Name: Wrike AI Agent Build-Off Competition | Part 2
Uploaded: 2026-03-25T16:46:05.161Z
Duration: 1 h 27 min 45 s
Description: Wrike AI Agent Build-Off Competition | Part 2

Transcript for "Wrike AI Agent Build-Off Competition | Part 2": Hello, everyone, and welcome to a very exciting day here at Wrike, at least in my head. We are at part two, the showcase showdown of the Wrike AI agent build off competition. I am Melissa Kovach. I'm delighted to be your host for this. I'm empowering all of my game show, host vibes, so I hope that, today's session is is gonna be a learning session for you all. We did extend today's session because our contestants have such great videos to show you of their AI agent builds. So I do hope that you can spare thirty extra minutes in your day. If not, don't worry. Of course, our today's session will be recorded, and you can catch up on that that last half when you have time. So let's get some housekeeping out of the way. Questions, definitely hit us up in that q and a tab. We certainly don't have time today to do any live q and a dedicated at the end. But if you keep sending your questions in, we'll be sure to to follow-up with you after. A lot of you are already leveraging that chat. Keep it alive. Keep it keep it happy, and keep it fun. There may be prizes for all of you in the audience for your engagement in the chat. So definitely cheer our contestants on. Tell us what you're what you're learning, something that, maybe surprised you. React with each other and network, and we'll be giving away some fun Wrike swag. Excuse me. As I adjust myself. Wrike swag for every for those that are engaged. Those that do get tagged with the cool giveaways, I'll reach out to you and let you know who's who's gonna score some some cool stuff. Of course, today's session is being recorded, so on demand access will be available by tomorrow. You can, watch the the showdown at any time, and or please do share with your company so you guys you all can learn a little bit more about agent building. If you are like me and you love a good subtitle or closed captioning, go ahead and hover over your screen and turn those on if you'd like. There are also some multilanguage subtitles available should you choose. Resources are located in your docs tab. Of course, those will also be relayed in your follow-up email. We have a great prompt guide available as well as a new public AI agent library, so definitely make sure you save those. I know I have them in my bookmarks for all things AI, and I look at both of those things every single day. Let's see. So like I said at the top of the hour, our lineup today is packed. I already kind of alluded to our live audience giveaways. So definitely, keep that chat buzzing as they say, and we'll see who who is acclaimed winner. So everyone's a winner today. It's a good showcase showdown. Marco and Nancy are gonna give us little bit of an update on the latest and greatest in AI. Of course, we met two weeks ago for those of you who joined us for part one of the of this build off series. And but just within that two weeks, we already we still have stuff coming out that is brand new. So Nancy and Marco are backstage and at the ready to to walk us through the latest and greatest. And then, of course, the showdown the showcase showdown begins. We have our contestants live backstage. They have been working so hard on preparing really incredibly thorough videos to show you all and teach you all a little bit more about, how they're approaching AI agent building, why they need an agent to do x, y, or z, and how they use it to save time and bring more ROI to biggest keywords in everyone's life that always makes us perk up a little bit more, save time and money. So I'm really excited to show you some of those, all of those, showcase showdown videos, and then we will open live judging. That's right. All of you in the audience get the opportunity alongside myself, Marco, and Nancy to friendly friendly, review, as we call it. Judging is kind of a harsh word, but you get the gist. We'll be judging, the the video submissions as into one of three categories. Our winners are competing to either receive the best overall agent, the biggest time saver, or the most creative use case. So three winners, those are the categories. We're really looking for clarity of the agent, technical build of the agent, the challenge it's solving, and how much it has returned or saved in terms of that ROI and the the time saved. So we'll review that judging criteria again when the judging round opens, but that's the gist of those three categories. And then we will also recap with a little bit more of those resources to help you. This build off series is designed to help all of you, just like our contestants, to help help them kind of also, teach you all how to be a better agent builder because that's the that's the age that we're in now. Whether we love it or not, it is our AI era, as they say. And with that, let's see here. So we'll go over that a little bit later, but let me bring our product experts here on the stage. And they will oh, right, before we do that, are we I would love would love to hear. open up a poll for you guys. Let's get us thinking in the mindset of AI. Go ahead into that polls tab and tell us, as we all know, you can't can't open up a feed on any social network or newspaper or news website and not have a new term of AI hit you. The buzzwords are everywhere. So I figured this might be a fun one to kind of start things off with when it comes to polls. I'm not sure why it's not showing up on your screen, but it is in your poll tab there. We won't tell anyone, but, you know, which AI term do you think you know and just pretend to know. Is it things like work slob or LLM or AI hallucination? It's not new slang from the younger generation. It's it is what we are doing right now, and this is new language that we're learning. So I figured that some of this would be fun. Go ahead and go ahead and cast your vote potentially. See here. Open that poll there. Oh, good. You guys are all voting. Let me share that. That's the button I was looking for. We won't tell you when we promise. I don't know what work slop is. Nancy had to tell me yesterday what work slop was. And a few weeks ago, Marco had to tell me what an AI hallucination is. So, you know, we're all here. We're all friends. So let us know what what you guys are thinking. It looks like cognitive sync learning. Yeah. That one, I did had to did have to look that one up as well to add there. So, Marco, Nancy, any any buzzwords that kinda make you kinda gale? Commission sync layering is where I go. Uh-huh. Oh, that? Yeah. one's good. Uh-huh. Yeah. Sure. Yeah. What you said. What you said. Love it. Cool. It looks like, yeah, cognitive sync learning. Look. Cognition. See, I can't even say it. Maybe we might have a whole session on that next. I'll go ahead and stop sharing that poll. And then before I I get out of the way here, we do have one more poll for now that I'd like to share. And that is, which AI application would you most like to integrate with AI agents? We came up with a few ideas here, Slack, Teams, Outlook, Google. But, you know, if there's others that you have that aren't listed here, tell us in the chat. And that should say Google Workspace, not Google Workshop. So that was a typo on my part. But, hey, just shows you I don't have my AI agent quite dialed in yet. Your AI agent isn't fact checking here. I guess not. No. She's tired. She's a tired mom. But it looks like Teams and Calendar, I mean, obviously, those would be great. Right? Slack's not getting a lot of love, but we all love a good Slack integration. Who doesn't? And Google Workspace, is coming in close to eleven. I'll go ahead and leave leave the poll open while we kinda go through the rest of the show here. I'll take this off the screen. And, Nancy, Marco, tell us what's new in AI. Sure. Yeah. So I think we talked last time about agent builder, and I'm not sure in the audience here who's actually played around with agent builder, but this session is for you because we've got a lot of great contestants, agent builders, and doing a little preview of their videos. I think they've done a really good job of showcasing some real agents in action using agent builder. The idea with agent builder is that it's got, right now, three prebuilt agents for intake, triage, and risk identification, but it's also, you know, it's all about no code. And, Marco and team are really working diligently. I think no one here will argue that the pace of AI is just crazy unlike anyone any that we've seen in technology. But, every week, Marco and team are