Name: Planning Agent in Action with TX Group
Uploaded: 2026-04-16T18:16:20.989Z
Duration: 56 min 58 s
Description: Planning Agent in Action with TX Group

Transcript for "Planning Agent in Action with TX Group": Hello, everyone, and welcome to today's looking forward with Workday planning agent in action. So today, we are going to do a deep dive into Workday Adaptive Planning and specifically our recent release of the planning agent. And we're not just gonna talk about it. We're going to see it in action today with a live demo, and we're also incredibly lucky to hear from one of our early adopters, TX Group, about their firsthand experience using the agent and the tangible impact it's had on their business operations. So just as a quick disclosure, safe harbor, before we dive in, we will be talking about some unreleased services features and functionality. Because of that, we ask that any purchasing decisions you make regarding Workday services be based on features that are currently available. And I will say this is especially important. We are we are definitely gonna be talking about some of our generally available, functionality within Workday Adaptive Planning, but also wanna give you a sneak peek about what we're developing for future innovation with Planning Agent today as well. For housekeeping, if this is your first looking forward with Workday, a few quick notes before we get into the agenda. Today's session is being recorded, so you will receive a copy of this within your in your inbox within twenty four hours. We also want today's session to be interactive. So as you can probably see, we have a q and a tab on the right side right hand side of your screen. We'll be keeping an eye on this throughout the event, answering your questions, and we're hoping to save some of the most popular questions for live discussion at the end. And then finally, please stay for a brief survey at the end of the webinar. Your feedback is really important to us, and it also helps me and some of the team make decisions about future sessions that we're gonna have here at Workday. So great. So agenda. So we've got a we've got one hour, and it's going to be really packed. So, we've already gotten the first one off the list, which is great. So welcome is done. And so next, I'm gonna do a little bit of level setting, around Workday Adaptive Planning and also what we're seeing within FP and A trends and around AI. Then I'll also be joined or then we'll we'll be welcoming Ketanjani, one of our fantastic principal solution consultants for live demo. And after that, Demetrius from TX Group will join me for a fireside chat about their journey, as an early adopter of Plan Me Agent. And then we'll close out with a few remarks. So should should go by really, really fast. And as a as a brief introduction to who you'll be hearing from today, so we'll be joined by Demetrius Nikolagu, who is the group head of cloud ERP planning at TX Group, Ketanjani, who is one of our principal solution consultants dedicated to Workday Adaptive Planning, and then myself. I'm Kelsey Vaughn. I'm the product marketing manager for Workday Adaptive Planning, and really excited to be with you here today. Okay. Great. So what are you gonna learn? Our goal is for you to walk away with three things. First, an understanding of the top global trends for AI within the FP and A function. Second, I think we we all we all know there is so much going on around AI that and and all the different functions today. So I wanna give a little bit of background and clarity if this is your first time really exploring what technology, what AI means, the FP and A function. Some some of the trends that we're currently seeing within AI and to be able to distinguish the difference between what we talk about when we're saying machine learning or generative AI or agentic AI. And then finally, we wanna give you a clear picture of how Planning Agent, one of our recent releases, works and then also hear about what it means to use a a real company. So great. So we're gonna just dive right in, for a few of the trends that we're seeing in FP and A today. And what are we seeing in the market? I I don't wanna spend too long on this because I think more than anything is, we can put numbers around this, which I always love, but I think we're sensing this at our own at our own companies, at our own organizations. And what we're seeing is is simply a massive surge in adoption. So according to global surveys, 44% of CFOs in 2025 said that they use generative AI for over five use cases last year, and that's a staggering jump from just 7% in 2024. So we're seeing huge huge increases in adoption. And I think what's really interesting from an FP and A perspective is that seven out of the top 10 AI use cases in finance for delivering ROI value, a separate study found are happening within FP and A. So when you kind of tie that all together, it makes sense why 58% of CFOs in 2025 ranked FP and A as the area in most need of transformation. And I think intuitively, it makes sense. FP and A really is data rich. There's so much we we can we can look at, data that impacts our individual performance, key performance indicators, and it's also highly collaborative as a function. So it's kind of that perfect place for AI to drive value within the finance function and the business. So AI used today, what are we thinking about that? Like, truthfully, when I'm thinking about AI, I try and think how how do I really translate that to something that I use every day? So I think about my car. And so when I think about AI and AI innovations within, within the planning function, I I think first about, machine learning. And that I think of as your kind of your rear view mirror. Machine learning, it looks at historical data to identify patterns and predict future performance. Within Workday Adaptive Planning, our predictive forecaster tool does exactly this, takes your actuals, your historic actuals, and other datasets to recommend a forecast, with within a clear confidence in your bill. That's based off of historic data, and that's that's just one part of the equation, but it's a really important part of the equation. Next, I think of generative AI as your kind of your GPS. It takes data and trans or, takes data and translates it into natural language. So, this allows you when you think about, for example, CHAPPGT or CLOD or anything else along those lines, allows you to submit a query and get a synthesized response