Name: Workday Rising Recap for IT Leaders & Developers
Uploaded: 2026-10-29T17:31:45.870Z
Duration: 48 min 4 s
Description: Workday Rising Recap for IT Leaders & Developers

Transcript for "Workday Rising Recap for IT Leaders & Developers": Good morning, everybody, and welcome to the rising recap for IT leaders and developers for Workday Rising twenty twenty six. We are so excited to have all of you here with us for the next hour today. We have a lot of fun things on the agenda and the docket. To start, my name is Hunter Collin. I'm a principal product marketing manager here. I'm gonna be playing the MC for the next hour. I'll also introduce the first segment of our agenda today and the rest of our speakers, but thank you so much for joining us. So along with the content I'll be, speaking to you about today, I'm joined with Ashley Green, a principal product manager, who'll be talking to you about developer agent and building custom agents on the Workday platform. I'm also joined with Jolie, another principal product manager, who will be working with Becky Newborn to be talking about the new agent harness and how we create lawful governed agents on the Workday platform. So let's go ahead and jump in. So today, this webinar is on demand, so it will be recorded, and you will get an email of the recording within twenty four hours. So if you feel like you missed something, you wanna go back and double check, you wanna rewatch the segments, you can go ahead and do that. You can also ask q and a at any time. There's a bunch of people backstage ready to jump in and answer any questions you have about the subject matter today. You can also tab over to the docs section where you'll see different downloadable content relevant to the materials we're sharing today. And then we'll also send a post webinar, survey for you all to fill out so you can give us a little feedback on the content, what you're excited about, and, what you'd like to see next from us. This is our standard product statement, so we'll leave this up here for a moment. And the agenda for the day. So we're gonna do a quick Workday Rising overview. I'm gonna talk to you all about open by design, how we are creating a headless future for operating and connecting your AI agents to Workday. Then we're gonna talk about how we are backed by builders, how you can start building and deploying agents on Workday, and then we'll we'll sort of end it with lawful by default. So these context guardrails governing your agents. And then if we have time at the end, we'll have a fireside chat and some more direct q and a. And then at the very end, we'll talk about what's available and how you can start building, after today's webinar. So a little recap on rising. This was the biggest rising ever. We had over 28,000 people on-site with us in Las Vegas, Nevada. Over 11,000 people joined online. And over the three, four days of the event, our attendees built over a thousand agents with us. We had a bunch of different, like, on-site booths where they could sit down with experts and build agents. So a lot of really great hands on hands on experience deploying and building agents on on Workday. So the core message, what we shared at Workday Rising, is that Workday is transforming from a system you log into to a platform you build on. This is related to what we announced at Rising, which is the Workday trusted agent platform. We wanna become... Or what we are becoming is a platform that's open by design, allowing you to connect, to Workday with the agents your team use every day, backed by builders. So supporting your builders with the toolings, the frameworks they use every day to build, great solutions for your teams. And then also Lawful by default. You expect Workday to be, the best custodians of the most important and sensitive HR and, finance data. So we wanna allow you to do these agentic tasks without you having to sacrifice any type of security or compliance to do that. And then you'll see that within this sort of, platform, we have four buckets. We have a data cloud, which we won't talk about much today, but we have more content on that in the future. We have agent harness, which is what Becky and Joe will talk about. Agent governance and then extend, which is, of course, how, you access the developer agent and how you build agents on the Workday platform. So I'm gonna start talking to you about open by design. We know that you might already have an AI agent that you've rolled out to your teams that you love using every single day inside your enterprise. Maybe this is Copilot. Maybe this is Gemini, ChatChippy, Cloud, whatever it is, something you built by yourself. Our philosophy is that we wanna give you the opportunity to connect those agents to Workday safely and securely without sacrificing the user experience that you and your teams have come to love using these agents. So, what this means is now Workday is fully headless. We are supporting a fully headless strategy, allowing you to connect and build on our platform in all of these different ways, the ways that, sort of you... You've you've become used to and you... You're... You you prefer, based on how your business runs. So to articulate this open by design, example, I'm going to play the role of Maya. She's an engineering director at Acme for the HR and IT team, and I'm gonna show you two examples. One, I'm gonna show you how Maya has maybe connected her AI agent, in this case, Claude, to Workday in the past in a lawless way, And then I'll transition to showing you how, Maya can now connect her AI agent, Claude, Gemini Copilot to Workday in a lawful way without, again, sacrificing the user experience operating in, in Claude. So I'm gonna go ahead and share, my screen here and jump on over to a quick demo. Alright. So we're we're hearing Claude. This is the preferred AI agent for the Acme Corporation. And what Maya has done is created a custom MCP server. Now this is an MCP server or a model context protocol server that is