delivering new enhancements, and there's a few of these, listed here such as location change, which you you'll see, and you'll even see these on our agent builder videos where you can actually just have a folder that drops down and specify the locations. You can now start and, have start and due date changes that your agents are doing. You can name individual actions because, you know, based on all of your early customer feedback, agent builder customer feedback, everyone said they wanted multi action agents, and now it's clear because you can name those actions saying, you know, at mentioning me or or changing this classification, for example, naming those different triggers. There's one submission that where you'll actually see those as well. And then, the parent fields and sib siblings and cross folder lookups and as well as the Wrike API. So I I won't speak too much. I'll actually hand it over to Marco who can show us a live demo of all these new great updates. Very well. Let me share my screen. Can you confirm if you see my screen? We can see it. Okay. Wonderful. So, what I'd like to show you today is, some of the new features that the team have, have been working on, and do that as part of a use case that I think a lot of us are familiar with, and that's, you know, the the importance and urgency matrix. And what we're basically gonna we we have this folder here, which I call the Eisenhower matrix. And inside of that folder, we've got four subfolders, important and urgent, important or not urgent, not important and urgent, not important and not urgent. Basically, they correspond to the matrix that that is familiar, I suppose. So let's go to the actual agents. I have two here. We've got the triaging agent, and we've got the mover agent. Let's first go to the triaging agent. Here, I'm basically saying, read the Eisenhower matrix folder, and I'm pointing to that matric, to that folder. You can see here. I clicked on the location picker, and here you have the Eisenhower matrix folder. This description has quadrant definition, default heuristics, folder IDs. So let me just go back because that's what I had open here. You see? So we're basically pointing the agent to another folder, and it's this folder. And then what we do let me go back to the agent itself. On new item created, update the custom field, agent reasoning quadrant triage log. Agent reasoning because I would like to know why the agent decided to put it in one of those four quadrants. And the triage log basically gives you a history of what happened. And then in the end, post a comment to in in the folder itself so that you have a list of all of the movements that the agents made. So that's the first agent. The second agent is the mover agent where it actually takes the item and moves it. K? So move the work item to the correct quadrant. Once again, you have the four the four folders that I found here, so I can put in zero one important. See? And then I selected that. The action is move the work item to another location. I could have chosen add work item to another location, but then it would be in two places at the same time. So for the purposes of this scenario, I decided not to. And then here, the comment moved the item to the quadrant. So let's go back here. As you can see, all of the folders are currently empty. I created a little request form, and the Eisenhower matrix request form is right here. I'm gonna paste it from from some of the sorry, from the things that I prepared. So let's say we've got this one here, fixed onboarding hallucination, and then here, I have a description. The due date should be yesterday, but I'm just gonna do it for today. It's high. And, we've got the VP of operations and the account manager that are waiting for this, and it's a bug fix. Right? So I realized that I skimmed over the content, but here, the client's onboarding agent is hallucinating field values, setting categories that don't exist, BP escalated last night, need to renew the agent. So we're gonna submit this, and let's open the request itself here just to show you that it's in here. And so now we have to wait because these are several items that are working at the same time. But my hope is that it will go into important and urgent, but let's see what is going to happen now. This was a common thing that we'll see in in most of the submissions as well when we get the contest is that there's just a little you go. little lag. Yep. Yeah. Great. So you can see it moved it right, and it gave it gives all of the it gives all of it filled in all of the important custom fields. And if everything is correct, it also told us that it did so in the in the parent folder. Right? I'll do one more just to show you that it doesn't only show to import move to important and urgent. So we go here, and let me take this one. So here, route the dashboard feature request to the analytics team and then the description. The due date is for neck yeah. For next June. The importance is low. Are you seeing that I'm I'm playing the system a little bit? Who's waiting on this? I'm gonna say no one. And the type of work is admin. Now I don't know what other signals I need to give to the agent to tell it that it's not important and not urgent. But as we probably all know, agents are nondeterministic. So Here we go again. Waiting. It's taking its sweet time. But at a certain point, hopefully, it will decide oh, look at this. Not important, not urgent. Route dashboard feature request. See? There you go. So with that, I hope that I've given you some inspiration on how to use those those new features. Please be sure that the team are already working hard on the next features, and I look forward to showing them to you. Thanks, Marco. Thanks, Nancy. I really think that, like you guys have been saying, the product team is always hard at work, and there's always something new that we're coming out with when it comes to AI. So appreciate the time that you took to show us some of that. And it's a precursor to some of the videos we're gonna be showing. Alright. So, guys, it's that time. It is officially the Showcase Showdown. I wish I had Johnny Carson's voice from Price is Right, but let's just pretend I am him. And with that yes. Exactly, Sherrie. I'll have you come on down here shortly. But first, let's have miss Ali Moses move to the stage here. She is a powerhouse, Wrike champion of all regards. She has worked with us many times when it comes to a lot of, you know, customer led innovation and just a great resource. If you haven't met her or seen her in the community, definitely connect with her. For her or any of our contestants today, you can go ahead and check out their bio by clicking on the contestants tab at the top of your screen. Ali, it's wonderful to have you here. Welcome to the showcase showdown. Thank you. What a wonderful intro. That was so sweet. I am excited to be here and share my video. And, more than that, I'm excited to see everyone else's cause I wanna steal their ideas. So, go ahead, I guess. Alright. Well, here we go, guys. Again, keep the keep the cheer in the chat going for for Ali and for the rest of our contestants, and let's let the showdown begin. Hi. I'm Ali Moses. I'm the process improvement administrator at Varsity Yearbook. We design, produce, and distribute custom yearbooks to schools all across The United States and in Canada. Wrike is really our centralized work platform, in the plant. It gives us great oversight into workloads, workflows, any potential bottlenecks. We even track our customer art proofs through Wrike, so it really just keeps the wheels turning. So we already have one agent called the taskmaster that updates a custom artist field to reflect who the original artist was that worked on the job. So that's this artist name field right here. I talked about this a little bit during the opening of this event. And basically, this field is used to help art management understand who should be the assignee if this task returns to art for either a reproof or rework. Art management was manually reassigning thousands of these returning tasks during a single production season. So it's just a lot of manual work when they should really be spending their time doing other things. I could solve this issue with automations, but it would be a little unwieldy. It's not really scalable. I would have to have an individual automation for each of our artists, and there are about 20 of them. So just kind of messy, not my favorite. I would love to be able to automatically route work back to the correct artist just using a single agent while also adjusting the timelines based on why it was returned. So, for instance, a