Name: Reshaping the Future of Accounting with Agentic AI
Uploaded: 2025-09-29T18:30:12.996Z
Duration: 1 h 30 min 8 s
Description: Reshaping the Future of Accounting with Agentic AI

Transcript for "Reshaping the Future of Accounting with Agentic AI": Hello. Good morning. Good afternoon, depending on where you're joining us. Thanks for joining today's session on reshaping the future of accounting with agentic AI. We're also really pleased to be joined by our partners at KPMG today who bring extensive experience and deep expertise in this area. And before we kick off, I just want to do a few housekeeping slides. So the looking forward with Workday webinar series is designed to give you insight into how your organization can do more with Workday, including new and innovative strategies to help you get more from your solution. And today, we certainly have a lot of new and upcoming AI innovations to share with you all. And just here are some, you know, standard items, to note before we proceed with the webinar. We are recording this session. We will be making it available to you via email within, forty eight hours, so please be on the lookout for that. Workday uses the gold class platform for our looking forward webinar series. So this allows you to submit questions and get access to resources as well as to provide feedback on the webinar itself. Goldcast operates most effectively when you're using the Chrome browser. So if you are having any issues at this time, just make sure that you're using the Chrome browser. If you do have any technical issues, we have a team here to support, so please use the q and a functionality to reach out to them. You can get even more out of this webinar using the tabs, or the docs tab, in the right hand corner of the screen. This has some related resources, including a link to reports, that works I published that I'm gonna highlight in a few minutes. And then q and a is where you submit your questions, as I mentioned. Our goal is to respond to all of the questions throughout the webinar via the live q and a functionality. If we don't have the answer here today, I will make sure that we do follow-up with you, via email after the event. Lastly, you will receive an email with a short survey following the event today. We do review and use this feedback, to create events and also to build content that's relevant for you. So please do let us know what you think. And just to note our product statement, before we kick off, into some of the content today. And this is just a brief look at today's agenda. So we're gonna explore the topic of agentic AI and AI agents and how they are fundamentally changing the accounting profession, naming teams to automate traditional processes and redesign how they work. With so many new developments, it's really exciting, but it's also challenging to know how to put agentic AI into practice. So we're gonna show you how, by exploring specific use cases, Workday and KPMG are building to improve efficiency and accuracy in your finance operations. We've got a couple of demos, of those upcoming enhancements to show you as well. And finally, KPMG will share their deep expertise providing practical strategies for applying agentic AI to help you to transform your finance processes and your finance organization. Okay. So with that, my name is Marion Carr, and I'm part of the Workday product team. I'm focused on Workday financial management. I'm delighted to share Workday's strategy and our use cases, in the session today for agentic AI and for AI agents, already with our early adopters, and we're seeing some really great results, which we'll also share with you. And I'm also delighted to be joined, by Brian from KPMG. Do you wanna do a quick intro, Brian? Yes. Hi, everyone. My name is Brian Anderson. I'm the Workday practice lead for KPMG. I have about thirty years experience implementing ERP systems, and over the last ten years, been focused on Workday. And the area that I specialize in the industry that I specialize in is financial services. Super. Thanks, Brian. I'm gonna look forward to hearing more, from Brian just in a few moments. But before I pass over, you know, to, cover off, some of what KPMG is seeing in the market, I just wanna, you know, kind of grind us in what we're gonna speak about today. So we're living in this moment where AI is really reshaping our world faster than we ever imagined. So industry skill sets and career paths are being completely redefined in real time. And for decades, you know, up to this point, finance and accounting teams have been really frustrated by that need to manage transactional work, limiting, you know, your ability to partner with the business and drive value. And in a world where AI and specifically, agentic AI can automate and transform work like never before, this role has to change, and this is the opportunity. Our agentic strategy puts people at the center. As a people first company, we're we'll always design our solutions around a future of human machine collaboration. We're not swapping people for AI. Instead, we are building agents to augment people and scale their impact within the organization. Our Genetec strategy and products are designed to work around this pattern. And at the heart of this transformation is the CFO, and we see huge potential right across FP and A, accounting, and procurement. But for today's presentation, we wanna focus on where IdentityAI can help drive value for accounting. And before I get into what Workday is doing in this space, I'm gonna now hand over to Brian to highlight some of the key trends KPMG are seeing in the market and where there is great opportunity for IdentityAI to deliver real value for accounting. So with that, I'm gonna pass over to Brian for a couple of minutes. Perfect. Okay. Well, let's start the conversation around the next chapter of finance. And, you know, what we like to say is, digital enabled transformation is sort of evolving into what we call next gen finance, all enabled by AI. And and so the way that we think about