widgets. So this is a capability that is generally available out there for those extend pro customers. Wanna show you that in action. So we're gonna go right back into our developer experience. We're gonna take that page that we have. We're in the UI where we can create and edit that page, and we are going to go ahead and add in some prompts. So I've got things like a feedback field, name field that are already on my UI, and I want to add an AI prompt. Click there. Give it a name. We'll call that refine or revise. Doesn't really matter. And then we need to put in the prompt text. So I'm going to go ahead and grab this guy, yeah, from my other screen so you don't have to watch me type. That's not fun for anyone. So refine the positive feedback for the employee using feedback field, which is the name of this field, and you'll make it more specific, ensure it's professional. So we're providing all that information, and that's all we need to do on that one. And then I'm gonna do one more for you, and that is there's another field to send a gift card like an email text. So gift card email message, we'll do the exact same thing again. So we will call this generate message. That's just the text that will appear and you'll see that here at run time in just a little bit. And same kinda thing, we have the prompt information. So all these prompts are just happening within Workday. We're calling Gemini, you don't need to hook up to anything, you don't need to provide any LLM access, you don't have to pay any tokens. It's just happening through the application. You don't have to pay any tokens directly to Gemini. So once we have that set up, we can go ahead and view this guy in action. If I've done my job well, very good. We now see that. So thank you for all the great webinar support. And let's refine that. So we're gonna call that large language model. It's gonna take that information. You get different responses. It's LLMs, right? It's probabilistic. So in this case, it's given us a bunch of different options. I see that pretty frequently, honestly. I'm just gonna choose option number two here. So I'm gonna delete the other text. I don't want all of that. And we'll go from here. Little shift control end. And we now have some text. Great. How much do we wanna give? Let's do €50 and then let's generate a shorter email based message. So congrats, the name based on amazing feedback. Here's a gift card, a small token of our appreciation. Enjoy. That looks awesome. We'll put that in and I'm done. So great thing about these AI widgets is you're just gonna be able to generate content, refine content, make it easier for your end users to produce things in the moment of your extend applications. Alright. So last thing we wanna talk about, as we move through the demonstration is now we've got this great app. We have some AI prompts in there. We've created an orchestration to talk to a third party system. It's ready to go. I could stop there but I don't need to stop there. I want an agentic front end. I want a conversational flow on top of that. So what we're using here is Sonna agent builder and we're gonna just see this in action from a video perspective. So first thing we'll do is say I wanna add a component. Actually let me pause and I forgot to call something really important out here. Notice where I am. I'm still sitting in the developer experience, the same place I was just working on my application, I'm now building an agent that experience. I'm not somewhere else trying to hook all this together. And within that same consistent experience within Workday. So we drag out an agent component. There are many components within this tool that we might drag out and this is a simple agent component. It's gonna be calling Gemini 3.1 Pro. I now need to tell it how to act. It doesn't know anything at this point. So I'm gonna give it a couple of tools. The first is a Workday provided MCP, tool that's gonna help it understand data. What can it pull from? What are those actions that's available through this MCP server? And this is about understanding what's in my application. So I set that up and then the next thing I'm going to do is add another tool that is tied to the orchestrations. What are the things I'm going to allow it to do? So what can it see? What can it access? And then what can it do? And so these are the orchestrations that are part of this application. We're gonna choose three of them in this case. And then the very last thing we do before we're ready to test this thing out is go ahead and give it the instructions. Tell it what we want it to do. You think about prompt engineering in this case. So you're a helpful recognition agent ready to provide answers. Here are the specific things I'm going to enable you to do. It's not gonna run rogue. It's not gonna be doing things we don't expect because we're guiding what it's allowed to do. We would would then have clicked a button to push this to Workday agent system of record. This is where all Workday delivered agents live. This is where third party agents can