reproof has a certain timeline and a rework has a different timeline. So that was kind of my goal in how I went about building this. So in the 2025 publication year, over 3,000 tasks were routed back through the art department, either for a rework or a reproof. Having that all that manual work automated by an agent would have saved us over a hundred hours of management time so they could really be focusing on doing all the other parts of their job that are really more important in this kind of little manual task that doesn't necessarily need their full attention. So this would improve the speed, consistency, and workflow reliability basically for us. I'm hoping it'll make a big impact. Okay. So now we can get into the build. I named this one Taskmaster two, the reassigning. The general instructions for it is you're a task assigner who assigns tasks based on information in the artist custom field. So that's just that same custom field that we took a look at earlier. It's got a list of all of our artists' names in it. My trigger is whenever a task goes into either reproof needed or rework status, I want this agent to wake up and take notice. And we're gonna say any sub item of where the agent was added because I'm going to add it to our work folders. So the first step that I would like it to take is the major one, which is to reassign to the correct artist. So first of all, I told it what the contents of the artist custom field contains. So here I've got our list of artists, and I've also got an other category just so nothing falls through the cracks. And here, I tell it that I want it to read the artist custom field and select an assignee that matches the artist name in the custom field. Here, I've also gone ahead and listed the possible assignees. Down at the bottom, again, just to make sure that nothing falls through the cracks, I told it to ignore any extra text in the assignee field. Our assignees sometimes like to put out of office in their name just to kind of flag their coworkers that they may not get a response super soon. So I don't want to pay any attention to that. I also want to ignore any differences in punctuation or capitalization. And if the artist field is listed as other, I don't want it to do anything. So this action is a change assignee action, and I want it to happen in the work item where the trigger happened. So in other words, in the task that changed to one of these two statuses. My next action is another one just to keep things from falling through the cracks. In that custom field, we have this other option. So if the artist field is listed as other or as blank or null, just in case, I want to post a comment. And basically, this comment is I need it to let someone know, hey, something's going on with this job. It may need an actual human to take a look at it. So the comment that it's going to post is this task requires manual assignment. The action is to post a comment. I went ahead and directed it at myself for this example. And again, I want that to happen in the work item where the trigger happened. These next two are date change options. Now, on it, whether or not it is a rework or a reproof for either one of those, I want to change the start date to today. That's kind of how we monitor how long jobs have been in the art department. And we make sure that things are getting handled in a timely fashion. So I'm gonna tell it to change the start date to today. My type of action is change the start date. And again, this is in the work item where the trigger happened. Now here we have a little bit of a difference. I wanna change the due date, but I want it to take different actions based on which status the task comes into the department in. So if it's in reproof needed, then I need the due date to be three business days from today. However, if it is a rework, then we only want it to be one business day from today because those are pretty hot. So the action is change due date. And again, that's gonna be in the work item where the trigger happened. So let's go ahead and get out of here and go see if we can make it work. Alright. So right now, we are in the scheduling department folder. So this is where jobs would kind of come back to them, and they would take a look at it and decide where it needs to go. So let's say Sunnydale High School maybe has a rework schedule. And let's say that the rest of these maybe have a reproof that they need. Okay. So in automation, went ahead and grabbed those and moved them over here to the art folder. We're just gonna wait a second and watch our agent start to wake up and take notice. So here we see okay. Cool. This got reassigned to Pablo Picasso based on this artist name custom field. It is a rework. So the start and due date are gonna be the same day. It's still thinking about some of the other ones. Hopefully, on this Rydell High School where the artist name is listed as other, I should get a notification in my inbox here shortly. Oh, yep. There. It just fired. So let's go over there and take a look at that. Taskmaster two. This task requires manual assignment. Perfect. That's exactly what I wanted it to do. So then I can go in there and take a look at why it needs that. I have added this agent to both the scheduling and the art department folder. That way, if the automation kind of works a little faster than the agent or vice versa, it it would take notice in either folder since we there our automation and our agents are kinda working together. So I added it to both folders. And in order to do that, of course, we would just click on the folder, go to info, and then we can come down here and use this cute little add agent button. You can see here this is where I have added it. So that is how it works. I would say the key benefit to this is that it is going to automatically add the correct assignee and adjust the timelines for our art department. So hopefully that is going to just make everything a bit more reliable and accelerate production work. Probably the best tip that I could say for building something similar is just to make sure that your custom field that all of this hinges on is nice and clean. Bad data can definitely break this automation. In closing, I would say automate the busy work so your team can do the fun parts of their job. Ali, amazing. I love the the the taskmaster reassigning name. Fantastic. You're such a nerd, I love it. I think a lot of us are jealous of your, click path demo skills, making it look like it's nothing. I've only done tried to do that, like, a 100 times, and they've never been successful. So great great work, Ali. Thanks so much. I do like to keep myself entertained with the little things, so, hopefully, it entertains some other people too. Yeah. Love it. It was great. I just a quick comment. I thought it was really great. I loved how you started on the back end, and you showed us all the kind of magic that went into your prompts. And then, and then showed showed it at your your agent actually in action. So, I think you could actually be a school teacher maybe in a if you have a different career path, Ali, because you did such a great job of kind of putting everything in layman's terms. So nice job. Love it. Thanks so much. You talk to a lot of kids in yearbook land over here. Oh, I can imagine the note that conditional date, I bet, is a true I hate the term, but true game changer for you, I I would imagine, with so many pages in every book and so many books you're making every year. So, love it. Well, best of luck, and we'll move move you backstage here and bring on our next showcase showdown contestant, Andres Serratos. I really hope I pronounced that correctly. I'm horrible at names even though we've been working together for months. Andres, welcome. He is representing team at AppFolio, on behalf of him, Jared Brize and Mark Hemis. Andres Serratos is here today to talk to us a little bit about his awesome AI agent, and welcome to the showcase showdown, Andres. Thank you. I'm excited to be here, but also just really excited to see all the builds. Ali, that was awesome. I've looked at my own build for hours and hours and hours, so I'm I'm really happy to see all the other ideas too. It's like in school. Right? When somebody goes ahead of you and you're like, oh, wait. I should have done that. Darn it. Yeah. But, Absolutely. with that being said, Andres, let's go ahead and see what your agent is all about. Hey, everybody. My name is Andres, and I'm a project manager with AppFolio. AppFolio is a real estate performance platform. Me and my team were project managers specifically for the marketing organization. And we use Wrike in a multitude of ways, but it's all