back that helps keep you going on the right path forward. And then finally, when we think about Adjemtik AI, which is the one that I think everybody is talking about today and where is the driving value, Adjemtik AI is kinda like your steering wheel. This is where we move into autonomous workflows. So agents that can execute multi departmental tasks to reach strategic goals. And that's really where I think why it's so important to be talking about AgenTic AI today within FP and A because FP and A is so cross functional. That is where we can see real value emerging, from the from from the AI, technology transformation. And so, I think just another clarity piece that we like to say too is that, AI is really it's not a magic wand. So it's great. It's an incredibly powerful tool. It can do a lot and drive value, But AI on its own can't can't actually solve some of the challenges that we see within FP and A, a lack of collaboration, lack of silos. So if your data is siloed in different systems, AI isn't going to be able to find the signal and the noise. And, additionally, without baked in security and governance, AI becomes a liability instead of an asset. So these are things that are really important when we think about AI and the tools itself is that AI is not your is not your magic wand. It's not gonna fix everything, but it is an opportunity to really drive value. And And so what do we think about that in terms of how we're building AI within Workday Adaptive Planning? We're really thinking about a shift. So making that shift requires a new foundation, and we're calling it limitless decision intelligence. At the core is one decision engine, not stitched together. With tools, your customers can really work wherever they want, adaptive powers, intelligence underneath. Second is boundless exploration. AI changes planning because it removes the cost of curiosity, the ability to layer in any data, model any scenario, test decisions instantly. And then finally, what we're talking about today is planning agents. So planning happens when business needs answers, and those agenda workflows and having that baked into the system itself, really allows you to do that when you need those answers. So our vision for agents is really to support planner planners and analysts and helping them be more effective at what they do every day. Human judgment will always be part of the process and will augment with AI driven intelligence, but in the end, we're driving towards faster, smarter, more connected experience to drive better outcomes for you and your business. And we believe, really when we come back to Planning Agent, what's our vision around that, it really starts with a set of configurable skills that mirror actual jobs. So we we've really designed planning agent to be one agent. Within that, we are designing skills to help you do your jobs, but we've kind of we've we've framed them in a way that actually mirror the work you're doing. So for analysts, we're looking at building skills that explore data, highlight anomalies, monitor KPIs, generate board ready insights. For planners, it accelerates scenario creation so you can respond faster to market changes. For modelers, it simplifies architecture, suggesting formulas, and optimizing your structures. And for admin admins, it makes the system accessible to everyone throughout natural language while monitoring system health. And so together, this is our vision for for planning agent. And where have we started? So today, where we've started, we already have some of this generally available within the product today. So the natural language interface, so you can interact with the platform conversationally. Contextual help, so for example, if you forget how to build a how to build a report, you can actually ask a query within the system and get step by step guidance directly from our documentation. Data exploration. So you can perform ad hoc analysis and get instant intuitive visualizations, variance analysis. You can ask where and why variances are happening, and the agent will automatically surface the drivers for you. And part of that, testing and that build of what we did around agents, we we actually did in partnership with over a 100 early adopter customers. And a couple of quotes that have been really, eliminating in terms of the value that it's brought for their team, that planning agents, the skills we've developed so far have taken the stress out of data. They that by getting in some answers, they will find their insights faster, and you can focus on strategy rather than a manual entry. So really bringing this all together as we're seeing a way in which our planners are changing how they plan. Really excited, to have planning agent finally generally available for our customers. And, the same is leaving. So now I'd love to invite Ketan Johnny to come up and actually show you Planning Agent in in a demo. Perfect. Hi there. So so here we are, Sam. And what we're gonna do over the next ten, fifteen minutes, so is I'm actually gonna take you through planning agent today, going through what is available and how our customers are using the planning agent to really help them drive decision making and drive some of their processes within Workday at Applit Planning. So at the current moment, you can see I'm actually logged in as Patrick O'Brien. Patrick O'Brien has access to one of his planning hubs over here, which gives him a number of different tasks that he performs on a given day or as part of a budgeting or forecasting process. Now with the planning agents, it is something that is fully embedded within Workday adaptive planning. And what this means is the user effectively just has access to the agent where, currently, they have a number of skills that the agent can actually perform. So here, what we're seeing is the ability to summarize a report, I e, do some very simple exploration, detect anomalies, where have I fallen outside of certain boundaries within the data itself, but also being able to identify top variances to understand what are the root causes behind any variances or significant anomalies, within the data itself. So with that in mind, the user is presented with a conversational type user interface, very similar to what we see out there in the world today. But what this does is it gives you the conversational interface over governed and trusted data. Alright? So all of the data that is sat within Workday Adaptive Planning itself. So the user could