digital enable is is digitally enabled is is modernizing financial systems on the cloud. So a couple of examples just to highlight on the wheel here. So number one, accelerated processing becoming much more continuous cycles and more real time. We're seeing a shift from automation centers of excellence to innovation hubs. And and the the workforce pyramid is shifting, you know, you know, to a diamond structure. So and and what that means is more, you know, people in the middle, managers really reviewing work that's done by both people and agents, you know, out in the workforce. So when we talk about the era of the Gentec AI, first thing I'd like to do is is try to define, you know, GenAI versus the Gentic AI. And and, really, GenAI provides the ability to sift through vast amounts of information to summarize, present, or maybe create new static content. You can think of, you know, writing computer code, writing emails, presentations, doing analysis, putting, data in in tabular results versus the AgenTek AI, which is more about, taking instruction and performing actions. And so what we see in the market is finance is really still, mostly in the experimentation phase, but it's moving, quite rapidly in terms of you can see two thirds of all organizations are piloting AI solutions or Genentech AI solutions, which is up about 37% from the last survey. And almost everybody has plans to deploy, but, what's interesting is only about one in 10 have actually done so to date. So, again, lots implement experimentation, but this is this is the field that's moving, quite rapidly. In terms of how do agents unlock value, I think this is really interesting because there's sort of four ways that we see agents unlocking value. The first is, these agents really open up the possibilities. So with current tools, you can't look at every transaction. Now with AI, you can you can actually go through everything. Number two is agents don't sleep. They don't get burned out. Twenty four hours of productivity means that, you know, potentially up to three times in productivity, and and that assumes that agents and humans will be equally efficient, in the future. So really big opportunity there. Agents are wired for change. They're trained and motivated to deliver specific outcomes. They can evolve to get better. There's no reluctance to collaborate. There's no office politics. They're not threatened by change. So, really, that's that's a benefit there. And then number four is being able to convert knowledge into action. And I think that, are you hearing me okay, Marion? I can hear you. Yes. Sorry. We just Somebody is saying nobody is hearing. I just I just noticed that in the chat, but, hopefully, guys Yeah. No. We can hear you okay. The technical team did check. I think we can hear you. I just I just suggest that if people can't hear you, maybe to adjust their volume. Okay. Great. Sorry, Brian. Okay. Sorry. And then, so so again, the last one here is converting knowledge into action. And and this is really the core difference between GenAI and Adjentic AI is taking action, executing tasks and transactions. So when we look at a, AI driven, finance delivery model, I think maybe a lot of you on the on the webinar here would agree that sometimes even when you put an ERP system in place, either on prem or on the cloud, a lot of organizations are still spending a lot of time on the left hand side transaction processing, doing closed end activities, filling out the financial plans. And now with the shift to AI, I truly believe we're gonna recognize the opportunity to to shift the workload into a couple areas. So the first is, shifting towards monitoring and exception based processing versus processing transactions and running reports. Number two, strategic advising versus just completing the budget and forecast templates. And then I think a new category of solution innovation will emerge, and create applications and models for the businesses to use. And so while the finance function is never really gonna disappear, I do believe the nature of work is gonna change. So it's very, prudent to be thinking about how to make that transition over the next three to five years. Now where are the teams starting with AI and and the finance function? So, I'll talk about three examples here. So first is forecasting and budgeting. So if you think about this function in in a finance organization, it's all about making predictions about the future. Very data intensive, and AI really thrives in in this kind of a use case. We're what we're seeing in the market is the ability to pull in different types of information, which we call signals that can really enhance the predictive and, strategic forecasting. So that's that's there's an area of of, experimentation happening there. The second and I'll lump these together, commentary and and market intelligence. This is really using AI to take a first pass on explaining what happened in the financial results. Also, it's being used to compare performance to other peers through market research, and that's really, that that's quite valuable and and a great task for these these agents. And then, the third one I I would call out here is anomaly detection. This is becoming very popular for finding issues in the core underlying data. Workday has a great solution here called journal insights, which, continuously monitors your journal activity. And what it does is it it it looks at the account string, the the FDM elements, and and those are things like accounts and spend or revenue categories that, are out of line or out of, you know, mismatch based on historical, historical patterns in the dataset. So, again, these are these are where some of the early use cases, have started. And I think what we are now gonna shift to a little bit is, like, how do we really turn these initial use cases into more the agentic framework? So, the model and the framework that we use in in KPMG to to talk about agents is is the TACO model, and it stands for taskers, automators, collaborators, and orchestrators. And so