with the goal of helping execute our marketing campaigns. So that means planning projects, seeing them through the execution phase, managing the work of our many production teams, sprint planning, gaining insights through building dashboards and building reports. And I'm really happy to be talking AI agents today. So the process that I wanna walk us through today that I wanna automate is that of closing the sprint at the end of the week and pushing active items to the following sprint. So our team does work in weekly sprints, which means that this is a process that I have to repeat at the end of every week, usually on Friday towards the end of the day, right when I'm ready to start my weekend. I'm actually doing this to to close out the week, and it takes me between fifteen and twenty minutes every Friday. Can take a little longer if maybe there's an oversight or an error, which is possible given that it's such a manual process. So I'm sure you can understand why we wanna automate this and me take that time back. So before I really get into it, I wanna quickly kind of explain our folder structure here. So again, weekly sprints. And I do repeat this process for every one of our production teams. So you see here, this is the automation example. But I do repeat this for copy, for our web development team, for database, and for our creative teams. But they all basically use the same structure. So I wanna bring our attention to the current sprint folder here. This is a crucial folder, and it's important that it's always up to date because this is a source of truth for our production teams. This is how they know what to work on in any given week. And this also feeds into our reports and our dashboards as well. So again, very important. What we wanna do, or what I do every week, is I need to identify all the items, all the tasks that are still active. That way I can add them to the following sprint week. So you see here, it's it's week 12. This is the particular week. So that was last week if you look at the dates. So let's pretend that this was last Friday. I wanna identify which are the active tasks, push them to week thirteen, which is our current week, And then and and before I move forward, you might be wondering why there's a few stragglers here that are tagged in the current sprint folder, but they're not nested in the actual weekly sprint folder. And that's basically just because it's a shortcut. And these are tasks that were kinda dropped on us midweek. So when there's a task that, you know, we have to kinda create midweek and add to sprint midweek, we kinda just add it to current sprint as a shortcut because that's faster than having to, you know, click here and then type out the whole naming convention, auto 26 q one, etcetera, way quicker just to add this tag. So that way, it shows up in our reports and our dashboards, but and it's just it's just under current sprint. And this is the point where I would take these, put them in the actual week 12 folder, and then we move forward with moving all the actives to week 13. And then I would have to archive week 12, put that in past sprints. I would need to take the week 13, again, stragglers here, put them in week 13, then take week 13, put it in current, and then go to future, take week 14, and put it in next. So that would end that would complete the process just for this team, and then I would have to repeat that for the rest of the teams. So hopefully, I've demonstrated just kind of how manual it is and why this is such a good candidate for us to automate using AgenTeq AI. Okay, so now to talk a bit about the build. This is actually gonna be a group of AI agents that that are working together to complete this process. It's not just one. But the very first part of the process is, like I had previous previously explained, just adding these stragglers that are in the current sprint folder, adding them to the actual week 12 folder because they are part of the sprint. We just need to add them. And so that's what this agent is meant to do. Roll is the sprint task mitigator. It's triggered by a custom field that we created. It's just called AI trigger, and it's a multiselect custom field. But, basically, this setup is whenever that custom field changes, go ahead and fire it off. I set this to the work item where the agent was added. So I've already added the agent to current sprint. So once I once I change the AI trigger field here, then it'll get going. And as far as the actual prompt, here it is. Basically, all I'm asking you to do is the agent should look at everything that's in current sprint, every task, and just add the week 12. This this tag here. Just add it to every single item. And if it has the current sprint tag, just go ahead and remove it. And that way, everything here will just be moved to here in in simple terms. And so the action is move work item to another location, of course. And it works in all sub items of the work item where the agent was added. So, basically, just everything underneath the current sprint is being looked at. And then for comment metrics, I decided that when I do this, since I'm getting everything nice and neat in one folder for the current sprint, I wanted to pull some metrics for me, especially because, again, I'm closing sprint. So I'm asking you to give me some data just on what tasks are in which statuses, what's overdue for high priority or launch related tasks, which ones are still in progress, because those are things that as a project manager, I need those insights. So it's perfect to nice to get that just kinda nicely wrapped on a Friday as we're closing sprint. And so this is in the form of a comment. It tags me. It adds it to the actual work item. So, again, the current sprint folder. And you'll see that in action in just a bit. But I wanna move on to part two or part b, I should say. And part b is moving open tasks to next sprint. So here's where the actual, like, action happens of moving the items. So basically, we're asking the agent to find everything that is active, every task under current sprint that's active and just move it to next sprint. And so here are the general instructions. The trigger is exactly the same. So these these agents work in tandem. And then here is the actual prompt. If you wanna take a look at it, I've already kind of explained what the what the result is that we want there. And then as far as the comments or the other action of access to comment is it is only if there's an error, basically. If for whatever reason it cannot access that auto next sprint folder, just address a comment to me so that I know. So that's part b. Now, c, add tasks to add tasks to next week. So by this point, we'll have all these open tasks sitting under next sprint, but they won't be in the actual week. So again, it's the exact same logic as that very first part, part a, is taking everything here that's in next sprint, but not in here in week 13 and adding it to week 13. So same exact logic as as part a. Same trigger, just in a different folder. And there's the, the prompt for you. Move work right into the location in the action. So that's, part c. Now part d is to actually archive the current sprint. So by this point, every active item has been moved to week thirteen. That's where we're at in the process. And we need to take week 12, and we need to push it to past sprints. So this is probably one of the simplest prompts that I wrote for this process. The trigger here is a little bit different. This is just the checkbox. The naming here doesn't matter. It's just the the checkbox. And, basically, this just says, identify the week 12 folder within current sprint, which is that's how it's set up. And then take that week 12 folder and push it to pass sprints. So one important thing I wanna note here is that I hard coded the folder into the prompt week 12. But, obviously, as we go week to week, it's not always gonna be week 12. That's something that, you know, eventually, just once the functionality gets improved to where the agent can look up, what's the one and only folder that exists in current sprint? Is it is it week 12? Is it week 13? Is it 14? Something else. It'll know that, and then we don't have to hard code it anymore. So that's gonna be the future state. But for maximizing the functionality for today, I just hard coded it, and then you'll be able to see it in the demo. So, again, that'll move that to past sprints. Now, for the next part, part E is moving that next sprint to current sprint. So general instructions, the sprint folder relocator. So by this point, everything is in week 13. On Monday, it'll be week 13. Again, in this scenario, it's Friday. We're