not endorsed by Workday. It's something that her team built themselves and connected to Workday sort of outside, our normal business processes and security user groups. So if we open up this connector, we can see that, there's a few tools in here. Some of them are read, some of them are write, but definitely enough to be dangerous. You could see that there's tools to change, change jobs, to to get job postings, and to do basic actions inside of Workday. So we'll look at an example where Maya asks Claude in this example to show me her open job reqs. So it goes... It looks at all of the job reqs in Workday and returns 30 open job reqs, to Maya. But if you notice, there's something interesting happening here because Maya is an engineering lead, and we're seeing roles come back for things like account executive, regional sales manager, financial analyst, roles that clearly don't fit into Maya's core competency. So you have to wonder why why is that? So let's let's try something else. Let's move, Casey Wall onto my team as a senior engineer. So it goes... It looks for, Casey inside of Workday, and then it says done. Casey Wall is now on my team as a senior engineer effective today, which at the base level, that's a that's a strange thing. There's very little business processes, and workflows set up by our customers that just allow you to, you know, just automatically move someone over to your team. That's that's that's not a that's not a normal, action. But if we open up the reasoning chain here, we can also see that Casey Wall in this, in this example, is an account executive in enterprise sales. Now why would Claude allow us to move an account executive to enterprise sales? Well, the reason being in this example is because it's a custom MCP server that the Acme customer corporation built themselves. It's not following the same business rules and processes that are baked into your your your Workday tenant. So in order for them to follow all of these business rules in this example, you have to rebuild all of these business processes outside of Workday and around this, this quad example here. It's a ton of work. And also this example, clearly, Acme Corporation didn't do that. So we've done something potentially pretty harmful, which is operate outside of our normal business processes, show job postings that are not relevant to this individual, but then also move someone to a role that they they should never have been moved to. Now I'm going to to jump over and show you what it looks like to to do something like this in a in a in a lawful example. So we're back in Claude here. But instead of Acme Corporation building a custom MPC sir... MCP server, instead, you can see here that they're using the official endorsed Workday connector, which is available, today for you to take advantage of. So in this example, when Maya goes to hit connect, she needs to log in to Workday. And then if we open this up, you'll see there's a bunch of other read and write tools in here. I think around 400 tools in here as well. All of these are, of course, built by us at at Workday, and then inherent the, security user groups as well as the business processes that you expect when, taking action inside of Workday or when your agent takes action in Workday. So we'll go through, and let's look at the same example. So I'm gonna ask Claude again to show me my open job reqs. Claude's going to think for a moment. You'll see that it starts using the official Workday plug in here, and then it returns 16 job reqs. And if we do a quick check, these immediately look a lot better. They're all in the engineering family. It looks like we are properly returning, the job requisitions that are, that are assigned to Maya. We're not returning any account executives here or anyone in finance, so that's a that's a great start. Then we'll go to the next example where we'll ask to move, Casey Wall onto Maya's team as a senior engineer. Now, again, it's looking through not just, employee data here, but also looking at, who this employee is, what department they're in, and following, the business processes that your organization has set up. So you'll see Claude return here that it submitted the move for Casey Wall, the senior engineer, onto the platform team to join the team. But, it takes effect once it's approved. So the move was submitted, but it's not completed. We didn't... We could... We can't actually just go and move Casey onto our team. One, we know Casey's an account executive, but but two, even if Casey wasn't, there's a process to this. There's... There are other things that have to happen before we can simply move Casey onto the team. So this is a a simple, but I think important articulation to compare a lawless example where we're using an MCP server that is not endorsed by Workday, that is sort of custom built by someone at an organization or or or by a partner, let's say, versus a lawful example where, this is an official Workday endorsed plug in that has tools that are built by our team that follow the business processes that your that your organization sets up, in order to complete the task appropriate within compliance and with the right security frameworks. Now I'm gonna switch back to slides for a moment here. So what what I just described there is what we call agent ready tools. So these agent ready tools have been in, early access for the last few months, and we made them officially generally available at Workday Rising. So the ability to connect your agent of choice, whether that's Claude or Gemini or Copilot, to our agent ready tools is is generally available in a secure way, just like we just like we saw. So not only is it less work, but it's also, you can have more confidence that these agents are gonna be taking the right actions, within your your right, policies and processes. And then next, what we announced at Rizing, what we're really excited about is we have over 400 GA tools today. So these tools allow you to do individual things inside of Workday. So