live. This is also where agents that have nothing to do with Workday can be registered so that you can see the totality of your workforce. All your humans and all your agents in one place showing up on an org chart perhaps, seeing all those together. So we now have our agent, it's in the list of unregistered Workday agents just like any other agent that hasn't been registered yet. We will register that, we'll pop in here. We don't have it in here but we would dig into the configuration. What security groups are allowed to see this? You go build in Claude or somewhere else talking to Workday, you're gonna be talking to APIs, you're not gonna have context, and you're gonna have to build all that security infrastructure yourself. By building within Workday, that really just comes along for the ride. So let's see this thing in action. We've set it up. We're safely within agent system of record. Where do we access this from? The same place you'll access Workday delivered agents or partner agents. So go in and choose the employee recognition agent, and we're just having a conversation. We're asking it to do things for us. So who are my top three performing employees this year? It's calling my application, my custom application, and it gives me a list of those individuals, suggests some next questions. Yeah. Sure. Show me their bonuses for this year. That's one of the skills that it has. And then lastly, you know, would you like to give feedback? Sure. Let's give Liam a $200 bonus in a little text that says thank you for all you do. At that point, the orchestration's running. It's going out and it's calling Tango. It's pushing that out. It's updating payrolls, doing all the things we asked it to do, and giving us feedback that that task is complete. So you can do this kind of work that you saw there on top of just about any eight, Extend app that's out there. There are over 3,000 apps in production now. Customers are using Extend significantly. It's been the fastest moving SKU for quite some time at Workday. Very powerful technology. So just think about the applications that you have and would it Would you benefit from a conversational front door? There may be an application you wanna build that has nothing but a conversational front door and some admin pages. That will be up to you from a design perspective. And then of course these apps span HCM, financials, everything we do at Workday. We've seen the top three on the left there. Custom agents through agent builder. That will go generally available targeting twenty twenty six r two, You will get to play with that at DevCon. You saw custom apps through through the app builder experience as well as the orchestration builder. Orchestration builder is generally available targeting June for that. That's all using that same copilot. It's using that same developer experience. We wanna be incredibly consistent so you don't have to think about a bunch of different applications. And then it's all leveraging the APIs and connectors within Workday focused on open standards, rest, soap, graph, but also evolving into new things, MCP, a to a. We're supporting all of those for our agent access. We're providing Apache Iceberg, what's that for? That's for that upper right box there called data cloud. We're not getting into that today. Also targeting GA for 2026 R2. The data cloud is talking to our key partners that are in the big data space. Namely it's Salesforce, Google BigQuery, Databricks, and Snowflake. Allowing you if you wanna see some data from one of those platforms inside of Workday for reporting, we'll be able to report on that without ever moving that data. If you wanna see your Workday data outside of Workday, we'll be able to report on that from other platforms without ever moving that Workday data. Keep your sensitive financial and HR data where it belongs inside of Workday. You're gonna cut down on your integrations. You're gonna improve your security by not moving a lot of data around. So big changes coming around data cloud, around the custom agents all in the DevCon time frame for testing and then GA 26 r two. Two more slide one more slide for me before I hand things back. So when we think about this agentic space, you saw we've got it looped in in orange here. That is the agent builder, that we just saw in action. And that is going to live within an ecosystem. There's the workday delivered agents, the partner delivered agents, the ones that you're gonna build that are incredibly specific to you, all being governed by the Workday agent system of record. Inside of agent system of record is what's called agent gateway. That is where MCP and a to a functionality comes in. So we're moving forward with those, evolving standards. All of that sitting together on top of the Workday that you know and trust, as we move forward. So thanks for your time. Dhara, I'll go ahead and hand back to you. I hope. Yes. Hey, Don. Thanks so much. Everyone in the audience here today, I know there is seems to be an audio bug with Goldcast, so I apologize. Whenever speakers