moving on to week 13. This will move it to current sprint. Here is the prompt for that. Again, keeping it super simple. Look for this week 13 folder within this one, and then go ahead and move that to current sprint. So that is part part e. And then the very last part is just manual. I've not been able to to automate this part. But just for the sake of showing you the process, I would just go to future sprints, whole week 14, and I'd push that to next. So that's kind of the the agents and how we set them up. So you guys, hopefully, you've had a a chance to take a look at the prompts and kinda see how how we got that to work. Okay. Now let's see this collection of AI agents in action. So I'll go ahead and trigger the current sprint folder here, AI trigger, as I previously explained. So this will trigger two different automations. So it might take a minute to start running, but we should see these tasks get pushed into week 12. At the same time, like you're seeing now, we should see the active tasks from current sprint get added to next sprint. And so we'll just give it a minute to do its magic. Great. And it looks like it added week 12 here. We did not want that. That's okay. But as you can see here, everything that was active in week twelve is now is now in next sprint. So awesome. Now, the next part is we want to take everything from next sprint. Sorry. We want yeah. We want to take everything from next sprint, and we want to move it to week thirteen. So we'll go to next sprint. It's the exact same trigger. Just change the change the value, and it'll get going. And that's really cool to see. So week 13, it's got all the tasks in there. One thing I want to go back and show you, actually, I almost forgot, is it did provide us metrics on the end of sprint. So you can see the different statuses here. This many items in complete, this many items in review, etcetera. Overdue tasks and ownership. You'll you'll get the names here of the task, the assignee, how many days it's overdue, published dates that are missed, critical launch and high importance items that are missed. So it's a really cool summary that you can look to as you close your sprint. Alright. Now the next part is to archive the current sprint. So we want week 12, and then we want that push to past sprints. So I have the trigger set up for request, and that's all you do. It might take a few seconds. I'll go ahead and I'll do the same thing for next sprint. I wanna move that to current. So that's already checked. I'll just uncheck it. And so current sprint's gone. That should be in past sprints. There it is. And now, just waiting for current sprint to move to oh, no, it did. Next sprint moved to current. So that's perfect. And then the last part of this process is we go to future sprints. And then we pull week 14 into next sprint. And that's it. It worked. So, yeah, I would complete this process for the rest of the teams. But like I explained earlier, this is like fifteen to twenty minute process, and it just just this one team took me, like, a minute or so. I mean, I walked you all through the process. I don't I wouldn't normally do that on a Friday, so it would be even quicker. But, yeah, that's basically how it works. Andres, fifteen to twenty minutes down to basically sixty seconds. Okay. Let's, like, talk in some r ROI of time. Really, really impressive work. Really love that your your sprint is automated in such a way that, you know, who doesn't love to spend their Friday afternoons working on sprint reports. Right? I love that automation factor that you brought in with the reporting and the dashboards and everything. I think that that really is full circle. But then you took it even further, and you made an made an archival step. Like, brilliant. Just truly brilliant. I think a lot of people in the audience enjoyed it as well. Yeah. Thank you. And one thing I wanna point out too is I I wouldn't even say that these agents are in their final state. You know, there's always opportunities to improve. Maybe I can consolidate some of them into one. Obviously, that metrics thing was awesome, and maybe, you know, you can rewrite and add different metrics and different steps that you wanna capture. So. yeah. And I think that's a really fantastic point, is that an AI agent is always living and breathing in a lot of ways. Just like us, we're always learning and growing and changing. So that's definitely the mindset that that we we approach with agents as well in terms of it's always getting a little bit better, a little bit more powerful, a little less hands on, even even more so. So thanks so much, and best of luck to you. Thank you. Alright. Andres, come on down. See, gotta get I gotta get that that voice down. But come on down, Nathan. I'd love to bring you on this stage. Mister Nathan Jones here from the College of American Pathologists (CAP). Quite quite the name, and I think your title is even more impressive, but I don't have that off the top of my head. But welcome, Nathan. We're super excited to have you and to see what you've got building here for AI agents. Great. No. Excited to be here. I'll say my title real quick because it's a relatively new title, It's still practicing title. as well. Yeah. Senior director, emerging digital experiences, and innovation. That's it. I've long lived there, a lot of AI. So, yeah, really excited about this contest, and looking forward to hearing some reactions on the video. Emerging digital experiences. Talk about a today's type of title. I love. it. So let's see what some of that emerging experience is, and let's go ahead and play what your video, is building. Hey everyone. My name is Nathan Jones. I'm senior director of the Emerging Digital Experiences and Innovation team here at the College of American Pathologists. We are a medical association and our focus is on laboratory accreditation and providing laboratory quality products. Personally, I've been using Wrike for over ten years now, it's one of my core tools in my arsenal, But specifically, what I'm gonna show today is we're actually in the process of building a new enterprise innovation program at the organization. So all things related to innovation got a lot of ideas coming in, so we really wanted to create an idea management repository. So decided to build that in Wrike. And while we're at it, let's try really using AI as a core piece of the work. So the idea was pretty simple. We set out to really build an idea intake agent that would ultimately review any kind of form submission data, automatically populate custom fields, and ultimately reduce the burden on the submitter. Right? No one likes to fill out really long forms. So without AI, the the the flow would really be like this. Right? A Wrike form, pretty straightforward. Then maybe a user would review that submission, maybe schedule a follow-up meeting to give some more information, then take that information back, maybe score that idea, you know, is it high impact or, takes a lot of effort, is it feasible, things like that. Then maybe we have a team meeting to then review and discuss that idea. Ultimately, the idea goes into the idea repository, And then depending on that criteria, right, we would then decide to prototype it, pilot it, release it, and ultimately build it. Pretty straightforward process, but I could already tell immediately we would get stuck in this review through idea scoring sort of step. Right? It would just take a lot of time. So for every single idea that comes in, we have to look at it, think about it, talk about it, meet about it, meet again, meet again. Ultimately, the goal is to actually make something, not just to to talk about it. Right? So the next step is to think, well, can we we make this intake agent to ultimately do a good portion of the project? Alright. So let's take a quick look at how this was set up. So here we have our idea intake review agent. We gave it a simple name, some pretty straightforward instructions. Right? You're an operations assistant to help our team complete details for new ideas, and the objective is to review all task details including the description, custom field, and any file attachments. For the trigger, I set it up to actually fire every time a new item is created, and ultimately, this is coming in through a form submission. You could also trigger it off of a workflow, and I've experimented with that some as well. And then it's really going to, work anytime a new item is created in that specific folder where I added this agent. So looking at some of the actions here, the first one, update the title. So based on the submission itself, it'll update the item name with the summarized