we have 400 read and write tools today, and we made a commitment to releasing another 500 in our next release wave, where more... Nearly half of them are are rights so that you, from your agentic tool of choice, can do more and more inside of Workday, in a safe and secure way with, again, without losing the experience and the speed you get from operating in a cloud or Gemini, or Copilot. So this is just the start to our webinar today. I'm gonna hand it over to Ashley Green to talk about our developer program. Wonderful. Thanks, Hunter. Alright. Hello, everybody. I'm Ashley Green, as previously mentioned, one of our product managers here at Workday. And I have gotten to spend the last couple of months working on developer agent, which we demoed and, delivered at Debt Rising this year, and so that's what I'm here to talk about. It has never been a more exciting time to be a developer. There is no more writing code line by line. Instead, our developers are executing swarms of agents to get all of their work done. Instead of taking weeks and weeks to get through projects, they're actually getting a lot done in minutes and hours, spending more time with the business and less time writing that boilerplate code. But there's another side to this. Right? As Hunter just showed us, headless done wrong is really easy and can be really bad for your business. And all of those developers that you have at your companies, or maybe your one yourself, are inheriting all of that lawlessness. They're being handed those vibe coded apps and being asked to rush them into production. They are watching agents descend into the system, and they have to make sure that those agents aren't jeopardizing your business. And they're doing all of this while reapplying those business rules and security that you have spent years cultivating inside of Workday. That is accidental architecture, and that leads to tech debt and and risk. And so developers like Maya face a really tough struggle every day. They're stuck in between speed and trust, moving fast or making sure something is secure. And anything that gets in the way is just friction. And so the challenge to us is really clear. Right? Remove that friction, gain that speed, keep the trust. And so that's why we took a clean sheet of paper to our developer platform earlier this year and reimagined it around three key principles. The first, build from any tool. For too long, we've kept Workday out of all of your favorite tools to build in, including all those agentic ones, Claude, Cursor, anti gravity, and maybe something that is gonna launch today, right now that's the next popular tool. So we are moving away from Workday specific proprietary languages, formats, and leaning into all of those open industry standards that your teams already know and love. Second principle is let your... We should be able to let your stuff run inside our Workday guardrails. We've made it really difficult in the past to have Workday be integrated into your solutions, and... But we know that having your solutions run on the right data and the right context is critical to have those trusted solutions. And so that's what led us to opening up our agent platform. So now you can build your own custom agents, have them run right inside Workday natively alongside all of the ones that we as Workday provide to you. And finally, anyone can build. For too long, our specs, our tools, our documentation has been kept behind walls and under lock and key, but no more. Now we are opening up our platform so that anyone across the globe with an email address can come and build with us. And So we're bringing our developer program everywhere to attract that top and emerging talent. All of this is really fun to talk about, but it's actually much more exciting to see it in action. So I'm gonna go ahead and jump over to a demo. Like this one. Here we are on the newly reimagineddeveloper.workday.com. As you can see, we have our open API specs, our MCP specs, all of our documentation right at your fingertips. Go ahead and check it out. But it's actually never been easier to create an account and get building, so we're gonna go ahead and do that. So you can see here I'm using a pretty basic email. Right? It's just ashley@gmail.com. I'm not using anything affiliated with my company or job. It's all just me on the Internet. And ta da, we have made it to probably a page that is hopefully familiar to many of you, but if it's not, it... This is the start of our developer site. Historically, just to get to this page required sales conversations, contracts, partner agreements, but now anyone can go and just create an account and get building with us on the developer site. Now, of course, to actually build anything in Workday, we do need a Workday tenant, and so let's go ahead and set up our own personal Workday tenant. We have been re architecting our tenant infrastructure over the last couple of months and kind of rolling it out so that our tenants come up in minutes instead of hours, as you can see here, and are already agent ready. So I've got my account. I've got my tenant. I need one more thing before I get into building, and that is my good friend, developer agent. So developer agent is our agentic companion to help us build on top of Workday. And as you can see, I can use this from a bunch of different places. In this example, I'm gonna use it from Cursor because that's my preferred place to work. Okay. We're inside Cursor. I've got my good old friend developer agent, and I'm gonna go ahead and give it a prompt. In this example, I'm gonna build an internal mobility agent so that it can look across my open requisitions and across people available in the company and give me recommendations of someone to move into my roles. Let's go ahead and give developer agent a prompt and get it going. Let's start by looking at this prompt really quick. It's very simple. I'm not telling it exactly how to build anything. I'm not saying use this API. I'm not giving it