Video: Less Troubleshooting, Faster Resolutions: Get to Know Admin Agent | Duration: 2592s | Summary: Less Troubleshooting, Faster Resolutions: Get to Know Admin Agent | Chapters: Welcome and Introduction (25.35s), Housekeeping and Agenda (50.41s), Agentic IT Teams (153.555s), ASOR Agent Setup (415.65s), BP Optimize Agent (579.345s), Adoption Agent (1206.65s), Key Takeaways Recap (1873.2s), Q&A Session (1959.48s), Q&A and Wrap-Up (2140.205s), Closing and Next Steps (2274.77s)
Transcript for "Less Troubleshooting, Faster Resolutions: Get to Know Admin Agent": Hi. Thanks for joining today's webinar. Looking forward with Workday adopt in advance. Today's session is less troubleshooting, faster resolutions, get to know the admin agent. During this session, I will share a few forward looking statements that are subject to change. Please note that these are covered under our product statement agreement. Before we get started, let's cover a few housekeeping items. This webinar is being recorded and will be e emailed to you when published. If you have any questions, you can select the q and a button on the right hand side. We will try to leave some time at the end to help answer any questions that come in, but we will also have people monitoring and providing responses during the entire call. We have a large audience today, and if we can't get to every question during the session, we will make sure to follow-up with responses by the end of the week. We have uploaded some resources for you to view. Just simply select docs on the right hand side to review and download materials. Next, we have a couple of polls you can respond to during the session, and you'll be prompted with a survey after after the webinar concludes. We appreciate your feedback. And a quick note on demos, we have a few different ones from a few different sources, so there's likely gonna be a varied volume across these different demos. The agenda for today's session will include the introduction of Workday teams of agents. We will take a deeper look at the admin agent, which exists inside the agentic IT team. After that, we will dig deeper into two of the admin agent sub agents, BP optimized agent and the adoption agent. Lastly, time permitting, I will respond live to a couple important or common questions asked by everyone during this session. I will be your host today. I am Scott Erickson, product marketing manager at Workday. I focus on the different sub agents that met make up the admin agent. That said, I am not alone in this session. We have a handful of Workday experts backstage supporting me to help answer any questions you may have as we walk through the content today. Teams of agents. Workday understands the evolution of agents over the past couple years has been difficult for some to understand. There's a lot of new things being announced. It seems like every week there's a new agent. It can be hard to gauge if a given agent is right for you. First, the teams. There are five core teams, agentic HR, agentic finance, agentic IT, travel and expense, and legal. Based on your role, you will likely going to fall into one of these teams. This session will be focusing on the agentic IT team, specifically the admin agent. So in the Ajentic IT team, we have a developer and admin agent. If your role is to build on the Workday platform, the developer agent is for you. If your role is to configure, deploy, and manage the Workday platform, the admin agent is for you. Up to this point, when Workday has spoken about agents referring to Workday built first party agents found directly in the Workday platform, these agents, the adoption agent, the deployment agent, BP optimized agent, these are what I will refer to as subagents. These sub agents are actually doing all the work, what you actually interact with, what you actually deploy in the platform. The admin agent isn't a new agent. It is not a SKU. It is not something you configure. It is the role based grouping of sub agents that combine together to fit a role. On this slide, the orange text are the agents, which are the groupings of sub agents that support a given role. And in some cases, there are many agents that support a general team across different roles, like a Jentech HR team. If you're a Workday customer, you likely fit into one or two of these different roles. So instead of asking if a given sub agent is for you, you can start on the other side and filter things out so you can easily understand which agents you can focus on. Today, this is a new way for Workday to communicate about different agents. But as we look forward, it will shift more concretely into a single unified experience. As these agents mature, they will be brought directly into the Sono core experience. Users will access all agents from a single location, and Sana will be behind the scenes orchestration tool that routes a given question or comment to the correct agent. Users won't have to go to three different places to interact with three different agents. They just open Workday and ask Sana. But this is something that is coming later. Today, the admin agent is the package of different sub agents. And now I would like to dig a bit deeper into the admin agent at a high level. The admin agent currently consists of three core subagents. The adoption agent will be side by side with you for biannual releases. It will help surface the most important features to adopt based on your tenant's configuration and give guidance on the testing and configuration of those features. The deployment agent gives detailed information, and with the r two release, it will be in tenant context to get you