switch, make sure to click the unmuted button. You might have to refresh your browser or when you right click on your browser tab, make sure to click unmute site. Hopefully, that works. I apologize for this little bug that we're dealing with here. And Don, while I still still have you for a moment, I know we have a pretty big audience here with us today. Just one bonus question for you live. Is there any advice you'd give to anyone here that's interested in everything you said and trying to dive in more before all of our technology is out? Yeah. Yeah. I think certainly so. Sana Agent Builder is a technology known as FlowWise that we acquired last year, and there is a version out there that you can tinker with. So I would go tinker right now. Now that does not mean the exact the exact capabilities there are gonna make it into our product. We're gonna add some things, we're gonna remove some things, but go play now. Drag some things out, tinker on, that's what we do, that's how we get comfortable with it. So that's one thought. Second is start thinking about your apps now. What are the apps that you've already built and extend that would benefit from a conversational front end and you can start to plan ahead so as the technology becomes available you'll be in position. And then the last one is go to DevCon. I don't get to go because of some personal travel. I'm incredibly bummed. You're gonna get access to data, data cloud. You're gonna get access to the agent builder. It's gonna be an incredible DevCon. So for those who can't go, I'm jealous. If you weren't planning to go, you should probably plan to go. Awesome, Don. Thank you so much. I love seeing the demos in an actual use case that resonates so much with our audience. I am gonna keep us moving. And for this next section of our session, and really quick, I'm going to take a pause again. As these speakers are joining us, please make sure to unmute. I'm sorry about the bug that we're dealing with with Goldcast right now. Try right clicking on your browser and clicking on mute. But this next section of our session is the part where I'm really happy that we get to highlight all the core pillars of Workday build. We're, of course, going to keep talking about extend. But I'm really excited to hand things over to Clint and Chris because these two aren't just certified experts in our partner ecosystem. They're also well known and respected figures in our developer community. Between them, they bring years of experience, not just building with extend, but also teaching others how to do it well. They live and breathe this stuff, so I can't wait for you to hear their insights. Chris and Clint, take it away. Alrighty. Hi there. My name is Clint Harwood. I'm here with Chris Humphrey. I'll let Chris introduce himself. Hi, everybody. I'm Chris Humphrey. I'm, well, the the slide says I'm an extend archmage. I haven't gotten that official title yet at work. I'm working on it. My title is, director, Workday Extend. I've been in the Workday ecosystem for over ten years. Previously, I was, at Workday and field readiness, field readiness architect for Extend. So I've kind of lived and breathed Extend for several years now. Clint? And my name is Clint Harwood. Many of you know me from having been an extended instructor for quite some time before I moved over to field readiness as well. And, you know, between Chris and I, we wrote the workday extend style guide and a bunch of other fun stuff. Check out developers and dragons if you haven't already. Now we're gonna be a little more focused on stuff you can do right now. I thanks to Dhara for, like, pointing out all those cool things that are coming. We're gonna talk about the stuff you can do right now, right here. And I want to highlight a couple of really key points, which is, first off, what extend does well. We're gonna get into, okay, why you'd probably be interested in extend pro, but I wanna dig into, okay, what does Xtend actually do well? Why would you want it in the first place? And the first one is that spreadsheet. So you know that spreadsheet, the one that, you know, sits on Mary Beth's laptop that you can't close payroll or do your month end close without getting that one piece of information that's sitting on their desktop? Yeah. Yeah. Extend is really good at that. Extend is good at taking that spreadsheet off of that deck top, off that last laptop and putting it all the way up into the cloud where it's visible, auditable, safe. There's so many awesome things that you get out of moving things off of those spreadsheets and into extent. It's a great place to go. The next thing that we see really commonly are Workday simplifications. Workday is an awesome tool. We all everybody here has at least some positive feeling for for Workday, I hope. The thing is, Workday needs to build a tool for, like, 10,000 customers, which means whatever they build has to work for 10,000 customers, not just you. So there's usually a ton of things inside there that, honestly, your org may not need. And one of the really