title of no more than five words, then create a short summary. Right? So look at whatever was submitted, create a short summary that's no more than one sentence long, and that that, you know, sort of a quick scannable description, if you will. Next, it actually created a, potential opportunity. This is where it starts to get a a bit fun. So same idea, look at the task details, and then update the opportunity custom field that I created with a brief summary of the problem, the opportunity, and any potential outcomes. I am pushing it a bit more to think deeply, propose a big picture opportunity, to sort of go beyond that surface level submission, maybe even try to get a little creative, have some fun, but still no more than one sentence. Still keep it brief. And then again, that's populating a custom field. Then we get into the actual scoring, right? And this is the part that can take a lot of time normally, or it can be very subjective. So you'll see here that I, really sort of qualified when we think of high effort. What does that mean? So I I wrote down the idea feels like it'll cost a lot of money and require a large team to work on it. Or if it's low effort, you know, or call it low if it's, something that maybe will take just a little bit of money or it could be completed by a single person. What's interesting here is this gives us a way as humans, right, to to calibrate that criteria. If if we think that high or low means something else, we add it here, and now the agent will just use that. Same idea with estimated impact. I wrote down tag it as high if it's going to impact an entire group of customers, right, like all of our customers and members or all of our, employees or medium if it's only a portion of a group or maybe low if it's really gonna impact only one or two individuals. So up next, we'll do a quick demo of how it actually works. So what I have here is a straightforward external, externally facing form. And what I'm gonna do here is actually grab a quick idea. I generated a a bunch of ideas for demo purposes. So here's an idea to create a a benchmarking dashboard for labs. Right? Develop a digital dashboard that allows laboratories to compare themselves against peer benchmarks. I'm gonna hit submit. As soon as I do that, we'll see the idea popped in right here. Right? And so you see this dedicated Wrike Space. I'm in the board view. You see what I submitted here. You see empty custom fields on the side. You see there's actually no title here. What's happening right now is the AI is actually working behind the scenes, really in a couple seconds, all of this should update. So we'll give it a few seconds here, and there it is. So you see it automatically renamed it. It automatically gave it a short summary. Let's take a look at this idea. It said to create a benchmark benchmarking dashboard for labs. It offers a strategic opportunity to differentiate our offerings. Right? So it's really going again, sort of reframing that idea using AI to really call out the opportunity. It's saying it's maybe a high impact and even high effort. Right? So now, of course, as humans, I can go in, review it, I can take a look and change it, or I can just go with this. No surprise, we're able then to use this to surface that same data then in dashboards so that we can quickly see, well, what's maybe an easy win, what's high impact, low effort. Maybe we do our review first there, or, maybe we wanna slice and dice the data any other way. So, ultimately, it starts to save us a a lot of time. I did try to put some math to this. You know, again, if we think of the original process for a right form, maybe it takes the user fifteen minutes to fill it out. Another person fifteen minutes to review. Maybe we have a quick thirty minute follow-up meeting. Maybe then that person looking at it goes off and does some of that initial scoring, has to do some additional research to figure out how to score it. Ultimately, we have a team meeting for an hour, hour and a half, maybe even more, all of that before it even gets into the idea repository. So really, we're looking at without AI, about three and a half hours best case per idea. Now we all know even that three point five hours might take, you know, two, three, four weeks, then you meet again, you forget what you did last time, and it just sort of goes on and on, right, before you actually build anything. So with intake agent, I really dropped all these numbers down to zero. I was able to make it a much shorter form. You saw it was only one field, not a long form that someone need to spend a lot of time with. And then ultimately, that team review is sort of sped up too because a lot of the data is already reviewed for us. So I like to say that essentially, the idea now only takes about fifteen minutes of of human time, right, per idea. So in other words, about three point two five hours saved per idea. So that, you know, adds up pretty quick. The last thing I'll just add is that now I'm going a step further to look at multi agent workflows. So I initially had an intake agent. Now I also have broken that up into a separate scoring agent. I have a routing agent that moves into different folders. And then an agent, depending on how it's scored and depending on how it's, routed, it'll actually start to create a brief of that idea. So that way when a human comes in, they already see a full project brief essentially ready ready to go. So I'm having a lot of fun. In closing, you know, couple quick notes. I always like to say, don't overthink your prompts. You know, test it. Update it as you need. You don't need to spend a lot of time getting too specific with language. Just sort of give it a go. Put it in there. See what's working. Tweak it. And I do like to say take notes as you go. I get excited as I make the ideas more and more complex and stitch together multiple agents. If I don't jot down to notes, I'll find that after an hour, I'm sort of lost down a path. Right? So that's my Wrike intake agent. I hope you had fun at least sort of watching some of it. I know I certainly had fun building it. Again, it's pretty straightforward. Pretty easy to build these agents, so I do hope you dig in and and have fun as well. Thanks, everyone. Nathan. Nathan. Nathan. Who has been how many of us have been in a brainstorm meeting where we're full of awesome ideas, and then you go and process them and three and a half hours per idea, that's insane. And he brought that down to what? My notes say fifteen minutes. Incredible. That's just you guys keep upping the ante. You're gonna make voting really hard. Okay? You're all overachievers. But, Nathan, I love that. I love I use request forms day in and day out. I mean, I'm in marketing. We basically run on request forms. So having it simple, reducing that friction, one field, and then submit. Done. I mean, come on. Genius. Genius. Well, thanks. I'll I'll just add, you know, as someone that's built very complicated forms, it's, sort of cathartic. Excuse me. Right? Sort of cathartic to even just be able to introduce a single field form, so I'm? enjoying it as well. easy to overcomplicate a form because you want to cover all your bases. Right? So you want to do this and do that and try to answer everyone's questions before they ask you. And I think it shows a a real level of sophistication when you can take something that you can really get advanced with in a very simple way. That's the Apple way. Right? Sophisticated simplicity. And I think that that's what I take away from from your agent. So so great work, and best of luck. Great. Thank you. K. Sherrie, come on down to the stage. Sherrie Besecker of Syneos Health. She is also I mean, guys, our contestants are just amazing. Hand over fist. Sherrie, awesome to have you on stage, and welcome to the showcase showdown. Oh, Anne, you're on mute. We were all one of us were gonna was gonna do it at one point. We've all been there, Sherrie. I've unmuted one to unmute the other. There hi, you go. every Hopefully, that works. Hey. And Sherrie, again again, our contestants are overachievers. So Sherrie submitted three agents, and we made a tough call, to only show one of those three agents. If you wanna see the other two agents that Sherrie built, we'll we'll be adding those to your follow-up email as long as well with everyone's agent so you can definitely take more notes. Sherrie, we are going to showcase your agent on task of the day, or is it task of the week? Task of the week. Task was originally the week. today. Task of the day for testing, task of the week for the real thing the real deal. Alright. Alright. We always love a good good task of or of