moving quickly for error resolution or configuration changes. The BP optimized agent analyzes your processes for configuration issues and gives guidance to improvement based on industry peers and best practices. These are the three core sub agents of the admin agent. We'll be adding more sub agents in the future to support the Workday admin. A notable addition will be the report authoring agent would is which is expected to GA next year. Those of us at Workday understand maintaining your Workday deployment can be a monumental task. We have 40 employees within Workday who maintain Workday on Workday, which is our internal Workday deployment for us. We understand the complexities and efforts it takes to maintain, and and it only gets harder with smaller teams. For teams with only one or two admins, it becomes a job of just keeping things working as is with little time to explore how to expand and improve the deployment. This is where the different agents the different sub agents for admin agent come into support. To address these issues, we have reimagined the ownership experience as a continuous cycle of accelerating, executing, and assisting. Before we get into the different sub agents, I do wanna give a quick demo of how to use ASOR as this will be speak about, numerous times in this presentation. Yeah. In this demo, I'm gonna walk through some of the basics of ASOR. This will help you understand what it takes to register, configure, and activate an agent. When you go into the agent management hub, you will see a tab for unregistered agents. Here, you can select a given agent to bring it into your tenant. Once you register it, it will now show up in your agent registry. But before an agent can be used, it requires some security settings to make sure you provide the proper users the proper access and to prevent the agent from accessing data it shouldn't be able to access. Next, you can walk through the security configurations for the agent. This will let you assign the different security groups to each skill the agent provides. In some cases, you will simply define the same users to all skills. But some agents have wider use cases and can support different teams. So per skill configuration is provided. This ensures that a given agent is only accessible to the users it needs to support. And that's it. Once you activate the agent, the end users who have access will see it in their agent chat or in the corresponding hub the agent operates in. K. Next, let's take a look at the first sub agent, the b p optimize agent. BP opt BP Optimize agent is available to all customers as of the r one release earlier this year. I will review the value of, use cases for, and some common optimization journeys you can take with BP Optimize. Then I will give a technical and pricing overview, and we will see a demo that shows where it can be accessed and how to use it. The BP Optimize agent acts as an embedded business process optimizer that continuously monitors the heartbeat of your Workday business processes. Unlike third party tools that require heavy data extraction, this agent sits natively within the platform to continuously detect configuration anomalies, flag performance degradation, and provide AI driven recommendations to resolve inefficiencies and avoid impacts to business agility. If you have managed business processes before, you understand the complexities. There can be a lot of them, and they can be a spiderweb of potential paths as users go through different processes. It can be very difficult to even understand the flow of a given process. The agent will look at the processes and surface specific steps inside the process that are increasing the cycle time. It will explain where in the process things are getting hung up, analyze why this is happening, and give direct step by step guidance on how to adjust the configuration to improve the flow of the process. In some cases, it may not even be the process configuration that is causing issues. I've spoken to customers who have had the agent analyze a process and have found a couple steps were taking much longer, days longer than expected. The admin then dug into those steps of the process, was able to easily see the owners of those steps, and reach out directly for them to input as to why things take so long. They realized it wasn't the business process that was broken, but the external tools and processes those people were leveraging during those steps that were the problem. The BP Optimize agent is able to not only analyze processes, it is able to understand how your current implementation of a process stacks up against well configured versions across your industry peers. The process may be completing fine, but it's taking twice as long as it should. This is time wasted by the workforce. It will guide admins on what to change to streamline the process and bring it in alignment with the industry peers. The agent is able to look at processes and figure out not just if a process is inefficient, but it can value the different paths the process actually takes based on the historical data to understand if specific steps may be inefficient that may be inefficient or even worth focusing on. The more often or likely a path is taken, the more important it is to get it right. This ensures admins are able to focus their efforts on the steps that really matter. I wanna frame a few different ways to think about what type of ways the agent can support process optimization. Reactive optimization. This is understanding where bottlenecks are and detailed analysis on how to resolve them. Proactive optimization, surfacing inefficiencies