really deep instructions. It's very basic. It's my requirements. It's the business outcomes I'm looking for. And then developer agent is off to the races here, building all the different components that I'm gonna need for my agent to be successful. And then also, as you can kinda see here, it's connected to my tenant. It understands what's available to me, and so it's actually understanding what needs to happen for my agent to be successful. Developer agent realized that I didn't have all the... Quite the right data model inside of Workday, and so it's extended that Workday data model by giving me an additional business object, business process, and security domain in order to actually have my agent be successful alongside building all of the skills. It's validated in my app, and it's deployed it to my tenant so we can test it out. But before we do that, let's take a closer look actually at one of these agent skills. So this is an agent skill. It is built on top of the industry standard skills.md framework. So nothing special, nothing Workday here, all just industry standard tools. Plain language, really easy to follow and understand. Then these are those same agent ready tools that Hunter was talking about that are being used in our example here. Again, because the developer agent is connected to my tenant, it knows what agent ready tools are available to me. It's not pulling in stuff that I don't have, and so it's gonna work right out the gate for us. Let's go test out our agent. We have it built. Let's go see how it works. Every custom agent you build with developer agent is automatically registered in the agent system of record. This is the agent system of record inside of Sana. If it looks a little different, that's why. But it's really simple to get started. Once your agent is registered, you just have to go and activate it, and now we're ready to get going. In this example, I'm gonna show using your custom agent inside of Gemini Enterprise, but you could also run this inside of Workday, Sana, any of your other favorite tools. So there's our internal mobility agent. Let's go ahead and see it. So again, this agent is right lawfully built. It's built on top of Workday. So it should know what I can and cannot do, but we're gonna put that to the test here. I'm gonna ask it to promote my good friend, Mark, to CTO, something I totally, 100% cannot do on Workday. I do not have the power to do that. And so this agent, because it's running in our Workday guardrails, it knows who I am, is going to say, hey. I actually can't promote work today. If you wanna promote work, maybe go talk to Dana, but you don't have the power to do that, so I'm not gonna do that. This is lawfulness in action. Right? It followed the policies. It knows who I am. It followed security and didn't move forward. We're gonna go ahead now and see what can happen when I actually can do something and have the agent do that. I'm gonna say, please promote Kelly to the senior director of IT. And so this is, again, exactly what we wanna see. The agent has started Kelly's promotion, but, hey. It actually can't take effect today because it understands our processes and policies, and we need to go through an approval process in order to actually have a promotion happen. Right? You can't just promote somebody out of the blue. It's gotta be approved first. And so that's exactly what the agent did here. It started the promotion, which I have the power to do, but it hasn't actually finished anything or circumvented the process. It's just started the process to go through. So that is developer agent and custom agents all in action. Hopefully, that was a fun exciting demo. Everything that we just talked about today, developer agent, agent ready tools, custom agent, is all generally available to our customers who have Extend AI. This is formerly known as the ExtendPro product. So if you already have ExtendPro, congratulations. You have Extend AI. You're ready to go. If you don't have Extend AI just yet, this is a great opportunity to check it out, learn a little bit more, maybe maybe get in there. Now, of course, all of this didn't start here. We announced all of these tools and features back at our developer conference, DevCon, back in June, and then over the summer ran an early adopter program where over 125 customers and partners built over 500 agents or built or modified 6,000 apps with Developer Agent. There's crazy numbers, honestly. It was such an engaging EA program. It was a lot of fun, and now I'm excited to see what everyone else is gonna build out there. And the early results are big. Right? What we learned from our EA customers is that those standard coding tasks were happening 12 to 40 times faster, and those development timelines were eight to 10 times compressed. So everything was moving, honestly, at the speed of light, and it meant that your developers are getting to spend more time at the business working on high value asks instead of just coding information line by line. So what do I leave you with here today? Go check out the developer program. Go to developer.workday.com, scan that QR code, make your own personal developer account, get your own personal development tenant spun up, build an agent, play with developer agent, do all that fun stuff. And then once you have actually built your agent, obviously, it needs to run-in a place. It needs to have somewhere to go, and you want that place to be managed, governed, and trusted. And to tell us all more about that is my good friend, Joe Lee. Thanks, Joe. Over to you. Ashley, that was awesome. Thank you very much. So hello, everybody. My name is Jolie, and I'm another product manager, but on our AI platform team. And I am proud to be joined by Becky. Do you wanna do an introduction, Becky? Hi. Hi, everybody. I'm Becky. I'm a software engineer on the agent platform, specifically on the agent system of record team. Awesome. So we have a lot of great stuff to show you today. But first, let's talk a little