list. So let's see what Sherry has in store for us. Okay. Now we're gonna look at another agent. This particular agent is called my task day of the week agent. What is actually happening and what's unique about our industry in health care and pharmaceutical advertising is that as we take our client through their legal review steps, they have very client unique prescriptive days of the week that their review meetings happen. So you might find that often happens like you thought you were scheduling it for the right day of the week, and in fact, you missed it by a day, and so your timeline's gonna be crunched, and you're gonna have some unforeseen risks that come up very quickly for timing constraints. So we've had you know, you can always go in and click and go, okay. This is happening on a Tuesday or this is happening on whatever day of the week. And so what I wanted to do was to be able to create an agent that flags days of the week for our medical, regulatory, and legal submission steps. So this is actually a kind of an interesting little agent. Very, very simple. It really literally just reads anytime that a task date changes because our task we know one thing about project management. All of our timelines will be changing all the time. So let's say our kickoff got delayed, and we actually didn't start it on the sixth. Let's say we're a week behind, so let's just say we're kicking this one off. We'll just reschedule this date for the thirteenth. Right? So everything else beyond that should be shifting. And what this will do is it's gonna start looking for those medical regulatory legal steps and applying the days of the week so that we can literally see what day of the week these particular days fall on. So I can see these are on a Tuesday. These are on a Tuesday. And just to check it, is that a Tuesday? Yep. And is this a Tuesday? Yep. So let's say my date needs to be on a Wednesday instead of a Tuesday. So I've got to make an adjustment at some point, so I'm gonna push this out just a little bit further out. So I know these are my submission dates, and then these are my review dates. So I need to push this out to add another day, so I'm gonna go ahead and give my creative team one more day. Since we have time to accommodate, I'm gonna make that a four day window instead of a three day window. So that will automatically then change and update my dates. It just takes a couple of minutes here, just a minute or so, to go through and reevaluate when they landed. And they'll go in Wednesday, Wednesday, and the other one should change to Tuesday here in just a second. In theory. So as we wait for it, there it is, Tuesday. So we can kind of see the other thing that's happening is I'm getting notifications because what I want to have happen is and this was a happy happy accident in true Bob Ross fashion, is I want the team to be notified when those dates have been rescheduled so that they can quickly check. And the happy accident was originally thinking it would go inside of the tasks that it lives in, but what I actually love is that it actually posts it here. So anytime something is actually changing, I can see those tasks days of the week right here when they were rescheduled. So I can make sure I have a little bit of a change that it notifies the author of what's happening, and we're pretty good here at this particular time. And so that's what my agent does, and let me show you how I wrote that agent. In the day of the week scheduler here, your purpose is to monitor for these tasks due date and find the actual date the due date occurs on and excluding projects and folders. And so it's basically anytime the due date changes, I need to, update the custom, field to match the real calendar day of the week, and do not change any other tasks day of the week fields, just this one. And then, of course, posting a comment, monitoring again to the author if it changes and if something adjusts. They can even customize it to say if it's not a Wednesday or if it's not a specific day of the week. So those were just my my test scenarios, and that is how my Oh, Sherry. Got cut off a little bit there at the end. Sorry. Sorry. Was so quiet. As you all know, I am not quiet. So that recording. Gotta love technology. That must have been a upload error, so I apologize. But we didn't miss the meat of it. I think all of us understood a lot of what you're trying to do there. Time line's changing. The notification, though, I think that's really the next level there. How many times have we found out that a time line has changed after the fact, and missed a deadline, because it got changed and you were never really known? So love that. Love the task of the week. Again, Sherrie had two other submissions we'll be, putting in the in the follow-up email, but this is the submission that is contention for the showdown. So, best of luck, my Sherrie. I wanna say too real quickly, like, this is the one agent that when I built it, I was solving a problem that I couldn't solve about eight years ago that I tried to solve and make really simple. So go into your archives and think about those problems that you tried to solve back then that since you kind of figured out other ways of working because there's some jewels in there that you could build some great agents in there. So Hey. That's a great tip. Definitely write that one down, because kind of kind of forget some of that stuff that you could now hit the easy button, more or less. So thanks, Sherrie, and, best of luck. We'll see you at live voting. Alright. Last but certainly not least, we have Szilvia. Come on down, who is joining us from Budapest, Hungary. She is from Avida Avida, and we're super excited to have you with us. Your agent video, just gonna spoiler alert, is pretty big deal. So excited to to showcase it. Thank you. I'm also excited to be here with all of you. I'm happy to see the other content contestant AIs. I'm already stealing ideas. They are great. So, well, no no speech for me. I will let my video tell the story of my IT PMO compliance page. Ninja, let's call it how I call it. Love it. Best of luck. Let's see what you got. Here we start with a new demand request entering the system. The request is first reviewed by the head of department, and then by the country IT manager, who provide the required information, including a clear project description, key dates and all mandatory fields. When the status changes to 'Waiting PMO approval', the PMO compliance agent is triggered automatically. The agent checks whether the request is complete. If any information is missing, unclear or inconsistent, it is sent back to the head of department with a structured comment explaining what needs to be completed. If everything is in place, the request moves forward without a delay. Once a new Dimitriq that is converted into an IT project, project managers prepare it for project board approval. Each project must include at least one benefit that meets the required standards. When a benefit is created and its status is set to validated, multiple AI agents are triggered automatically. The Benefit Classification Ninja analyses the description and classifies the benefit as a financial or non financial and assigning a standardized category. The financial scoring agent calculates the total value based on this categorization. Finally, once both value and impact are available, the benefit scoring agent combines them to calculate an overall score. Once individual benefit impacts are defined, changing the project status to PM Board review automatically triggers the benefit impact roll up. It aggregates all benefit impacts and calculates a consolidated project level score, automatically but consistently. Switching the status to PMB review triggers the PMB compliance agent to validate the IT project. Complete and accurate projects are marked complete with a clear comment. The agent evaluates data integrity across all mandatory fields, and verifies that the project structure, scope, clarity and benefits are present and compliant. If any required data is missing or inconsistent, the request moves back to info requested, with guidance on what needs to be corrected. The short description is also automatically updated to have the project management board review. Okay. Szilvia, let's be real. Okay? First off, kudos to your videography skills. Second, incredible way to spice up the world of compliance. Really awesome. Awesome work, especially since it's such a complicated and nuanced thing. You've really show showcased how to how to simplify that. That's amazing. Incredible work. Thank you. Thank you. I'm trying to make our life easier, you know, with this intake