common extend use cases is to take that big, heavy, all these options process and trim it down to something nice, neat, and small that your manager self-service team can absolutely pick up and totally run with. And the last one, I'm gonna go a little off script in this last magic is in the intersections bit because not two hours ago, I was talking with the company about the magic in the intersections and actually used the phrase. They're an organization that has, like, Prism and Adaptive and Workday HCM, and they wanna do some forecasting. And I was telling them what the joys of the forecasting API and the a AI gateway. And what they needed was they just needed a nice clean place that would allow them to take all that information and put it into one spot where they could do the forecasting. And extend is a perfect place for that. It's a place where you can take all these different disparate pieces of information, all these different processes, bring them together, and show them to whoever you happen to need to show them to. Now I wanna talk about a couple of the apps we're actually building right now, which is always the fun part. Like, we're building apps for real people. And one of them is, you know, to springboard off of, do people ever actually need to present data in Workday that exists elsewhere? And the answer is yes. So right now, we've got an organization that is that they build a bunch of financial information. So they take all that finance information that's in their wonderful application, but all the people that need to see the summary information are over in Workday. So I've got information over here. I've got customers over here. How do I put them together? Extend. Extend is a great tool for this. Now there's another really cool tool that we're building that is a a really good simplification of an IT landscape. This allows an organization. They get information from a third party. And right now, they feed it into a system in their IT landscape, and then they take it out of that system and bring it over to Workday and create POs and invoices and do all sorts of, you know, cool Workday finance things. But what's that third system actually doing? All it's doing is complicating the world. So what we do is we take the data directly into Xtend and then allow Xtend to create all the Workday artifacts behind the scenes. So this allows us to build these really cool use cases in something that looks and feels just like it's Workday. And, I mean, I gotta at least show you one screenshot. Come on. So this is an app that I can actually talk about, show you a screenshot, and this is one that we've recently built. And the the story behind this one is kind of fun. So a really long time ago, I was a manager of HRIT at a large college up here in Canada. And I'm sitting, I'm talking with a bunch of my employees very first day, and every single one of them said almost the sum same line, some version of, yeah. My your predecessor keeps track of that on a on a spreadsheet. About the fifth time I heard that and had to tell them, yes. I don't have the desktop, and I don't have the spreadsheet, I realized the depth of my hatred for spreadsheets. You know what spreadsheets? They they they they hurt us. They burn us. And this application allows you to flexibly collect approvals, query workers, do all sorts of these really cool things that will allow you to at the manager level, give this to your entry level manager folks and allow them to talk to their workers. Now I'm about to engage in some foreshadowing here. Foreshadowing is your sign of quality literature. So the foreshadowing is this. Let's say I took an app much like this one, and I rolled it out, and I gave all my managers the ability to do surveys to their workers or start collecting, approvals from different processes that they can define themselves. Getting all those spreadsheets and all that stuff out of OneNote, you know, all the the crafty stuff that's sitting in OneNote. Your your OneNote, my little OneNote, they probably look the same. But you take all that stuff out, you put it up in this extend app, but the problem with this is, if it is a problem, this is like potato chips. Like, once you have the first one, then you realize, okay. This is awesome. I'm going to have them all, which starts to really cause a lot of extra data volume. Like, this this can really result in a ton of data volume. So let's talk a little bit. Let me go up one more slide and hand this over to Chris. Let's talk about some extended professional features. Yeah. So we wanted to cover some of the things that we love about XtendPro. And just to kick it off, like, if you were back and remember the olden times back when we, struggled, with the limit where we could only have 1,000,000 objects, which maybe may may sound like a lot. But a million object instances for your extend application, it's not gonna cover a lot of those use cases that you might wanna tackle. Now more recently, that limit was, increased to 10,000,000, and that opened up an incredible number of use cases. But I still probably don't wanna be tackling my, Fins cases or my time tracking. There's some Fins cases, obviously, you can cover. Suppliers, you know, invoices. There's a lot of the stuff that's really heavy volume. And with ExtendPro, we have the ability to have a 100,000,000 records, and that unlocks an incredible number of use cases that we couldn't tackle before. And if you marry that with Prism, you have this ability to load all the historical data as you go, and you you keep ahead of ever hitting that limit. And it just it makes for an incredible feature enhancement that, like, honestly, this one alone, in my opinion, makes pro worth the value. Right? And we have some examples here on the slide too. Right? 