let's go through each of these one at a time. The first is a tasker, and this this is a type of agent that responds well to well defined questions or singular goals. So there's a couple good examples of this. One is, there's a third party called Auditoria. They've built a series of smart smart smart bots for the finance function. So they have something called an AP help desk agent, that attaches to your AP inbox. It reads incoming requests and then dispositions dispositions them accordingly. Another example of this Workday has got something called Expense Protect, which audits, expenses in real time using machine learning and and really looking for duplicates, incorrect classifications of expense line items. So these are things, that we would put in sort of the tasker category. The second category is automator. You can think of this as the next level agent that addresses multi system workflows. It draws on asset knowledge that's, you know, human based. And a good example of this, we'll see, later on in the presentation is KPMG is building a financial close companion, that that helps with the year end close process. The third category is collaborators, and these are higher order agents. They're gonna interact. And the key thing here is they're actually reasoning with the end users. So there needs to be a lot of con contextual awareness. Long term memory are are important for these types of attributes, and and and these attributes are colleagues. You you can sort of envision them as colleagues helping to solve complex problems. And then finally, we have our orchestrators at the bottom. These are the most advanced agents. You can think of this as the control tower, design, handling the most complicated, path in the finance organization. And you could you could envision this as potentially a procurement agent that goes out, sources, either good in service, negotiates price, and actually makes the purchase for an organization sort of all, independently. So that's that's sort of an introduction. And and when we think about this model, and then this is a little hard to read, but, just a couple key points. I wanna talk about how we sort of compare these. Number one is that most of the agents that we see today are either in the tasker or the automated classification. You'll see a couple of these, from Workday. Marion's gonna go through some of those, and we'll show you one that that KPMG is working on. The second point about how we compare, agents in this model is as you move, from left to right, you start with efficiency savings on the left and doing more repetitive work. And then as you move on to the right, you're actually, using these agents to to provide more insight, more value across the organization. So that's a key difference. And then the last point is complexity. So, again, as you move from left to right, the complexity is is not linear. It's it's more exponential. So I I truly believe it's gonna take some time to really roll out collaborators and orchestrators, but this is the vision for where the world is moving in terms of agents. And, you know, I think this is a useful framework to think about how you develop these. It's something that we've, we've come up with over the last, you know, since the agentic AI revolution, has started here. So, why don't I turn it back to Marion to, provide some sort of work what Workday is doing around agents in the finance function. Super. Thank you, Brian. I see lots of great questions, in the q and a, so we can address some of them. Brian, there was a question, around what Workday is doing, around collaboration and commentary and reporting. And that's something that we are working on, and we will highlight it in a few minutes. And we'll try and get to the rest of the questions, shortly. Okay. So I wanted to, you know, highlight a report that work that we've been recently produced. And, you know, as Brian noted, AI agents hold immense value for maintaining, you know, that competitive edge. But with any new technology, there's, you know, gonna always be gonna be apprehension. So we did a recent study, you know, a global study to shed light on the actual adoption sentiment and the role of AI agents in the workplace. And the report confirms that AI agents are rapidly integrating into the workforce. And whilst many employees are comfortable collaborating with agents, very few are willing to be managed by them. There is huge optimism around AI's transformative benefits, and this is accelerating adoption, but, you know, doubt does persist. And you can see a link to the report, in the docs, section of, the webcast there. There is apprehension around, you know, ethical barriers, security, and the potential impact on meaningful human interaction. And these are kind of front and center that came up, as part of the research that we published. But, you know, one thing to note, in the report, you know, for the finance sector, long struggling with that talent shortage, that's particularly growing over the last number of years. AI agents aren't viewed as a trap, but rather as a solution that could help to reshape the industry. But to drive success, it's really critical that leaders ensure that AI enhances that employee experience rather than undermine it. And that's really at the heart, of Workday's AI strategy. So the real power of AI agents lies not in simply automating tasks, but in amplifying human capabilities. Workday's AI strategy is focused on building the best agents that deliver real tangible benefits for our customers. And then we embed these agents into functional areas tied directly to real business outcomes. And then we open up the platform so that you can build and extend in ways that matter most to your own organization and to your own business processes. And everything that we build starts with one singular goal, and that's delivering outcomes for our customers. We know that you have to see the ROI on AI and agent use, And that's why every agent that we build will clearly check one or more of the outcomes that matter most. Things