effectively just simply click on summarize this report where the agent will now go in, look at this report in context of what is currently defined within the report. For example, here, we're comparing the working budget against my actuals and where we've got a number of different variances. We're seeing this at a total company level against a full set of income and expenditure accounts over on the left hand side. So my revenue, my expenses, and allocations, as well as any operating expenses within the report itself. So what the agent will do, as opposed to me having to either go through the report, try and identify or summarize what I'm actually viewing, The agent will come and explore and analyze all of the data within the reports for me against the context that is defined within the report. So, effectively, what has come and done is it's given me now a reporting summary where it's analyzed the overall performance. So net income for actuals was x exceeding a particular budget, which shows or highlights a 9.47, positive variance, as a percentage. It can also then go and identify any key drivers of this variance. Okay? And what it's really effectively doing here is analyzing all of my lines of data within revenue over here without again, if we think about what it how what the user experience might have been previously, I would have potentially had to drill through this and view the data and come up with my own type of analysis. However, here, the planning agent is enabling me to automate that typical repetitive task of mine. What it also means is the ability to ask questions in natural language means you're able to take planning out to your non planners. Right? Users do not necessarily have to learn how to navigate within the reports. They can simply ask questions of the data, and the agent will come back and analyze those, analyze the data and provide the insights for the nontechnical type of user. It also eliminates the requirement for anybody to potentially learn how to write a report. Okay? So, typically, if we think about a given example in an Excel type reporting world where the user wants to ask more difficult questions of the data or needs to summarize something or needs to drill into something, they have to effectively write a new report. Right? The planning agent will help reduce that reliance on having to learn report writing, having to learn how to configure any reports themselves. If we take the summary of top takeaways, what we could do is effectively from within here, I can copy that. And what I might wanna do is send this out as a notification. We'll send it out as a comment against the data, and I'm gonna tag this for Rachel Knight. Oops, Daisy. Tag this against, Rachel Knight. Paste it in there, and that then is now held within this report itself against the intersections of data of of the comparison in q one twenty twenty seven at a total company level. Alright? So, again, use as a planning agent, using what we've taken as a result from the planning agent, we can then start collaborating and tagging, users with this. And this can then come up as a notification in the likes of a Slack message or indeed in the future moving towards a Teams type message. If the user wants to, they have the ability to further ask questions against this. So what we could have somewhere here would be what products drove 04/01/2000 product revenue for q one twenty twenty seven. So, again, just simply asking questions of the data, and the planning agent will go back and analyze that and provide certain explanations of this. Again, what this means is the user does not have to learn to write a report. And here, I've got that analysis, and what I could do from within here is take that even further. So here, it's taken the product revenue line, analyze that further for me. What I could do is maybe get it to analyze a couple of different revenue lines. So, again, it could be compare 4,100 product revenue versus 4,200 services revenue for q one twenty twenty seven. Alright? So, again, just very, very simple natural language providing me the ability to converse with the agent and enable the agent to answer those questions for me. And what this means for our customers is reduced time. Yeah. Time savings in being able to provide those answers to them with the ability to thereafter say go and auto generate all of this analysis into a presentation for me. And if I just open up the PowerPoint, what I've now got is the results of that analysis available to me in a PowerPoint document, which when it eventually opens up, here we have it, and I've got everything with the respective narrative. And as an end user, all I'm gonna do now is maybe spend a bit of time, formatting this to my individual requirements. So what we've seen then is how a user can use the planning agent to explore data within a report. Alright? As you've seen, the AI assistant or the planning agent is fully embedded within the application itself. Okay? Furthermore, we could even say, right. What I want to do is go and identify my top variances. Okay? So simply saying, identify my variances. It's now identified those for me where I could now, from within here, start drilling down and say, go and analyze that further. And me simply clicking on analyze this further, it's automated or actually written that into natural language for me to provide the variance analysis for 4,100 product revenue q one total company, and it's produced this nice little pie chart that, again, in a similar fashion, I could simply say going on to generate this into a PowerPoint, or I may say show me this in a different type of visualization. With the ability to drill even further. So go and explore some of the various drivers that it's picking up, from within here, and it could be, like, go and look at this by, I don't know, let's look at it by location and see what it picks up from that perspective, and it's taken that further for me. So what you've seen here is how the planning agent really can help our customers, one, today, really drive speed to insight. Yeah. Empower your business partners. Enable them to come in and self serve themselves information without having to learn a given tool, without having to learn a particular report writer. And thirdly, key takeaway here is what it will provide back is a set of trusted and governed answers. So, again, the solution uses governed data. So data that has been governed when it's been entered into here or actuals have have come into the