bit about our trusted agent platform today. So our platform is built for global scale. Just over the past few months, we've made over a 515,000,000 HCM and self-service recommendations. Over 11,800,000,000 AI interactions made to the platform also. We're managing our platform across 11 global regions, making sure it's conforming to regional, sovereign, and federal compliance. This is the platform you can trust to run and scale your critical agentic workloads. And adding to our trusted agent platform, we are announcing the GA of our Workday agent harness. All of our Workday agents built by Workday, like the decision intelligence agent, the payroll agent, deployment agent, self-service agent, and even the internal mobility agent that Ashley demoed, these all run on our agent harness. So Ashley showed that the agent that can act lawfully using agent ready tools. But the harder part of an agent is maintaining context, operating safely, and prove... Improving those decisions made by the agent. So that's where the agent harness comes in. So let's take a look at some of the core features for this. With memory, on our harness, you could start a conversation and step away. And when you return, the agent picks up the conversation right where you left off. With episodic memory, the agent remembers across multiple turn and distinct interactions, learning what it should or shouldn't do. This combined with your tenant data, your agents get smarter with every interaction. Next with guardrails. These are always on keeping your agents safe. These block outside attacks and make sure every response is appropriate and on brand, and the guardrails can be tuned for catering to unique interactions. Next, sovereign routing. Keeps your data in region where it belongs no matter what region your tenant is in. Every request routed is routed in region to an approved foundation model partner or Workday hosted model. Speaking of models, custom agents like Ash... The one that Ashlee built, developers can select the model the agent uses. We're partnering with foundation model partners like Anthropic, Google, OpenAI, or open models like NVIDIA, Meta, and Mistral AI. Evaluations. These ensure every agent is accurate before it actually gets rolled out to users. Traces how the agent actually performed and prove that its reasoning is sound. So all of this is within the agent harness. This is all agents that run on Workday's platform. You get all these all these features. Everything is built in and ready for all your agents, but we're not done yet. Two more features coming soon. The first, agent contact service. Becky's gonna do an awesome demo with this, but this hands the right information to your agent before it even starts work. Next, grounded reasoning. So today with agents, permissions only tell the agent what it can see. Grounded reasoning tells the agent what it's allowed to do with the data. So with agent context service and grounded reasoning together, that solves the problem of agents knowing when to stop and escalate. Alright. Let's see this in action. Paraway. Give me just a second. Okay. So I get to show you all agent contact service and ground reasoning in action. So I will pull up the internal mobility agent, and this is the same agent that Ashley was building. All I'm gonna do is change how it runs. So first, we can... So everything that Joe mentioned that the agent harness provides, the memory, the guardrails, the routing, all of that is on by default. We're just gonna layer on the news... The new things that are coming. So first, I'll turn on the agent contact service, which allows the agent to detect information relevant to the current task and inject that into the context window, which results in faster decision making, quicker responses, and overall less token spend. Then we'll turn on grounded reasoning, which enriches the agent with business process and organizational data straight from your tenant. So in real time, it allows the agent to check whether any given action should proceed, be declined, or be escalated. And that's really the whole setup. The agent contact service and grounded reasoning are built into our agent harness. There's no extra services. There's no extra code required to be written, and so we can see it in action. So we're gonna give the agent a little bit of a harder task. So let's pretend that I'm a director of platform engineering, and I have an open position on my team. And I'm looking for an internal candidate to backfill a senior engineering role. So I will prompt the agent, and I'm gonna suggest that it take a look at our historical performance reviews and identify the strongest candidates. Here at Acme, we have, like, ten years of performance reviews and thousands of potential people. So let's see how this goes. We can see what happens before the agent really gets started on its reasoning. The agent contact service is pulling everything that the agent needs into this... Into the context window. And we're talking specifics about this job requisition, so that's the required skills, the level, the location. Our agent now has the full picture upfront, which gives us faster and more direct results. And then the agent can dive into reviewing our candidate's performance history, which we will see in a couple seconds. So most agents would be reading ten years of performance reviews, which is slow, it's expensive, and it's pretty risky when you're moving sensitive employee data around into the agent's context window. With agent harness, that data is staying exactly where it is in the tenant, and the agent is only reading what it needs. We use the agent harness here at Workday. And with our own Workday delivered agents, we've seen the agent contact service provide seven times fewer reasoning loops, which just gives us faster results, which is really awesome. We can see here that the agent found me two candidates. We have Gavin, who's a senior engineer on the data team, and we have Anisha, who's a software engineer on