like increasing employee productivity, creating a resilient and an innovative workforce, getting to insights and decisions quicker, and keeping your organization safe and compliant. And I'm gonna come back to a couple of these as we look at some of the use cases that Workday is building. So the AI innovations that we'll show in a couple of moments, sorry, bring unprecedented speed to your teams. They enable you to not just foster compliance efficiency and insight, but also to go beyond that driving strategic operational excellence for your entire business. I wanna take a look at some of the areas that we're bringing agentic AI and AI agents, into the key finance business processes. And I'm gonna start with the opportunity to cash process and move through to what we're doing in report to report, as well as in financial reporting. So first up, we acquired Evasort last year, as you may know, and we launched two AI agents. The first was Workday contract intelligence agent, and then the second was the Workday contract negotiation agent. And both of these were powered by Evisource AI. And this incredible AI native Evisource technology is the foundation for a new agent, that is currently with early adopters, and that's our revenue contract agent. And this agent automates the next steps in your document journey. And if you're a Workday customer, you'll have heard our vision around document driven accounting, and this is that, you know, kind of first, release of one of those agents. So the revenue contract to agent transforms static revenue contracts into dynamic strategic assets. It automatically extracts and analyzes critical contract data and automates the downstream activity by recommending things like billing schedules, revenue recognition, and it can create the accounting entries for you. And this frees your team off from doing all of that manual work, and it gives you the insights you need to make better decisions. So I have a, short demo that I'm gonna play now, to give you a look at the solution in action. In the revenue hub, we see all contracts and billing, and we see document intelligence in action. Let's take a look at the documents requiring review. So for those of you who are not creating contracts via integration with your CRM, your revenue accountants review the contract in PDF form and manually capture this data in a spreadsheet. But as we all know, anything manual is error prone. Fortunately, the agent has done this for us. It's extracted the data from the original customer contract and created one in Workday, and we can compare the two side by side for easy review. To the right, as Workday is indicating the contract may be eligible for a discount. But first, I want to check the order effective date listed in the original contract. So I'm confident the date and Workday is correct. Next, let's go ahead and review the suggested discount. Our agent can traverse multiple documents that impact the customer agreement, including the MSA, which is the source of the discount information. It indicates that this new contract exceeds the minimal annual revenue volume threshold, and a 20% discount can be applied. But before we accept the discount recommendation, we can navigate directly to the MSA to validate if the information is correct. Now let's go ahead and accept the recommendation, and the agent takes action to apply the discount rates across the customer's billing schedule. Okay. So that's the first, agent, that I wanted to highlight. And, you know, while we're highlighting the agent's ability to draft contracts, and address volume discounting, In the future, we see the agent using multiple documents to enrich contracts with details that may not be in your CRM, make revenue recognition recommendations as I mentioned, create accounting, and even update it when amendments happen, addressing some of the most painstaking parts of your contract management. And then with the help of the revenue contract agent, we expect to improve time to billing readiness by 35%, thanks to expediting that contract setup that you saw at the start, and reduce errors when customers create contracts by 60%. So time back, fewer errors, and faster forecasting. And this goes back to what I mentioned a few moments ago around, our strategy with AI agents focused on delivering value. So just as with the opportunity to cash, cycle, we're focused on, you know, supercharging the record to report process with automation and AI. We often hear from our customers, you know, account reconciliation is a major headache. It's really manual, time consuming, and full of risk. And our upcoming AI power reconciliation enhancements are designed to fundamentally change that. Our vision here is to deliver real time accurate and autonomous account reconciliation. And because it's embedded in your ERP, it maintains full end to end data lineage, which leads to a faster, more accurate reconciliation process and greater audit confidence. In fact, AI can handle most of the heavy lifting for you, reducing the time your team spends on manual reconciliations by up to 70% dependent on the industry that you're in. So with the new reconciliation workspace coming in '25 or two to early adopters, we're delivering the first phase of this vision, and we're gonna continue to build on this in future releases. And at rising, we announced the financial close agent, which will build on our AI power reconciliation to kind of further automate that downstream period close task to give your teams more time for strategic work. And if we look at another area now so a few minutes ago, we spoke about, you know, the key areas that we're focused on from a product development perspective, when it comes to agents, compliance, efficiency, and insight. So we've seen some, areas around efficiency, around account reconciliation, as well as the revenue contract agent. And now, we're gonna talk touch on the topic of compliance and how agentic AI can help you to mitigate risk and achieve unparalleled financial confidence. So this is using, our financial test