application, but it will also respect the security layer of Workday and Planning. So with that in mind, that that's what I wanted to present to you today in terms of the planning agent itself. I'm gonna hand back over to, Kelsey. Excellent. Okay. Great. Thank you so much, Ketan. That was great. And I do know we had in the chat, it might come through again. If for whatever reason you can't hear me, it's that same unmute tactic. We'll have someone put that in the chat here. Augustine, if you wouldn't mind actually dropping that in, that would be that would be great. We we heard we heard last week that that that could be going on, so appreciate the the the patience with the technical difficulties there. Excellent. So we have so many questions coming in, both in the chat and q and a. So if you have a if you have a question that you want to be answered, please feel free to post it in the q and a. We're we're starting to respond to those. So, please keep them coming in. It's really great to see. And, yeah, hopefully, that demo was a good, introduction to what the experience of actually using Planning Agent looks like. So now we're just we're going to move on to the next portion of today's session. It's it's my favorite portion is actually being able to hear from someone who's been in the weeds with planning agent for quite a while. So, I'd like to invite Demetrius to join me. We're gonna talk a little bit about, not only TX Group's story, but also your experience, being part of the early adopter group. So you you are actually helping the task, and we're one of our customers actually helping the task and get feedback to make sure that we were delivering something that was going to, you know, not just work on paper, but actually work for customers. So, one, really appreciate all the work you did there, but also appreciate that you're here to share your story today. So thanks, Dimitris. Thank you very much, Kelsey. Hi, everybody. It's a pleasure to be to be here, to be honest. And, I'm just think some some things about me. So, basically, my name is Mitrice. I'm the group head of clientele before planning at TX Group. Some things about TX Group. So, basically, it's a company It's a media company that is based in Zurich and Switzerland, and, we've got approximately 3,300 employees, a turnover of approximately 900 thou million, Swiss francs. And, basically, our focus is, media and publishing. Although the last couple of years, we've expanded in other areas of of this. For example, now we're kinda entering a real estate. We have spent a lot of time, kinda investing in, in accelerator for startups, And, basically, we're kind of to also, become the leader when it comes to the job. The digital company for the switch of market, for example, have several websites that we're kind of owning and, you know, people can basically go and apply for a job. So it's a diverse portfolio of businesses, and we're trying to kind of, you know, steer the business, from the FP and A in us as well as from the workforce planning point of view here at this group. So we we look at the planning in general, and, basically, the main the main problem we used to have in the past is basically that the the the the business are very diverse. So, basically, the advertising business is very different, for example, to a media business. And the media is very it's very different, for example, to to a marketplace business where basically the focus is is very, very different. As a result, the affinity teams have the same problem. So, basically, we we sitting at the group level, we need to basically have a a way to understand, and translate, into numbers, of course, all the different business and to support them throughout the common platform, which for us, for example, is adaptive planning. And, for this policy, biotechnology departments, you need to kinda move from this additional reporting approach to a more modern approach, which is should be focusing more on modern solutions, like, for example, the planning agent that that that that the plan has has has launched or, for example, to be able to create more value for our top management and for our shareholders. In the end, what we're trying to do here is basically to we're trying to to to make better decisions as a group. And, for us, the most important thing is to serve the business a posteriori, of course, but most importantly, forecasting the outcomes for for the group. Yeah. Great. Yeah. And I know, yeah, I know that that that's something that I've always found really interesting about your story in particular is just the the the the complexity and the like, how many different individual models that you have to build as as an organization to really drive performance. So yeah. No. It's it's so interesting to hear that. But, yes, would love to hear a little bit more about your early adopter journey. So I I think I've mentioned this a couple times. So TX Group was one of, a little over a 100 global customers we had in our in our early adopter group, and would would love to hear from your perspective. What was that experience like? How did you and your team work alongside our prop, product developers to ensure that the agent was actually, you know, solving, those kind of real world skills and complexities that you and your analyst fill, you and your analyst face. So things like, you know, either data exploration or or variance analysis. We'd love to hear about that. Sure. Sure. First of all, I would like to highlight that you have participated in more than, I think, five early adapter programs and more than seven, design partner groups in the past, which I think is a is a great opportunity for us, you know, sitting at the customer level to try and be able to test in advance future that are gonna come in production within the next couple of months or years. That's very important. Now regarding to the planning agent, as such, we started this journey last year, and, it's basically, it came with within in stages. So, basically, first of all, we had the opportunity to get to kind of this basic for example, you know, the this work work work in support window. For example, Kevin was showing previously, for example, we tested. We tried to kind of understand the prompts there to see how the prompts work, how they respond to what we get as an answer. And, basically, this was, of course, with the people in the in in The US. So, basically, we try to bring them real FP and A problems we have on on our daily kinda