10,000 staff with, a single daily interaction per staff member, you know, three and a half million instances. You're nowhere close to the limit there. Right? Of course, there's design considerations when you're building apps that are that large. We definitely want don't wanna try to build an app that is, creating a spreadsheet, in your extended application. And don't get me wrong. We love spreadsheets. Right? Clint's not saying spreadsheet's bad. I like hammers. They're very useful. I'm not gonna use it to change the tire on my car. Right? It's all about using the right tool for the job. So 100,000,000 records, it's gonna unlock a lot of value for you. Clint wanted to say a few things about the AI widget. I wanted to talk about this one. So you you saw Don demo this particular AI widget a bit earlier, so I'm not gonna demo it. But I wanna highlight that this is a very cool thing. Like, this this is a line. So all of you, this one here is absolutely the line where we're now getting into what people think of as AI. This is going to be this widget, this one right here, mark my words, you put this one on the calendar, this is the one that's gonna start drawing AI out of the shadows. Why? Because this is now gonna allow your people to be generating that text inside the extend application where, previously, they were going out to chat g b t, typing some stuff in, copy and paste it over. Now they don't need to break that. They can just be right here, and this is awesome. But it's also, like, every good hammer, you know, you can hit your thumb with it. So what I want you to think about here is I want you to sanitize your inputs. For those of you who are x k c d fans, little Bobby table still applies almost twenty years later. And the what this means and this led to, like, a spectacular conversation between Chris and I when we first saw these things, and the conversation was about governance. The the awesomeness about the governance here is what are you willing to share with that LLM? So what you saw earlier in the demo is you can put something in. You can just type in whatever you'd like. However, because this is an extend app, we can control that if we want to. So now you'll notice in this particular screenshot, those of you who are sharp eyed, you can see that that safety first, that's an instance list. I can lock down exactly what gets sent out to that LLM and only send very specific information if I want to reduce the amount of flexibility that my users get, but increase the amount of safety that they live with. This led to a fantastic governance conversation about, well, what is the risk reward that we're willing to put in here? And if you're gonna be working with these AI widgets, what I want you to do is this is now the time where I want you to start having this governance conversation. What stuff do we wanna share? What stuff do we wanna keep in house? What stuff do we wanna let our users put in there? This is going to be the the feature that absolutely drives that conversation. This is the one where you need to have it, and you need to have it right now. Alright. Data readiness. So let's let's, let's talk about what what the whole topic of this this webinar is. Right? How do I prepare for creating AI apps and and AI agents in Workday? And this is one of those funny ones that you'll see the the the meme on LinkedIn. Nobody's talking about this. And I feel like everybody's thinking about it, and maybe there are some conversations happening around this. Right? Hopefully, there's a lot of conversations happening. The thing is the technology when you're building an agent, that stuff, anybody can do it. Anybody can learn it. I feel like it's pretty, you know, straightforward solving that part of the technology problem, But it's not always the right answer. And even if it is, I think there's things you need to do to prepare before you even start, you know, implementing anything. Right? So data readiness, what I mean here, there are some key pillars we wanna talk about. Is your data clean? And what does clean mean? So to me, when I think about clean data, I think about consistent structure, right, and consistent data and, the data being trustworthy. Right? So some of the examples where I would say the data is not clean. I have duplicate records in my system. I have inconsistent naming