love. to learn as you've rolled out your agent. I'd love to hear later how much time you're saving from all of that and, you know, a little bit more of the impact of that because I feel like that that must be huge to scale out and roll out to a company, of in this size. you know yeah. The, you know, the projects are just jumping, you know, between the country IT managers and the PMO sending back Rick before the required information. So, actually, it saves us a lot of time. Yeah. I can imagine. That's amazing. Well, that is the showcase showdown. Szilvia, hang out on stage here. I'm gonna bring unbeknownst to them, I'm gonna bring every of our all the rest of our contestants on this stage because it's that time, guys. It is time for voting to begin. So, Ali, Andres, Sherrie and Nathan, let's let's bring all of our lovely builders here on the stage. Sherrie. I was seriously fangirling backstage about all the other builds. They're so cool. I have so many ideas. Right. You guys really really made it hard here. You're making it hard. As we open up a polling here for our winners, let's talk about the first winner, in terms of the categories. So we're talking about best overall agent. Which agent do you guys think out of all of our wonderful ones, we'll recap. Ali had the taskmaster, the reassigning. Andres had the automation team sprint. Nathan had the intake idea agent. Sherrie had task of the week, and Szilvia, last but not least, she had compliance ninja. So let's look at if we think one of those was the best overall agent. I know, guys, it's really it's really hard. We picked a lot of really fantastic customers here, to be our contestants, and the videos just kept getting better and better. I know it will be hard to vote, but the voting is open for best overall agent as well as I will also open and share our bet biggest time saver. So think about a little bit more around those time savings. Like, Andres agent for the automation team sprint, we're thinking twenty minutes down to one minute. Or Nathan's agent where, he took three and a half hours, essentially, and brought it down to fifteen minutes for ideation and process development of that idea. Think about Ali reassigning, how much time all that artist reassigning takes. And think about kind of I always like to think about time savings as the less you have to click, the more time you get back. And that's a lot of what these agents are doing is taking away a lot of those unnecessary clicks. So? think about oh, Sorry. yep. It looks like people are not seeing the poll. Oh, well, that that would be that would be because I forgot to hit open. Mhmm. Again, this is why I need an agent, and this is why I like having you, Ali. Okay. Poll is open. Let's see. Close this one. And open. Best overall agent is open. There we go. We'll leave that open for a little bit as you vote for best overall agent. The next one, again, is gonna be biggest time saver. I'll leave the poll for the best overall agent open. We can only have one poll open at a time. And then we'll move on to the most creative use case. Voting will also be opened in your follow-up email. So if you don't for people who had to leave early, or if you wanna sit back and rewatch some of these amazing builds, I know I will, and vote later, you certainly can do so in your follow-up email that you will be receiving from me, over the next few days. Go ahead and vote for that best overall agent, in the poll now. And then our next our final category is going to be most creative use case. Which agent do you all think was the most original and inventive application of AI? Depend no matter the use case, no matter the industries, our contestants are from across the board, from yearbooks to health care, and everywhere in between compliance. So, it's really about how they got creative with it. What is something that made you think, oh gosh. That's a magic moment. That's a creative use case. So think about some of these some of those criteria as you as you vote. Again, since we can only do one at a time here, I'm gonna close. I won't share yet who has been voting for who, and I won't share if you vote for yourself. Kind of a tip there. I will go ahead and close that. Open up biggest time saver. We do have, speaking of time, four minutes left, and I've already held you thirty minutes over. It's just such a fun day. Showcase showdown. We might have to do this again. I feel like it it has been a good time. Nancy, Marco, join us on stage if you'd like to. Tell us a little bit more about what you saw when it comes to some of these agent builds. Like, we knew they were gonna be good, but this good? No. Not this good. I'm super impressed. I mean, kudos to all of our contestants. Just really great great use cases. You're so generous in in kind of explaining your process as and this is gonna be really hard for me as. as it will be for our audience, I think, to to decide on. the let me just echo that. I find it completely impossible to rank the submissions. They are very, very impressive. Some are impressive because of the workflows they cover, the other because of the ingenuity in the in the actual prompts. And I'm not even gonna speak about, you know, the quality of the videos themselves. I'm I'm just blown away. Kudos, guys. This is fantastic. And that's coming from the product man himself, so I would definitely keep that on your brag wall. And, also, happy belated birthday, Marco. Marco celebrated yesterday. So you're starting off your next year. round trip around the sun with the showcase showdown. What more fun can you have? Also, Exactly. audience, our most creative use case poll is open. Go ahead and cast your vote for which agent our contestants submitted that makes you think, wow. That is really just the most creative. I can see why on competition reality shows, the judges always say this was really hard. I always thought that that was just, you know, lip service, but I get it now, man. I I don't know. I might have to take the rest of the day to rewatch and revoke. Because I thought I had my favorites for each category, and then I rewatched, and now I'm like, oh gosh. It's gonna be diff difficult. And I'll give you one hint. There's lots of close tie across the board so far. So go ahead and cast your vote. And in the interest of time, I'll go ahead and leave the most creative use case open here for a minute more. To all of our customer contestants, thank you. Really, you blew it out of the water. You, as they say, you understood the assignment and then some. So if I had a little noisemaker, could clap for you, but I hope that everyone in the audience felt their energy, felt their passion, felt their creativity. And I hope that you walk away with a few new ideas on AI agent building. Again, you will be able to vote in your follow-up email. I'm gonna go ahead and close that and just wrap things up real quick and then let you guys get on with your day. Of course, if you have questions, I know a lot of you sent in some fantastic questions in the q and a panel. You still have time to submit those. Go ahead and and hit us up with those questions, and we'll follow-up with you after. If you wanna learn more about AI agents, prompting, demo, you know, demo building and stuff like that, please reach out. We're always here. We'll have more sessions like this, hopefully, if you all enjoyed it, if our contestants are not jaded too much by by being on the first ever Wrike Showcase Showdown. Hope to all join you again soon. And go ahead and check out those that docs tab for some resources, including our awesome Spotify playlist, for this build off series. And part three, we will announce all the winners on April 8, so please do join us live for that. Our winners will be receiving, a pretty sweet limited edition exclusive swag box only available for this build off competition and lifetime bragging rights, which honestly, I think is kind of better than the swag box, but the swag box is pretty sweet. So join us live April 18 or April 8, for the final contestant winners. What do they call that? The winner's circle. That's right. Join us for the winner's circle on April 8 where we'll be naming and crowning three winners. Sorry. Two two of you have to I don't wanna say you're losers. Runners up. Okay? We'll have a sportsmanship or sportswomanship type of award because they're just all so good. Go ahead and, let us know what you think of of each of these categories again by voting in your follow-up email. And without further ado, Nancy, Marco, always a pleasure. And we'll meet back here April 8 for the Winner Circle. I hope you all go out and have a great rest of your day. Take care. Thanks, everyone.