early before they become critical failures. Passive optimization, guidance on how to streamline current processes even if they aren't causing issues. Strategic optimization, enables organizations to define KPIs then optimize and measure to meet those. Next, let's start to dig a little bit deeper into the agent from a technical perspective. Data security is always a top priority when deciding if a given agent is right for your organization. All data is processed locally. Your tenant's configuration details are not leaving the Workday platform. The benchmarking of data is always done with fully anonymized data scrubbed of any identifying details to protect customers' privacy, and the data the agent is using to inspect your processes is never shared across customers. Let's look at what happens when a BP event is processed. There are four key steps here. Data ingestion. All BP related data is collected from the customer's tenant. This is more than just the processes, but also the security configuration, organizational structure, anything related to what the process is actually doing. ML model analysis. Our internal machine learning models analyze the data to detect anomalies, trends, and create benchmarks providing comprehensive insights around the given process. GenAI recommendations. Root cause analysis and details are sent to the BP optimized agent to create recommendations and detailed testing and configuration plans. Then delivery. Insights and recommendations are sent back to your tenant through the agent's interactive chat. And, of course, this is built with Workday's foundational trust and security. Customer data is never shared between tenants, and benchmarking data is always anonymized to protect protect customers' identities. So how can you activate the agent? The agent is configured in ASOR. Please note that the UMSA is required before you can do this. Once activated, you will need to configure it with a few key points of consideration. The scope of a historical BP data from your tenant that the agent can ingest for analysis, the look back period. Obviously, the more historical data it has, the better, but the agent does consume one flex credit per 10 BP events it analyzes and a maximum of 2,000,000 events per month. So you and you must have at least one month of historical data actually in your tenant before the data can be enabled. So depending on your tenant, you will wanna be aware of these trade offs. You can also select different product lines you want to include. You don't have to include all processes across your deployment if you don't want to. When does it consume flex credits? At this point, we are waiving all flex credit consumption the first time you enable the agent. If you later disable and reenable it within twenty one days, no new charges will be made. It will still be considered the first enablement of the agent because it will not have to reanalyze the event data. But because the agent only holds previously analyzed event data for twenty one days after deactivation, if you disable it for more than twenty one days and then reenable it, that will be considered a new activation. Historical data will then consume flex credits. The second point, the flex credit consumption, is triggered on the first day of the calendar month. On the first day of each month, the agent will look back at the previous month's BP events across your enabled product lines and bring those into the historical analysis. Again, these will be charged at a one to 10 event rate per, flex credit. Because of this structured timing of the new event consumption, it is advised that if you wanna test the agent out before committing to flex credit usage, you enable it early in the month to give you the most runway for testing before any flex credit charges hit. The agent doesn't charge credits when you interact with it. It only charges either on the second or later activation of the agent and each month on the first day of the month based on the total number of new BP events that were created in the active product lines for that previous month. BP optimized agent is only supported currently in the prod and prod sandbox environments, and the customer first needs to activate it in production before they can access the agent in the prod sandbox. Next, a quick demo of the BP agent, BPO agent in action. In this demo, I'm gonna walk through a couple of different ways you can use the BP Optimize agent. It can be activated in your production or sandbox tenant. You can reach the main dashboard by typing BP Optimize agent in the Workday search bar. This will load a high level overview of the different processes you currently have active inside the agent for analysis. You can configure the dash to show other processes or to adjust your industry profile. You can switch between process time and send back counts among other configurable parameters. Next, I wanna show how you interact with the agent. In the conversational window, you can ask the agent questions. What processes are causing problems? Where can I focus efforts to improve? Or ask directly about a given known process that is causing friction. Here, I am asking for any processes with high anomalies. It locates the most problematic processes and asks if I wanna see other processes with high anomaly counts. I asked to get more details on the change job process as this is one we have heard internally is causing some friction. I asked its advice on how to improve the process. It shows a list of different ways I can improve the process from changes like conditionally bypass steps for mass actions to configuring templates to streamline the