suite, which is powered by Agentic AI and powered by an agent. And it's designed to elevate the financial assurance that you have today within Workday. And a financial audit agent then will automate all of the audits audit evidence collection. And together, you know, the goal of this product is to help you to validate compliance of your business processes and your transactions, with your company policy, regulatory requirements, and standards, proactively detect errors and fraud to prevent overspend, losses, or reputational damage, and then to achieve data accuracy and completeness. So with trend analysis, and more to help you accelerate that close and bring greater confidence in an accuracy into your financial results. And the last one, the audit agent will expedite that audit preparation and response by automating all of that audit evidence collection in support of an auditor's PBC request. So in a moment, I'm gonna share a short demo, of the financial test suite as well as the financial audit agent, and just to give you a look at how these can help you gain a greater level of scrutiny over your financial data as well as to identify and resolve anomalies and, accelerate the audit process. I'm just gonna play a short demo now for a second. So here we are in the oversight dashboard, which provides insights into financial operations, transactions, and data. Key insights are at the top, which require attention, and the operational areas below provide insight into the control, confidence, and risks. IC Procure to Pay has identified four risks. So as we drill in, we see there is a duplicate invoice that could put us at risk for overpayment. And as we review further, we see the duplicate payment has been stopped by the agent. I follow the agent's recommendation to shoot off an email to the suppliers to verify the invoice history and payable balances. I also see a recommendation from the agent to explore additional tests that can prevent this type of issue from happening again. So here I am in the test marketplace, and we can see the workday provided tests as well as tests recommended by the likes of KPMG and technology provider, Canis. I go ahead, and I'm gonna enable the recommended high risk supplier verification test to continuously monitor supplier master data going forward. So not only is the agent running tests to look for unusual activity, but I can also review requests from my auditors and begin to assemble sample data, essentially preparing a PBC response. So now I'm in the audit request work area, and we can see all the sample data requests that require review. And while we're in the work area, we're receiving emails from our auditors. I'm gonna go ahead and initiate a new request with the agent. So to do this, we engage the agent in a conversation. I simply enter the details I'm looking for, an AP cutoff sample, and I want the agent to include all of the related details like invoices, payments, delivery dates, and contracts. The agent reviews the request and immediately begins to collect the sample data and quickly see surfaces of a sample selection. It looks like a great start, so we notice the next portion of the request, the parameters and assignees where the agent pulls in the company, year, reporting period, and creates a name for the request and assigns the preparer and reviewer. We see the agent has selected the company based on the description. We complete a review, and the agent begins to assemble and complete sample data request. And then within seconds, the sample request is complete, and I can complete the review. Alright. So, I hope that gives you an idea of some of the value that the financial test suite can bring. So this comprehensive suite of tests will not only spot errors from their inception, but also helps to prevent future errors and fraud from occurring. And then, as you've seen in the demo, our financial audit agent, which is, in early availability with customers at the moment, will help you save significant time when it comes to those auditor requests. I'm gonna now look at, one area. I know that somebody was asking, a few minutes ago, and this is the area of financial reporting. So if you think about how many times, you know, you've needed to pull a quick data point from a report, you had to wait for it, or you weren't able to find the information that you were looking for. And that's where conversational reporting comes in. So you can now, using the ask Workday assistant over on the right hand side, ask any question you want about your data in plain English. Not only do you get a clear answer, but you also get an explanation helping you to build trust in that data. And it even predicts what you might want to know next. So I've got a short demo, for a couple of seconds that I'm gonna play for you now. I'm in a supplier contract activity report. Now I click on the ask or to icon and start asking which software contracts are out of compliance with contract terms and by how much. I get an answer right away with a helpful table showing all of the details. If you want to know how I figured that out, just click the see detail information and check the logic at any time. Next, I ask, show me consulting contracts with higher than expected burn rates. I can see that there are a few items. It also helps me calculate when the spend would exceed contract terms so I can prioritize where to follow-up. I can also get to a full screen immersive experience when I really want to get into the numbers. As you can see, each question gives me quick, clear answers. Okay. So as you can see, this new solution will allow you to get to the information that supports you in making those decisions, and can help your decision making process and get to the answers faster. You know, when I spoke a few minutes ago about our strategy, it was about developing agents that deliver transformative value for customers. And there is tangible, measurable value for every single agent, and this value will continue to increase as those agents evolve. So our early adopters, of both, the