the mobile team. Gavin is our strongest match, and the agent is going to prompt me to initiate our transfer. So the agent is trying to move Gavin into position, but we can see here the grounded reasoning stepped in. So let's take a look at the reasoning log to see what happened. So it turns out at Acme, we have a mobility policy, and this says that someone who is in a critical role cannot transfer out without a ReadyNow successor. In our tenant, Gavin's role is flagged as critical, but he has no successor named. So as the hiring manager of this open position, I know about our mobility policy, but I don't know much about Gavin's position. So I would have filled the gap on my team by creating a gap on someone else's team. And so what's so cool about this is that any agent can read a policy document that tells us about this mobility policy. But with grounded reasoning, our agent had access to what lives in your tenant. This... The important information about Gavin's role being critical and not having already now successor names, that lives in your tenant, and grounded reasoning provided the bridge that allowed us to prevent a big succession issue. So grounded reasoning caught our mistake, but it doesn't give up, and it suggests that Anisha would pass our, mobility policy, and it prompts me to initiate the transfer. So we can do that. And we started the transfer, and it launches into a business process just as normal. So overall, we got to see the harness provide the context and the grounding to make this whole flow a a success, and our agent was able to make critical data driven decisions faster and ensure that we're complying with company policy and our business rules. Back to Awesome. Thank you, Becky. Appreciate it. you, Okay. Doesn't stop there. We're still. going. We are also announcing that we are an early access for building your own agents with your favorite runtime or frameworks. So this is essentially being able to use external frameworks and bring your own code to run on Workday's agent harness. Our engineers already build this way. Now you can too. So to make this happen, we partner with industry leaders, AWS for critical infrastructure and sandbox isolation, Google for the agent developer kit and massive agent scaling, Langchain for Langgraph runtime and traces, and NVIDIA bringing the latest Neutron models to developers can choose with... That developers could choose with custom agents. Okay. So with open developer program, custom agents, all these agents running around in your business, the more agents you have, they're built more ways by more people. Every one of them must be governed no matter who builds it. But how can you trust an agent? How do you protect against things like prompt injection? Let's take a look at what we're up against. So looking at this example, this is a jailbreak injection example. It's hidden in an ordinary file. In this case, just a resume. You can't see it, but it's in there, and the agent certainly won't miss it. Without the right defenses, one hidden line like this can lead to an unfavorable outcome or even a lawsuit. But Workday's got a solution. That is Agent Passport. Agent Passport, we introduced back at DevCon in June, which is amazing, by the way. Everybody should attend DevCon. But Passport allows you to test agents against attacks like the one you just saw. It uses policies to make standards consistent across all of your agents. Today, we're bringing passport stamps that are available for all Workday agents. What do I mean by stamps? So every single Workday agent built by Workday now carries stamps, and you know it's assessed before you go live with the agent. And this isn't just Workday agents. Your agents, partner agents also can earn c stamps. These are tested against the same standards. Now a stamp is only as good as who's behind it. Cisco, the assessor of our agents, and we're also adding in Stryker and OneTrust. These are industry leaders in agent security and AI governance. All of these agents tested to a high standard by many assessors. That's agent passport. So agent passport is EA today and will go GA in December. Becky, you wanna show us a demo how it works? Yep. Absolutely. Alright. So let's get back into the agent system of record in Asana, and we're gonna navigate to the trust and safety page. And here we can see that we have a trust policy that includes the three stamps that are required for any agent running on the Workday harness. I'm gonna edit our policy, and we're gonna add the prompt injection and the jailbreak resistance stamps. Once I add them to our policy, I can see that our internal mobility agent is marked as noncompliant, and that's because it's never been scanned. We can see that it's missing the prompt injection and jailbreak resistance stamps I just added. So from here, I'm able to kick off the scan, and quickly we can see that our internal mobility agent passes the scan, it is compliant, and it is active and ready to be used. So full transparency, these scans take a couple of hours and we sped them up for the sake of this demo. But what's nice about knowing that it takes so long is that it's so thorough, and we're really testing everything that could possibly go wrong with the agent. Back to you, Joe. Awesome. Thank you, Becky. Alright. With that, I think I will turn it back over to our fabulous MC Hunter to, to walk us through a panel discussion. Hunter, you might be on mute there, my friend. You know, you you... If you know what I've learned by now. Thank you. Thank you, Joe. Thank you, Becky. Thank you, Ashley, for the fantastic demos. Everything we just showed here today was announced at North Bay Rising, and have various degrees of availability as as we just shared. So some are pre early access, some are moving into early access, and then some, are are general availability today. I wanna throw the first question over to Ashley. Ashley, you spent a ton of time at Workday Rising with customers, building agents on the floor. What is, like, an