routine. For example, as you said previously, Kelsey, variance analysis is a very important topic. I mean, everybody does BBA during the month of close and, especially when she or she sits in finance. And then for us, we tried basically to understand how the variance analysis of of a planning agent works and functions. And if, for example, we get the same results versus our actual analysis, basically, that, you know, our IFNA guys do on a on a monthly basis. So, basically, we focus heavily on that. We focus heavily on the, for example, how the agent can analyze P and L during a monthly close and how an agent can basically compare a forecast, for example, version or a planning version or budget version versus the actuals. There were some, of course, back and forth. We, you know, provide our feedback. We, for example, focus on materiality. For example, sometimes, as everybody knows, you know, there's a material variance, but the numbers are not material. So, basically, we will end the highlight on topics. And, at at the end of the day, we have a lot of degrees of freedom to kind of, do this iterative testing and, you know, go back to the to the people in in in Workday and, you know, discuss what for us would make sense. And this is, to be honest, something, you know, for example, as a customer, it's it's very important to kinda do in advance and then seeing this, for example, feature going live. And we also have, for example, the opportunity to discuss with them and kinda analyze, for example, even the the the TX Group results with them. So that's very interesting. I mean, the the the collaboration is very is very good. It's very hands on. We have very regular feedback sessions. We're still working on it. I mean, there are some stuff coming now. I know that it's gonna come out of disconnect. So there are things that are coming are coming as well. We're testing a lot of functionalities. In general, we're planning, to kinda launch this to the whole finance community here at this group so everybody can get, can do the best. Yeah. No. That's that's great. And I think, yeah, I think that's one of the things that, particularly, like, I I really I really enjoy about, you know, being here is is that working on Adaptive is just how, you know, how closely, you know, our product development team has been working with our customers. It's it's really not, you know, designing in a silo and then and then producing. It's it's actually figuring out, okay, how can we make this, how can we make this feature when it actually does go live, like it is today? How can we make it as, robust and and effective as possible? So, so happy that you've been a part of that. I know you've been a big part of of those groups over the years. But, yeah, you're kinda moving into that next phase now. So you you mentioned that, you're moving past early adopter. The early adopter phase, we do have some of our some of those features I mentioned are are now available to go actually go into production production environment. So, your, your data exploration skill, for example, can can go into production, and I know you're talking about bringing that out to the rest of TX Group. So, what is that transition? Kind of how are you looking at that and thinking about that? I think that's a question we get very often from from folks, who are thinking about AI adoption within their own company. So what does that transition look like? And, yeah, what are some of the, what are some of the early kind of time savings you've seen from from your early adopter phase that you're looking forward to expanding out through the rest of the group? Sure. The first thing was basically to build the right tools. So for example, when you have a variance analysis, it's good to have, for example, the the multi dimensionality aspect. So you can, for example, understand and translate the the the results, into the dimensions you are looking into. And that, for example, took us a bit of time. Generally speaking, we, we basically try to to build usability here. And, when you see visibility, for example, the the reports that we have built and the tools that basically we kinda want to kinda run on a on a weekly or a monthly basis that they are kinda very friendly for the for the users. And that's what we're trying to try to do test here. And, basically, our team is also responsible for, other kind of AI tools within Adaptive. And, for example, this process of, you know, the variance analysis and the whole kinda analysis on the on the several dimensions helped us a lot when it comes to analyze the actuals because then the whole predictive forecasting, for example, process becomes much easier because we don't have to kinda go and, you know, amend some data to for example, there are some wrong bookings. So, for example, there are some wrong, some adjustment that, you know, kinda distorted big datasets. So this kinda planning agent approach helped us kinda streamline a bit this to this this area. And, we managed to basically save, I would say, on a weekly basis, approximately half an hour to kinda when we focus on this topic. And how we do this? Basically, now the the reports are done that, you know, we run them once we we take it basically the data. And then following following the following following, the run, we basically understand we have clearly from the from the planning engine all the highlighted variances that do require, some some, some, for example, control or some somebody needs to check them. And then we can basically identify directly, you know, the the outliers. So that's that's very important and help us a lot when it comes to you wanted to scale this across the team. If you can imagine, for example, adding adding this up throughout a yearly period, it can save a pause for the team and then basically gonna focus more on things that do make more sense. For example, to improve the planning process, to understand the data data better when, for example, it comes to kind of produce a forecast for the for the for the FNA teams. And, in general, we kinda try to to improve a lot of the planning quality here, which, we have done already so far throughout the year when, for example, the the predict forecast that was launched last year, We have made a little progress on this, as well. Yeah. Amazing. Amazing. Yeah. Okay. So predictive forecaster, great great call out. So, yeah, I think I talked a little bit earlier today about different types of AI and having different AI within the