and categories. So let's say, for example, for states, in The US. I've got CA and OR and, you know, TX, rather than, you know, the full names in one area. And then in another area of my data, I have the full names, California, Oregon, Texas, etcetera. That's inconsistency. Right? Those are the kinds of things we want to address. Perhaps you have fields that should technically be required, but they're empty in the system because they were empty when they were loaded in. Right? That's a problem. Contradictory values. Right? If you think about policy, old policy versus new policy, which one gets followed? So those things need to be fixed before we apply AI to it. And, honestly, these things are problematic for humans too. Right? It's not just an AI problem. When we think about data and the data conversations have been happening for decades. Right? First, it was, you know, just computer systems and analytics. And then the big data wave came, and we started having the conversation again about, like, data cleanliness and readiness. And here we are again with AI, and it's being amplified because the AI is not just summarizing, it's now taking action. And we don't want it to take action on unclean data because that that dirty data, is going to lead to, bad assumptions and inferences. Right? Data relevance. So let's assume that your data is clean. It's immaculate. It's the most pristine data of any system in the world. Is that good enough? Maybe not. Is the data relevant? Is it fit for purpose? Okay? So let's say, for example, your company wants to open a new store in a new region, and you've got tons of data available, sales history and inventory, staffing patterns, but not for that region. So how can you utilize that data to make a decision? And better yet, how's can your AI make a decision? Right? As a human, maybe you've got some more context and reasoning ability that they I may not have, and you can do an analysis. So that, well, this isn't relevant to that region. Right? The AI agent might not be asking that question. And lastly, assuming you have clean data and it's relevant, you have the data about that region you need, is it fresh data? Right? If I have the staffing information and and, you know, the the market, pay information, cost of living information for that new region I wanna move into and it's twenty years old, well, it's clean. It's relevant, maybe. It's not fresh. And, again, as a human, maybe I can put some reasoning and say, oh, this isn't right. Right? This this doesn't feel good. There's a problem here. Your AI agent might not do that. And so, you know, I'll and it's gonna be important to have humans in the loop when a lot of these decisions are made as well. But the human in the loop, might not even work for you as if it's surfacing the data up and not providing you the, you know, the fact that the data wasn't clean and it made some assumptions and it filled in some blanks for you or that it wasn't fresh because maybe you're not seeing that. Right? So I feel like we need to prepare and make sure the data is ready as we move into this new age where AI and agents are becoming such a a big part of of the machine as we operate. Now I did wanna give some practical tips, for extend when you're designing applications. Things I don't know if an if a lot of people are thinking about right now. When you go in and and you do your design and you're early in the process and you're thinking about your data modeling, right, you're thinking what fields do I need to to have, what data do I need to capture to, meet the business case and and, you know, have those user flows work the way we want and the analytics that we're gonna wanna run. Are people having a conversation? How is this data gonna be used by my AI agents in the future? And there's a couple things we can think about that, you know, are, I think, easy wins. Workday recently enabled, rich text fields on model objects, which means you can have markup. You can bold and make lists and and all that sort of thing. Right? Which is great for a human analyst who's looking at the data. Right? It's not so great for the AI agent, though. It's gonna make inconsistency, and the extra formatting is just extra data that's nullities. Right? So maybe think about having a secondary field that's more structured, and it's going to be the data that an agent can take advantage of. Right? Another problem we'll see sometimes where numbers are stored in a text field. Well, unfortunately, there's no enforcement at the model layer that says this has to be a number. Right? There's nothing I can do as a validation at the web service layer. You have to do it in your code, which means somebody could use an EIB and slip in, an a, b, c in a numeric value and suddenly you have unclean data. You're introducing a potential problem there. So make sure that the field type is fit with the data that's gonna be stored. And there's plenty of cases where, like, having numbers as text makes sense. There's reasons for it, but think about that. Just be careful and be thoughtful