insight that you walked away from after after working with those customers over the last three days? Any trends or patterns you saw that you can share with this group here? Yeah. For sure. So as Hunter mentioned, I was at Rising. I got the opportunity, and I worked at the build an agent booth. So right in the middle of the expo expo floor, we had, like, 25 computers, and. people could come and do a quick little activity to build their first agent in about ten minutes. And so I think one of my big takeaways was just how many people went from sort of, like, being nervous about coming up to the booth and, like, trying it out. And we'd be like, oh, do you wanna come build an agent? And they were like, oh, I don't know if I can or what's up. And I'm like, you totally can. Come sit, try, developer agent is here. And every person who came was able to build an agent in ten minutes. I know several people said how fun it was to build an agent. Many people walked away from the discuss... The event saying that it was really fun to build an agent. It was really easy. We also had some really interesting agents. We had a, activity obviously to do, but people like to kind of explore on their own a little bit, and so we also got to see some really interesting agents. I think this might have been the one... A couple of, like, payroll agents, finance agents. Somebody built a very silly one where they had, it give performance reviews, but they talked like a pirate in the responses. So lots of really good stuff coming out for the build an agent booth. I think we had almost 700 people come through and build an agent across those three days. So really exciting to see, like, just the energy and how easy it was for everyone to get started. Yeah. That's great. Yeah. It was always the most packed booth. I mean, a thousand agents over the course of, you know, three days is is crazy. Joe, Becky, I have a question for the two of you. This kept coming up in the q and a during your presentation. Just, like, questions around human in the loop. You know, where does human in the loop come into the process that you just described? Can you talk a little bit about that? Yeah. I could take the first crack at it. So, yes, human in the loop is something we take incredibly seriously here at Workday, and we wanna make sure the experience is really good for all of our agents. That includes custom agents. That includes our Workday agents. So we are working very hard to ensure that any action that could be not potentially harmful, but any action that is going to modify your data, change your data, that we're making sure we surface human in the loop. So that goes all the way down to our agent ready tools. That's why the agent ready tools are so important. It's because they're annotated properly that this is this is a right operation. This is something that could change your data. So on our harness, again, we're able to identify exactly which tools are part of that, which ones have those annotations that need that human in the loop. So we're improving not only the identifying which, actions need human in the loop, but also the user experience. So we're looking at things like ADY, things that make that actual user experience a bit easier to use. And that's not just on running on Asana, running on the Workday chat. That's across the different platforms. So ADUI is an open standard. But, again, it's it's making sure that not only your experience inside Sonar and Workday, but your experiences if you bring those agents, Gemini Enterprise, to Microsoft Copilot, that it's having, like, the same consistent experience and and being being, not efficient, but it's it's it's it's being, consistent across every single interaction you have with any of those potential operations. It's great. other thing the other thing I'll add is that, especially with something like internal transfers, we're not... Just because you're using an agent to launch something in Workday, we're not skirting the business processes you already have set up. If something's an initiating action, the agent is the one who initiated it, but we're still going through your normal business process flows. So there's no magic happening here. The agent is just a little bit of an intermediary. Yeah. Yeah. Yeah. It's great. thing applies with our agent ready tools. Yes. It's all back, like Becky said, with our business processes, everything. Yes. And I'll throw, one final question to whoever wants to answer it or or, multi view, want to answer it. Workday has a bunch of agents that that we build ourselves that you can, surface in different places and can use to get real work done in HR and finance. When do you see is the right time to build something custom, to go in, to use developer agent, and build a custom agent? Can you talk us through a little bit of that? Whoever wants to take it. Yeah. I I could take the first stab at it, and then I'll turn over the others. I mean, this this... That bigger question of build versus buy is really what that is. This is not a new problem. We... When we came out with the extend product years ago, it it... The same questions came up. It's like, look. What if Workday build something in in six months that that does what I need it to? So this question is is not really new. It's just we have new new toys that are that are solving these problems. But the question kinda remains is is build versus buy. Again, we went through this with extend, but, essentially, the the way I like to talk about it is, first, look at Workday's agents. See what's out of the box already. See what's already built so you don't have to reinvest to buy something. So I would first take a look at what's available in Workday, because it it could solve 80, 90% of what you need. But looking at that last 10, 20% possibly, what do you do there? If it doesn't do exactly what you need, then I would start looking at at custom. Another way to look at it too is, let's say, it... It's a road map agent. But is that agent gonna do