revenue contract agent, and the financial, audit agent are already seeing huge value with great opportunities to reduce contract management costs by over 150,000 with revenue contract agent and then nine hundred hours saved annually with the financial audit agent. So these are really transformative outcomes, and we expect, you know, to be able to share more about this, as the solution, expands out to the wider customer base. And, you know, our strategy is more than just building, you know, amazing AI capabilities. We build to deliver value, and there's an entire agentic ecosystem. And I'm just gonna, you know, wrap up kind of what Workday is doing before I hand over to Brian, and he can share, what KPMG is doing in this space. So Workday will build agents for our customers, and then customers can build, deploy, and manage their own agents with the Workday FlowWise agent builder. And they can also discover prebuilt trusted partner agents through the Workday marketplace. They can manage all of these agents regardless of their origin at scale and with security and compliance. And I know somebody was asking you about that a few minutes ago, at the core with the agent system of record. And then they can orchestrate across agents through the agent gateway. So I'm gonna hand back to Brian now. So he's gonna highlight, an agent that Workday, or that KPMG is building, you know, to further add to, Workday's own product investments, as well as to share some practical steps, around AI, strategy and adoption. Great. And before I do that, there was a really interesting question in the q and a here around, how do you manage to make sure an agent is doing what you asked it to do to ensure business and IT controls, make sure it's working as intended? This is a great question we get all the time. And I I sort of just think back to, you know, what would you do if it was a real person? So the first thing you you think about is what security are you gonna give this agent to make sure they have the same security that you would expect, based on a function of what you expect that agent to do, monitoring what that agent is doing. So through audit logs and and and and and making sure they understand what what's happening, putting in the right, control points. So so I think, Marion, you in your one of your examples, you talked about applying the discounts. And so there was a step where someone had to approve that that they wanted the discount to go out the door. So inserting those, control points, if if you're posting a journal, even if a human's doing it, you always have a a second person potentially reviewing and approving that. So so it's really designing the controls very similar to, you know, treating them like like their individual people and this agent system of record. That's why this agent system of record is so important because it allows you to track both the your people, assets, as well as your digital workers as well. And so, that that's kind of, you know, how I would sort of answer the question. I don't know, Marion, if you had anything any other wanted to make on that specific question. Yeah. No. I completely agree what you're saying. And, you know, again, it goes back to that human in the loop strategy. So there always is somebody that has to approve it, and you can always validate the results of the agent. You know, in the example with the revenue contract agent, or you can go in and you can validate, by looking at the source document. So you can always validate the information. A human always has to approve it. And then, also, we build reports into the process. So there's a there's a workflow, behind each of our, you know, agent capabilities, and there's also a report there so you can see each of the steps that were taken so that you're able to, you know, ensure confidence, in that data and what the agent did. Great. Okay. So let's talk a little bit about our, the future of financial close. And I want to introduce what what KPMG is working on. I'm really excited about this. This is something we're calling the KPMG financial close companion. So once again, this sits in the automator category of our Topco framework, and it's a role based agent really designed to help financial controllers process tasks on the month end close checklist. It's that simple. And it's really designed for for three things, drive efficiency and consistency, reduce time on repetitive tasks, and reduce mistakes. So, before we get into a quick demonstration, I wanna talk to you about what it it actually does. So so the first thing it does is it reads your financial close checklist. So right now, it's trained to look at the workday period close business process, BP. In the future, we are looking to incorporate things, such as, like, Trintech close, because we recognize that not everything will happen in Workday for a financial close. But it basically it it it, it looks at that Workday period close, business process, and it processes the task on that checklist. So the agent does this by accessing Workday through its APIs to execute task and retrieve the results. And then once it's done that, it forms analysis of the information received similar to what a human would do. So on the screen here, you can see the first set of features that we've built out. And and in a moment, it will show a demo of two of the features. The first one we'll show is fixing operational journals. These are from the Workday subledgers. And then we'll show an example of how do you analyze the trial balance and how the agent will do that for you. And so the, you know, the agent's trained to analyze errors and make recommendations. You'll see two examples. You'll see, journals that are in error status that won't post the GL and how to fix them. And then we'll do things on the trial balance review. We'll do, we'll look for simple expected results such as, control accounts being zeroed out or correctly signed asset liability, accounts in terms of, you know, positive or negative balances. We'll look at expected month to month fluctuations. So