product itself. So predictive forecaster is our machine learning, tool that we have within Adaptive. So you've you've also been a really powerful adopter of predictive forecaster. Would love to hear maybe maybe talk about that just for a moment. How how have you used predictive forecaster, and what are some of the, how have you used the tool, and and what are some of the things you've seen by using that particular module? Absolutely. I mean, we, we're basically waiting for this because, the but if forecasting is a big topic for us because of the rest of the business, we cannot basically have a kind of an output that fits for example, a more approach that fits everything. For example, the the marketplace business has a very different approach when it comes to advertising or to publishing. For example, in publishing, when you, for example, you want to build newspaper, the driver models the driver the key drivers are very different, so we have to build different driver models. And for us, it was very kinda, convoluted to build a process that works for all. In the past, we just had a process. Basically, it was working with for example, we had they're getting data from Workday Finks, flowing this to into Adaptive. And then from Adaptive, we're working with another software, which we're very trying to produce the forecast there, and then getting back the data to Adaptive and then starting with the. This was taking approximately one to one and a half days depending on the on the on the month and if there is no kind of technical issue. And, basically, since the forecasting came in in production, we basically used it because everything sits in in I mean, the whole data is unified and basically sits in our platform, which basically adaptive planning. So, basically, we completely decommissioned the old software system. Now we're using adaptive, and the forecasting the forecast that it has using all these different algorithms, for example. Of course, we have tested everything. And, basically, we have seen which algorithms work for for which can, with the unit, we can basically diversify between algorithm and business units. For example for example, we can say, for example, the experience and model works with this business. The other one works with the other business. So that's very, very, important for us. And now we basically have streamlined a lot of workflows. We have basically improved the load wind capture time. So basically now within, I would say, three to four hours at the at the worst case scenario, We can have our forecast there. We have replaced fully some of the forecasts. For example, some rolling forecasts that we had in the past and were kind of produced by the by the people are now made with, the forecaster. Forecaster is kind of the one driving, for example, the revenue forecast. And, the top management is looking is looking a lot at these numbers because, for example, we achieved last year for I mean, I'm when I say last year, I mean, 2025 full year, actually. We achieved a forecast accuracy of 98.5%, which I mean, it's a great thing to achieve, for example, for a very diversified portfolio of businesses. And now it's part of the yearly, budgeting and forecasting process. It's a it's a it's a it's a I would say it's a very important aspect of, of our monthly forecasting. And every every every month, we go. And once we have a new actual thing, we kind of go and implement the forecasting. Yeah. No. That's that's that's huge. Yeah. And and great to see. And I love one of the things that you you called out there too is that, you know, within predictive forecaster that there are the different algorithms that you you've tested out. We have different algorithms that exist within predictive forecast or itself that you actually went through for each section of your business and actually tested which algorithm is producing the most accurate results, within that kind of span, that confidence interval, and giving you the most reliable one. So I think that that that's also such a good good, best practice for anyone who is looking to bring in any kind of AI tool in into their practice or into their organization is really, is testing before and after because I think humans being in the loop is really important. You obviously know your business better than AI naturally does, but being able to have that kind of testing approach, it can help you do your job a lot faster, but it also makes sure that, you know, we're not losing that that role that we as planners still inherently have. So, so love that. So, I just wanna touch on a on a couple of those that you've shared from both, from both the predictive perspective and also planning agent. So you've you've, you know, maybe even specifically touching around how you reduce troubleshooting time from from hours to seconds. So when you're when you are dealing with things like account variances or, deep dive analysis, so how does having those answers instantly change the the the vibe of of your finance meetings or, the work you're actually doing? Do you find you're spending do you find you're spending any time less do you find that you're spending less time arguing over data and and more time really on strategic decisions now? I mean, everybody believes in finance. I understand that, you know, the discussion about who owns the data and the whole data accountability topic is very important. And, you know, I've been in several discussions in the past that, you know, we're kinda arguing over who is this on my data. You know, I cannot recognize my data, and that's not the data I loaded in the system. And it's basically you know, this happens quite often. I mean, the the good thing is that, when we have a data repository that is unique, there is no discussion who owns the data. So the planners for us are the ones loading and owning the data. So there's no discussion about this. Discussion is over. The discussion is, who basically is gonna report the data is very clear as well. Who is producing the forecasting, data is also very clear as well. Although, in the beginning, we'll have all the challenges to to persuade them. For example, this forecast, although from a machine, it's still, you know, focused that you have to kinda explain. This took some time, but, again, now it's part of the of the of the process. There is no discussion about who, at the end of the day, is going to kind of steer the business. So now, for example, every every if any person or even the workforce