exactly what you need? Is that process critical enough? Because every single company has their own very unique, very special ways of handling data or their processes. So we respect that. That's why we give them the open tools. That's why we give them the ability to create screens, apps, and extend, and their own agents is because those critical processes that Workday may never build out, we're able to give that to them. We're able to give them the tooling. We're able to give them that infrastructure on Workday's agent harness and through the tooling through Xtend dot ai to be able to do that. So it's really looking at the process down exactly what you need and how critical that is. But, again, I would always first start with the workday agents first and then and then take a look at at what you can do on our platform to build. It's great. Fantastic. The only other thing I would add to that one really fast and then... But I can let if anyone else has anything. The other thing to consider is, like like Joe said, look at those first party agents. Do they fulfill what you need? But then also looking across your ecosystem, is there stuff that you're doing in a spreadsheet or over in a corner or, like, in some weird tool that's the only thing it does? Is that an opportunity to, like, bring it into Workday? Is that an opportunity to put an agent on top of it? Could that help solve a problem somewhere else that you you can now consolidate down to? That'd be the one other thing I think to look at as well. Yeah. One additional point. Sorry. You brought up something interesting. The the work to agents, yeah, we're we're up here talking about the work to agents that are great, but we also have a big partner ecosystem. So the partners are all on board with everything that we're doing, and they are building agents. They are building agents. And with, the frameworks that we just put out around rising time, you're gonna see a lot more partner agents on the marketplace. So that would be the next stop is if a partner has something that's already out there, absolutely engage with that partner. And then, using our frameworks again, we can we can get you rolling pretty quickly with their agents that they've already built, and that goes through our partner program. So all those agents you see from our partners are all, they're all QA'd. They all pass, quality checks from our from our own team. So, yeah, definitely take a look at the the, the marketplace for for partner agents. It's great. Thank you, guys. I'll send you all backstage to answer some of the q and a that's coming in, and we'll close out with some slides. Alright. So big thank you to all of our, our panelists there. Just a little summary of the announcements that we we gave at rising. We gave a brief overview of today. So I'm gonna go down each of these. Generally available, there are 400 agent ready tools that are available to... Today through the Workday, MCP connection. So that's how you can sort of connect your cloud, your Gemini, your Copilot to Workday through the official Workday MCP. We have 500 more coming. Next, Workday data cloud, which we didn't talk much about today, featuring live data query and agent lake. There's there's plenty more on that, online that you can learn about. And then I believe we have other webinars, coming up in the future where you can tap in and learn more about data cloud. Next, developer agent is generally available as well as building custom agents on the Workday platform. Agent harness. So some of the elements that Joe and Ashley... Or not Joe and Becky just shared, are GA today, so how the infrastructure that allows agents to execute on Workday. Agent stamps via the agent passport is generally available. And then, the personal developer tenants, privacy is beta, generally available. Next, early adopters, third party frameworks for building custom agents, is now an early adopters, early accesses. This is just getting started, but we'll expand it more, over the next few months and into the next year. Also, agent contact service and grounded reasonings. The... That example that Joe and Becky just went really deep into around agent harness, both of those are an early adopter. And then coming soon, Agent Passport will be GA December 2026, as well as the next wave of agent ready tools is over 500 new ways to do things inside of Workday from your preferred agentic platform. Last call to action. So one, start building, try out the tools, get your own developer account for the Workday developer program, and just start building. Number two, keep learning and explore what's possible. So this is the Workday dev site. This is... We just did a big site refresh. It's it's beautiful and has a ton of great information there. We also just launched a developer YouTube channel where our head of developer relations, Nick Morris, does a a great job with his team breaking down all the great stuff coming out of the developer program. And next, if you have use cases in mind, you're curious about cost, talk to your Workday sales rep. They have all of this information, and they can walk, through all of this specifically tailored to you and your business. The QR code there, you can scan that to to just jump on to the site. And then lastly, join us for the next webinar. So in just a few weeks, on November 18 at 9AM PT, we are going to have a get to know extend AI. So this is how you'll learn how to build custom agents with Workday Extend AI in a deeper way that we showed today. It'll be it'll be a lot about that. You'll also hear from One Digital speak about, custom m and a agent they built with Extend AI and the Workday's trusted agent platform. So great hearing from a customer there. And then you'll walk away with practical next steps, for how you can start building and and get started with your own custom agents on Workday. But thank you so much for joining us this morning, everybody. We hope you have a great rest of your Thursday, and hope to see you at the next one.