that's what I mean when I say look at the trial balance and make sure it's suitable for the month end close. In terms of the, the technology here, so we've actually built this agent using something called Google agent space. So Google agent space is their platform for building the most sophisticated and complex agents. And so what the way this this actually works is the UI is is Google, and you, you you can see here this is this is a screenshot, but this is, the the agent would receive instructions from a user. Now this agent could also just move through the the closed checklist on its own and start executing things that they can do. But, there's also an interface for the users to interact with it. And then, what it does is it calls the workday APIs that are exposed to the agent gateway that Marion just mentioned on the previous slide, and then it and it looks at the results. And on this screen, you can see we've got 11 different items that this agent can actually handle right now. So with that, why don't we take a look at this agent real quick? Today, we are thrilled to introduce the KPMG Financial Close Companion, our innovative concept designed to streamline and enhance the month end close process. On the screen, you see a Workday home page. I will navigate to the period close event business process to show the to do steps within our month end close process. With the help of the close companion, users can easily request virtual assistance for monthly close processes and manual workday tasks. The overall goal of the agent is to be able to boost productivity, enhance efficiency, and allow accountants to focus on more value add activities during the close. In this scenario, an accountant is able to engage with the financial close companion to correct operational journals with errors. By asking the agent, what are the operational journals with errors? The accountant can obtain suggested solutions for addressing any journals that have not yet achieved posted status. The agent will then provide information on journals with errors. This information includes detailed analysis by journal ID and shares that there is a missing ledger account with the spend category, software rental leases. The agent proposes the ledger account that should generate with this spend category, software expenses. The agent also recommends for the spend APR to be updated and to rerun the fix operational journals task. In the next example, the accountant asks the close companion to review the trial balance, and the agent will perform variance analysis. Now that our results have populated, you can see that the agent identifies clearing and suspense accounts that should have a zero balance, but do not. Additionally, the close companion suggests solutions for these errors. This view and level of analysis allows for the accountant to ensure the trial balance is thoroughly reviewed and all discrepancies are addressed. Okay. So, hopefully, that gave you a little bit of, a vision for how this agent will can work in practice and really accelerate and and and and help the controllers legal and aid controllers and accountants with their with their month end close activities. Alright. So finally, the last section here, we wanted to you've heard a lot of information today from myself and Marion around all these different AI agents and models. So you might be thinking a lot of information here. How do I get started in a practical approach? And so here's here's six things that we advise in terms of how do you get started. So the first is to to think through your strategy. How much do you want to leverage Workday AI versus third party versus rolling your own type of solutions? Most of the organizations that we've been working with, have followed a hybrid. They wanna have some basic capabilities coming from their core ERP, but they also want more. And so I think, you know, really thinking about how much do you want to to be in each of those different category buckets is is really important to start out. And then the the next step in this is is really experiment. Run pilots, see what works, find areas where you can get good return on investment with with low risk, predictable results, and and that's really every organization, is doing you know, most of the organizations are doing this, today as as we spoke earlier. Number three is looking, very carefully at KPI and performance measures, and this sort of ties back into that question. How do you assess whether this agent's doing a good job? We think that you should be evaluating these agents. Are they getting the answer that you intend more times than not? Are you, how are they hand are they getting the the outcomes? Are they learning over time? And so, you know, carefully deciding how you want to manage the agents is something to really think through from a performance tracking perspective. Step four is to really think about then how do you scale these. So one of the challenges we see is you you you do this in a small scale. How do you replicate it? Because oftentimes, where you get the real value is being able to take small amount of savings and apply it more holistically or more broadly. And so, oftentimes, maybe you'll pilot this on one business segment and then roll it out more broadly. So think about how you wanna scale this across your organization. Step five is all around testing and validating. So part of this is is part of the scaling process, but it also means that you you need to come up with a capability to certify and test these agents, almost stamp them with seal of approval so that these agents are are legitimate. They can be rolled out in the in the organization, and they've they've been through a, control and a trust verification process. And then and then last but not least is looking at a robust process to make sure you can operate and sustain and grow these agents over time. So, you know, I I would I would I definitely encourage, everyone to be thinking about these six steps as you're thinking about AI more broadly and specifically for this conversation, within