process for users. If you want more information on a given step, just ask it for more concrete details. It can dig deeper into any of its suggestions to show exactly what the admin needs to focus on to get the improvement they are looking for. Next, we can go back to the dashboard and view the actual processes one at a time. Click into a given process for more details and explore how the process is actually working. We can see stats, trends, anomalies for each process, and the agent will outline different steps for improvement with details around documentation for each suggestion. Next, we will dig into another subagent, the adoption agent. With the adoption agent now available for all Workday customers, we have made it free to use during the twenty six r two release cycle. You can activate it in your preview tenant and use it without consuming any flex credits. I will give a quick overview of how it works under the hood, dig into the skills included today and what's coming soon, and we will see a demo that shows where you can access it and how to use it. So first, where do admins access the deployment agent? The agent is accessed directly in the adoption planning hub. This is where a lot of admins manage their relay cycle process. It is where you actually manage the work from tracking backlogs to assigning specific testers and configurators. This is a native tool included within your subscription at no additional cost. While some use Smartsheet or ServiceNow, this allows you to manage your Workday innovation directly within the platform. Use the hub to create a clear road map for upcoming features and rolling adoption campaigns, making it easy to share progress and priorities during steering committee meetings. The hub allows you to discover new features specifically relevant to your tenant and tracks retiring features so you can plan for necessary changes. You can test the agent in the hub in your preview tenant right now. So what is the adoption agent under the hood? The adoption agent is really a combination of three pieces. The LLM, a curated workday of lie the curated workday library of content, and your tenant's current configuration found in the preview tenant. The LLM is the reasoning engine. The adoption agent will take a given release note, inspect it for needed tenant specific configuration details, combine that with curated Workday document details, and then get a clear, concise response on how this configuration change will impact your deployment. It will give the step by step guide on what to do in order to turn on the feature given there are configuration changes needed. The RAG, the retrieval augmented generation. This is the library of Workday curated documents. The agent leverages these documents to build a complete prompt, including all relevant content from these curated documents. This is the shared data used by the agent. It is the set of Workday documents, user guides, community content, admin guides, and release notes. This is what's used by the agent across all Workday customers. This means that your tenant's information is leveraged only for your agent's analysis. The agent isn't trained on Workday data. We simply use the general LLM. We simply send the general LLM a prompt that includes only the relevant information needed to analyze the release note. We never train this agent on your data or use your tenant's data for other customers' agents. The adoption agent currently uses Gemini, which is a general purpose model that comes pretrained on a wide range of public data. It is very good at understanding natural language, reasoning through complex questions, and generating clear responses, but it does require us, Workday, to provide clear, concise details for the specific release note, user guide, admin guides, etcetera, to narrow down the context to make sure it understands what is needed to guide the admin for a given feature. In the simplest terms, probably oversimplified a little bit, but the adoption agent is building and submitting a highly curated LLM prompt to Gemini, including all of the most relevant information structured in a way that allows Gemini, which is only trained on public data available to Google, to explain exactly how a given feature will impact your specific tenant and what is needed to configure, test, and implement that feature in production. So moving past what the agent does under the hood, let's shift to how it can support Workday admins. Admins often struggle to keep up with the volume of features and code changes delivered in Workday releases. The process of reviewing release notes to identify necessary testing, feature adoption, deprecated items can be overwhelming. You may not even know without spending time reading reading the different release notes if a feature is even relevant to you. The adoption agent addresses these challenges by analyzing your tenant's configuration against incoming release notes for automatically available features. It will find the features that fit your needs and surface ones that really need your attention. You won't have to dig through the console or community to figure out which new features you can or should bring into your tenant. It makes release changes actionable, mitigating the stress of release cutovers and allowing teams to focus on the features that matter most. Administrators start the adoption planning process within the preview tenant to trigger an impact report. The agent generates a report tailored to the user's selected release and functional areas. You can go through a few critical functional areas or run the report on everything at once. It's up to you how you