client, because we're also doing workforce planning here, knows exactly what and how he will be going to comment, analyze, and, focus on. So in general, I would say that the whole flow is, is now very, very, very, very unified. The the focus is on the best outcomes, good or bad. And then the challenge you would have basically in the discussions are are solely on these topics. So we don't discuss about, you know, there's an excel, mistake or, for example, the formula didn't work or, for example, the loads did not take place. The more the whole month end closing and the month end forecasting following the closing is shortened severely. And, basically, we have now over the whole iterative process, is very, very, quicker. And we can also now do a lot of scenario modeling, which, for example, in the past was mainly through, you know, Excel files and the manual, I would say, individual efforts. Now, for example, it can also be done in adaptive and people can go and, you know, amend directly with data and see the results instantly. Yeah. No. It's good. It's good. And having everybody, yeah, being able to to, you know, answer those questions and you're not people aren't chasing you would be like, how do I build a report? Being able to to actually have that surface with contextual help, I think, is great, but it's it is getting everybody on one system. It's it's so it's so funny. I had a had a friend friend once say, it's like, gosh. We we all we all still we also live in Excel. I was like, well, you don't have to. That really like, there there there is an opportunity really to, like, to really, yeah, keep everything in that one source of truth that you don't have to be moving things around. So I think that makes a huge difference. So just before we we wrap, we I know we've gotten a lot of questions. Some of a couple have been answered, some we've we've held, but for everybody who's out there who is who's thinking about, who's thinking about bringing in agenda capabilities, or even rolling out AI within their own company. What would be, you know, the four steps that they should take to get to where TX Group is today? I mean, what we experienced at least, at TX Group is that, basically, we need to keep the the human in the loop. Regardless of the machine advancements and whatever comes in through the AI, this is, for example, Gen AI or Genic AI or whatever. I think that, you know, we have a lot of lot of challenges when it comes to the adoption of this. So I believe that if you keep the human in the loop, it's very important. The human need to understand, and I've seen some relevant questions here on the on the chat. For example, the human needs to understand and, you know, at the end of the day, approve or kinda not approve what the machine says. The machine just provide you a commendation or kinda and I will say that it kinda maybe has a different perspective. But in my opinion, at the end of the day, the human should be kept in the loop, and the human should be the one kinda, you know, that is gonna, for example, focus on the on the on the result and because he or she has to explain this. And, that's very important for for in my opinion. And the the second thing is that, we need to have a data lake that is uniform across the board. Even, for example, if you have the HR data, final data, operational data, it's very important to have one platform. Everybody depending on his kind access rights or permission sets, he or she can access the data he or she wants. And the analysis will be based on the same datasets. That's very important because there is no discussion about, you know, going back to basics. Why this, why that is not correct. I don't agree with this. And, you know, then discussion is over, and we'll just focus on the on the real kind of business outcomes that they finish to do, which basically connect planning and more strategic support. Yeah. No. Exactly. And I think, you know, at at the end of the day, if your data isn't clean, you know, AI AI isn't going to to to fix your data or or get your data unsiloed. You really still need to be creating that data lake, bringing all of your data into to one place. And I know really great for you to call. I know you've got, you know, you've you've got Workday financials, and and you're also doing workforce planning as well, which is really cool to see. But I think also important to note, we do have, we have, you know, over 7,000 customers around the globe. You know, about half of those and even it might even be slightly more than half or actually not necessarily using any other Workday products besides Workday Adaptive Planning. So they're connecting with other ERPs. So it is important to kind of also meet wherever you are with your current tech staff being able to actually integrate with with whatever data sources you're using, is really important as well. So, I know we're we've had a lot of questions. We flagged some to answer. So now I'd I'd like to invite, skip ahead, invite Ketan back. We're gonna have we're gonna ask a couple questions live. I'm trying to kinda summarize a few of them up. So would love to welcome Ketan back up here and give him a moment to come on stage. But in the meantime, I do have one that I think we can answer for I think we can we can direct towards you, Demetrius. So, yeah, curious. Someone someone asked the question, how many how many people are on your on your your team, that have been using Planning Agent so far? Yes. So my basically, my team is three people. We are basically all you all adaptive admins and have a kinda we're kinda when we do we do the modeling, we do everything. So, basically, those those three people, was were kinda testing the the planning agent and provide the feedback back to the to the states. And then, basically, we also try to kinda, you know, bring aboard some of the pending people. So I would say another two or three were in were kinda involved in the whole testing of the of the planning agent. Great. Perfect. And, Katana, I got one one for you. So you you showed, you showed a little bit around variance analysis and data exploration, but would love to hear, a a little bit about the vision for the next six to twelve months. Are there any specific other skills that are being built that are on the planning agent road map? Yeah. Love to hear about that. Yeah. Absolutely. Thank you. I think the the skill that is probably gonna become available in in the near future is around scenario modeling, which will