the finance function. So so with that, why don't we turn it back over to Marion wrap up, and and, I think we'll open up for some q and a as well. Great. Thank you so much, Brian. So thanks for sharing your perspective. I know a lot of you know, based on the q and a, and some of the chat, I know a lot of people are really interested, in some of the practical advice and tips that you were, willing to share with the audience, so we appreciate that. So today, we explored agentic AI. You know, it's not just a technological trend. It's a real shift in how we work. The key to success is not about simply adopting AI, but embracing it to actively create future value. As Brian mentioned, prioritize AI use cases that solve real world problems and enhance existing processes. Don't just implement AI for the sake of it. Ensure it directly addresses your organization's pain points. And then lastly, before you begin, define your success. So Brian mentioned, you know, have clear KPIs and performance measures so that you're able to track the impact of your AI initiatives. And this ensures that you can demonstrate a tangible return on your investment. And then, you know, one thing we always talk about, and this is key to our strategy, Identic AI is most effective when paired with a skilled human workforce. So upskill your team, build pilot programs, and this is where we've seen customers have a lot of really great success. You You know, they can put an ambassador in there, to drive, part of that program, have a road map in place, get feedback from the accounting team, and really foster that culture of continuous learning, you know, so that they can get, the best value out of these tools. And that kind of brings us to a conclusion for today's presentation. There was a lot of really interesting questions in the chat. Brian, I don't know if there was any that you've seen. I've seen a question there around, you know, security, when using Workday's AI tools. You know, like any Workday product, it's all built, using Workday security model. The agent system of record, you know, has a really secure platform, to ensure the confidentiality of your data. So, you know, you can read more about that on workday.com. And, Pam, if you want, I'm happy to share some more information with you. Brian, I've seen a question that maybe you wanna take. I'm just trying to, you know, find in here. I think it was around, you know Randomness of the AI agents, maybe. Yeah. Yeah. That was the one I was looking for. Yeah. There's a question about randomness and hallucinations and how do you rely on the results even if you've tested it a and I I think that's been one of the issues certainly in in in financial services and finance about it's not good enough to be 80% correct or, honey, you'll be a 100% correct on every transaction. So I think the, the the randomness in the AI agents at this point, the way to best address it is to make sure you have those human control steps before. Like, you'd never would let I I would never let an agent actually post the transaction or commit a transaction through a process. There would always be somebody that's reviewing that, which again might sound like, well, then what's the whole point if if I gotta have someone there double checking the work, then am I really gonna get the savings? But but I do think that, at least early on until these things become more deterministic, then I think that's that's a good place to to sort of start and and, even if you even if you sort of tested it a few times. But, that that's a that's a really that's an important consideration is is try to is is trying to make sure you understand that. I think I think what I've observed in in some of these agents is you get a lot you can get a lot of false positives where it's it's doing the analysis. It's pointing out something, and you say, oh, that's I expected that variance to be higher for this month, or I expected this. There's a reason why this this, control account is is is not, zero yet. So I, you know, I I understand that, but it that's that's a good question. Great. And, Brian, there was a question some people have asked, you know, to learn more about the financial close companion. I will make sure that we get the details from KPMG, to be able to share some information so you can continue the conversation. And there were some questions around There's a link yeah. I was gonna say, man, sorry to interrupt you, but there is it is on the marketplace. Like, there's a link on the, the Workday marketplace for the agent. I don't know. Maybe we can I don't know if there's a way to drop that, link in the chat? I I can maybe just drop that in the chat here, and you can always follow-up with me directly on that. But it is it's something if you're interested, there's a like everything else on the marketplace, there's a place to submit a request, and it'll get accepted back to our team. Okay. Perfect. I think that was all the questions. I saw one about early adopters. Somebody was asking about how they could be, on the early adopter program for the financial test suite. Oh, yeah. Ali, I did see that, and I will get your details, after the presentation today and, make sure that I can, link you in with the right product team to follow-up on that conversation. I'm trying to see, There was a question around the financial test suite and the audit agent, Somaricio. These will, follow our agent, flex credit model, and I can share more info about that, as well. So I think there's a there's a couple of niche questions, in the chat. So what I'll do is I'll make sure that we can get a link to all of these questions, and we can follow-up directly. And for anyone who's, you know, wants the KPMG information, we can link you in with the KPMG team as well. But I just wanna say thank you all for your time. We hope you found it valuable, and we look forward to seeing you again at future sessions. And thank you, Brian, for sharing your expertise in the call today. Yep. Thank you very much. Great. Bye. Thank you. Okay. Bye, everyone.