break that down. It will keep track of what has already been processed so you can ensure not to revisit functional areas you have already viewed. Unlike generic guidance, the agent evaluates the changes specifically against your actual setup and provides recommended net next steps, including testing and validation steps. Users can access a step by step breakdown of the agent's work, save details directly into adoption items without exporting data, and admins can assign tasks to product owners internally. The agent helps amplify the value of the adoption planning hub so teams can focus on the high impact features and efficiently manage potential issues. With the current release, admins can leverage the adoption planning hub to get details on feature by feature basis to understand what impact, if any, each new feature carries. It will also give detailed testing plans and explain exactly how to enable the feature. The features will be ranked by impact, and this will allow admins to leverage the adoption planning hub more efficiently, letting them quickly focus on the highest impact features, assigning them to the proper teams, and tracking their progress through release adoption cycle. As we look forward, the adoption agent will also include analysis on retiring features, giving admins clear visibility into what features will be removed in the future and what impact that will have. This will excuse me. This will ensure you are not caught off guard by feature removal you weren't aware of or you weren't aware was would have a high impact on your users. Likewise, it will also eventually show different fixes that are coming in the next release and the impact of those fixes. While a current release does not include feature adoption recommendation skill, this will be added soon. This will take feature analysis to another level. It will take into consideration your organization's strategic outcomes, further refining which features will have the most impact. As we look further out on the feature adoption recommendation skill, we will add the ability for the adoption agent to craft detailed implementation plans, 10 taking your tenant's configuration into context. Now that we've sco covered the values and capabilities, let's walk through how to get started with the adoption agent. First, again, you will configure the agent in ASOR. It is critical to know here that the UMSA is required to access ASOR and, in turn, activate and configure the agent. Next, if you haven't, you'll want to configure the adoption planning hub. Ensure that users who have access to the adoption agent via security group configurations in ASOR also have the adoption planning domain. You can launch the release impact analysis skill directly via the adoption planning hub, which you will see in a demo coming up soon. Lastly, I wanna highlight two important points regarding usage. We are making it easy to test this agent out in this release cycle. Flex credits will not be charged during the full twenty six r two release cycle. We do expect the agent will consume flex credits by the twenty seven r one release cycle early next year. Please keep in mind that the release impact analysis skill is currently available only in your preview environment. Post GA, the agent will remain in the preview tenant, though you can manually migrate your configuration to production when you are comfortable with the implementation plan. A few comments from customers out of our recent EA program. It was able to cut down the complexity of the release cycle. It helped admins focus their efforts efficiently, and it gave them time to move past the required updates quickly, opening the door to new features that wouldn't have been possible with the restraints placed on them. Now let's take a look at the adoption agent in action. I have come into adoption planning to access the adoption agent's new feature, release impact analysis. On the overview page, a banner highlights that impact analysis for automatically available features in the '26 r two release is ready to generate. To analyze 26 r two, I select generate impact analysis, click 26 r two, and select the functional areas I would like to analyze. All functional areas that are enabled for your tenant and have at least one automatically available feature can be analyzed. Before creating the report, users can quickly see the total number of release notes that will be analyzed based on the selected functional areas. Once ready, generate your impact report. As the report is progressing, the adoption agent is analyzing each change against your tenant's configuration. Once complete, the adoption agent suggests actions based on its inspection. Users can filter the report based on these actions and review details generated by the adoption agent via the feature summary and impact summary columns. For more information, users can click to see a suggested test plan and a clear step by step list of the actions performed by the agent. Now that you have reviewed the analysis, users can capture these details in adoption items for an in tenant release management experience. First, select the changes you would like captured as an adoption item. Click create adoption item and assign a status, user, and priority to the work. Now with adoption planning, track the progress of each update as reviews and tests are completed. This end to end in tenant experience allows for a simplified and streamlined release management experience. So a couple takeaways. First, we are delivering Workday agentic teammates. This isn't just about individual features. It's about bringing together agents with common specialized