as we're moving into design where we're thinking about agents and AI. This was one. I think we have the the slides swapped around a little bit. Another pro feature we really like, developer copilot for API search. It's incredibly handy and documentation search. This is stuff that Don talked about earlier. Don't sleep on these. These are fantastic features. Alright. I'm gonna wrap up here. I'm gonna talk about this one real quick, and then I'm gonna hand it over to Clint. The developer community is fantastic. There's a monthly developer program update call. It's gonna keep you in tune with what's happening, where Workday is moving. You're gonna get, you know, tips and tricks, and you're gonna get the Workday team there showing it off, what's coming, how we're doing things. It's fantastic. It's a great opportunity to stay in touch with everything. The dev forums. If you have problems when you're developing your apps, you need help, people are will trip over themselves to help you. It is a fantastic way to get questions answered and to share things that you've learned. Make sure you do that as well. Right? You gotta reciprocate. We gotta keep the community healthy. Finally, LinkedIn. There's a lot of good stuff out there. I heard rumor that there might be a a post coming today about a a podcast interview with, yours truly. So keep your eyes open if you wanna, take a look at that. Good stuff. Great, great, opportunities there to, you know, learn and build yourself as a developer. And, of course, don't forget our, you know, shameless plug for our wizard sessions. Chris and I run a once monthly wizard session. The next one's gonna be on agent system of record, so there should be some fun had at that one. And the last thing that I do wanna talk about, as Don said, I'm gonna echo everything you said. If you're not going to DevCon, go to DevCon. Seriously, it is that awesome. It is that useful. It's the time that you get to see everything. It is literally my favorite conference out there, and I'm definitely going to highlight, so there's all this stuff. You get to play with these things, and it it's hands on. It's not, you know, us talking up at the front. DevCon, you're playing with the tech. That's the cool part about DevCon. The other cool part about the about DevCon is brain dates. Like, I I'm gonna be there. Chris is gonna be there. We put out brain dates every year. I I've been brain date champion at, like, two DevCons. It's having a hallway conversation with a total expert. How awesome is that? So I really do think that DevCon should be on your radar if you haven't already got it there. Dhara, I believe you're up next. Last item on DevCon is that there is a twenty four hour hackathon for those of you with a bent. And even if you're not trying to be competitive with that hackathon, I'm gonna recommend think about going because how often you get to sit, play with all the new tech with a ton of experts who have nothing else to do but sit there and help you. They have. an individual track this year. You don't have to compete. Yeah. Absolutely. Go have fun. Thank you so so much, Clint and Chris. I loved all the insights you just discussed, especially all those stories about the spreadsheets and those reminders about governance and data readiness that are more important now than ever before with AI. And, seriously, thank you for continuing to be such huge parts of the developer community. I know that we discussed a lot in this session, so I wanted to leave you with three big takeaways. First and foremost, Extend Professional is your key to building with AI on Workday's platform. It doesn't just supercharge developer productivity. It unlocks entirely new levels of functionality for your end users too. Second, agentic apps unlock the next level of impact. Because remember, an agent is only as good as the data and context it sits on. By building on Workday, you're ensuring those agents act lawfully within enterprise guardrails from day one. And finally, as Clint and Chris said, join us at DevCon in June to get hands on with our technology. If you're interested in learning more about building apps and agents with extend or all of our other build tools, DevCon is the place to be. You can hear how others are developing with our build tools in the theater section sessions and very live demos. And there's so many ways to get hands on and test the newest tech like Sonna Agent Builder. You can try building an agent within our expo, Express Labs, and especially our hackathon. And if you have any devs on your team that are just starting out with extend or some of those other build tools, we've got a beginner's track at Hackathon this year too. It's the perfect opportunity to really dive in and learn alongside peers. As Clint and Chris mentioned, there are so many experts on standby to help you. DevCon is truly a Workday event like no other, so I really hope to see some of you there. And that brings us to the close of our session today. Thank you all so much for your time and take care. Hope to see you at DevCon.