enable the agent to actually come and create change and compare a number of what if scenarios. So I know there was a question in the q and a and in the chat around this. But what what it will what the user will be able to do is kinda like prompt the agent in, again, natural English, you know, the idea of saying, you know, create a scenario where revenue is 10% lower than last year. Right? Being able to put something like that together and the and the agent will go and actually create that scenario, and then provide a set of reports that will then compare that against other scenarios. Or it could be, you know, on the workforce side of things, you know, what about go and model out a hiring freeze in q three for me, being able to do that. But it's important to note that what we wanna do is within the conversational UI is to almost tell the agent a business story. Right? So, you know, it could be I wanna model a hiring freeze in q three, but also tell me the impact of a change in the national insurance rate, the employee years national insurance contribution rate to from 13.8% to 15%. What about if if the agent could go and model that out? But it's really about, you know, being able to put a business story together. But, that is one of the skills that the agent is currently getting skilled up on, to be able to create these scenarios. And, yeah, we look forward to bringing that out. And, of course, there will be additional skills in in the future around, you know, being able to get the agents to write a formula, get the agents to actually build the whole model from scratch within Workday Adaptive Planning. But the that that's where we're going with the solution. Yeah. No. That's that's so great. And and love love to hear that. Yeah. Scenario modeling too, I know, is, we're our own head of FP and A at Workday. I I I know that it's top of her list, to the point that she's really had we've been really lucky enough that someone on her team on within our own FP and A team is really working very closely with our product product team to be able to, to get that that skill up and running and robust because we want it for ourselves. So, excited that that one's being prioritized too. Excellent. So, another question. So lots of questions around, what is what is currently what is currently available in production today. I think we probably have a mix of folks and people who might be new to adaptive. We probably even have some customers on here, so very welcome to both of you. So if you are a Workday Adaptive Planning customer, so those skills that we we talked about today at the beginning, talking about contextual help, being able to actually go back and, ask questions about how to build a report, data exploration, and some of our variance analysis skill components. Those are all generally available today, that you can put into production. You probably you you you if you do have questions about that, I would encourage you, let us know. Happy to follow-up. Happy to connect you with, someone in customer success to help navigate how to turn that on. We we also had some questions around, pricing. So, yes, definitely, unique to everybody, but I would say that what we've done, for for using planning agents, the flex credit system, so it's work day wide. So if you do have questions around if you are a customer, you are using, you are planning to use planning agent, you've already had a a large number of credits allocated to you, flex credits allocated to you. And then planning around that, your customer success person can can really help, with that, help help with that. But yes. So we'd we'd definitely follow-up with them there. I think we have time for we're getting close on time, and I need a couple minutes to wrap up. But I have one last question I think we'll answer. If we do have any follow-up questions we didn't get to, I'm happy to follow-up with folks one on one as well. But, yes, I think we had a question around, that, restricted, data. So in terms of so, Ketan, maybe I can ask you this one. So, can you maybe explain, like, what the, if you are using planning agent and maybe you're not, not able to see everything, what what data is restricted how data is restricted so you can't see the wrong things if you are asking planning agent something? Yeah. Sure. So the planning agent respects the security aspect that has been set up in Workday Adaptive Planning itself. So for example, if I was to ask the planning agent to go and analyze the salaries line from a p and l, it would only analyze it down to the level of detail that Ketan is allowed to go and view. For example, I could probably I could maybe see salaries at the highest level, but I would not be able to break it down, down to an individual level or anything that identifies an individual. Right? So it does respect all of the security and, governance set within Workday. Excellent. Oh, fantastic. Well, great. Just wanna, say a big thank you to both of you. Really appreciate it. Appreciate seeing seeing Planning Agent Live, and then also thank you, Demetrius, for sharing your story. It's really great. I'm really excited to see too, the six months down the line, what what it looks like at TX Group when you have it rolled out across the business and production. So thanks to both of you. Excellent. So just just to wrap up for for a few key takeaways today, so really hope that you, have have have felt that this is we're kind of at a really exciting moment. Scott Goldilocks inflection point. AI and f and a tools drives value, but it's also, where we have a big opportunity to to to really to really see the future of work happen within FP and a. I hope you've learned a little bit more. If if you are new to AI and FP and A, I hope you've learned a little bit more around, what trends from a technology perspective are happening, the difference between machine learning, generative AI, and agenda capabilities. And that, you know, also that those capabilities that we've talked about today, are generally available and in production today. So it's it's not, in the future, it's now. So really excited for for you all to, really, learn more and and and and join this kind of new era of limitless decision intelligence that we're entering. So, really wanna say a big thank you to everybody for your time today. Thank you for all the questions. I know we'll have some more to follow-up on, but, really appreciate it. And, again, thank you to Demetrius and Ketan. Thank you very much. Thank you.