skills that directly support you and your role and your goals across the Workday ecosystem. Second, with the BP optimized agent, we are helping agent agent we're helping admins refine and fix their process workflows quickly with expert level analysis, accelerating efforts at all phases of the life cycle. Third, with the adoption agent, we're helping you cut through the noise. It focuses your efforts on the most important feature updates, providing clear, actionable guidance on what you need to test and enable with actionable guidance you need to test and enable them with cut with confidence. Next, a quick note about the docs. This is a few of the docs available plus many more in the right hand tab, underneath the docs. There are some links. There are some documents. Feel free to peruse those and see if there's anything you wanna take home with you. And last, I'm going to go through a all of the questions that were asked today and answer a few of the ones that were asked multiple times or, or further emphasize content that was overlooked in this session. So first where was that? There was one about adoption agent. So from Tara, how far does the adoption agent look back in its, for current releases? Right now, the adoption agent will look back to twenty six r one. So you can do the last release or the current release with the adoption agent. Yeah. Can, from Kate, can you clarify the piece about flex credits and activation for BP optimize? This is a complicated process. I know there's a lot to it, so let me run through that one more time. When you initially configure BP Optimize, you'll be charged one flex credit per historical 10 BP events that is ingested either three months or six months, look back. But the first time you activate it, you're not charged. It would be the second time or later that you activate it. And then on the first day of each month, the agent will look back at the previous month and look at all of the different, BP events that were created by the organization and bring those into the historical analysis. So month by month on the first day, you get charged for, say, if you have 10,000 BP events across the organization in the active product lines, that month will charge, 1,000 flex credits for 10,000 events, and then that will be part of your historical analysis as you move forward. So, again, the first enablement of the agent is free. If you deactivate it for more than twenty one days and reactivate it later, it will be charged one credit per 10 events of historical data. And then every month on the first, you will be charged for the previous month's, event count, as it brings it into historical analysis. I know it's complicated. I'm trying to explain it as best I can, but there's there's a lot of of, moving parts in that. Can BP optimized be activated in sandbox preview? No. Right now, it is just, production and production sandbox. And you enable it, in production first. That gets copied over in the sandbox, and then you can use it there. And, again, you do not get charged to interact with the agent. You only get charged for historical BP event analysis, as it moves forward. So as you interact with it, ask it questions, get advice from it. That is not a a charged event in any way. Are there any about adoption agent? Troubleshooting. So troubleshooting help somewhere for why adoption agent is failing to analyze all the items in the preview tenant. It's from Kelsey. I have seen this occur. It typically has to do with security settings of the agent in agent system of record. The two, bits of advice I would have would be for adoption agent to, disable the agent and then reenable it, and then go through the security configuration and make sure the person that is interacting with the agent is set as an approved user of that skill that is being leveraged. If you trigger an analysis, but you do not properly have the agent configured for yourself, it will fail. So so one of the it's it's my best guess would be, if they're failing, it's a security issue, and either reactivate the agent or make sure the configuration is in line. See if there's one more that was asked that doesn't cover content that was covered here. Back. A lot of BP optimize. Do these agents require UMSA from the SIT? Yes. Yes. They do. Both of these agents require UMSA. They do not require the flex credit, contract, but they do require UMSA, and this is because these agents are activated inside of ASOR. And the only way to get ASOR active in your tenants is to sign the UMSA. So you sign the UMSA, ASOR gets turned on, and then you can go into ASOR and configure the individual agents. Alright. So one note, we have a couple of sessions looking forward with Workday adopt in advance sessions coming up. Simplify your path to Workday innovation. This is all around UMSA and data contribution. If you are confused, and that's quite common right now, around flex credits, the UMSA, and and everything around it, we are having a session that we hopefully will, that we hope will be very informative and help answer questions that are created from this new process. So please join that. That is September 2. And then the next session, enable confident AI adoption, looking at how to actually adopt AI across the Workday platform and how to think about these types of decisions that are made across the organization, not just, looking at how to turn on agents, but what to think about, how to how to, make these decisions, across the organization and some some best practices around, those entire processes of AI adoption. And with that, I thank you for today's session. The recording will be sent to you shortly.