Transcript for the video 'Work smarter with TeamworkAI':

welcome to working smarter with Teamwork AI. Nice to meet everybody who's jumping in here. We like to see everybody that's joining from around the world. That's so cool. My name is Renee Demery. I am a CSM here at Teamwork dot com. And joining me today is Steven O'Neil. We are both here at Teamwork. Steven, if you wanna introduce yourself as well, go for it. Yeah. Hey, everybody. I'm probably a familiar face to quite a lot of you. I've appeared on probably too many of these webinars than I can count. I'm group AI product manager here, responsible for both internal and external AI initiatives at Teamwork. Amazing. So the both of us will be running today's session. Like always with our sessions here at Teamwork, these are recorded and will be sent out afterwards, so no need to worry there. And then to as well, throughout this session, you have areas where you can drop in questions, and you can also, put in chat if you have any questions as well. Feel free to utilize that as we go. But here's what the next sixty minutes are gonna look like. So you've probably seen all of what Teamwork has been releasing with a bunch of AI features over the past year or so. Know, maybe that excites you, and maybe you're already a Teamwork dot com AI expert. But, if you haven't dipped your toes in the water of AI yet, that's totally understandable. Regardless of where you're at, we are here with you on your AI journey, and you've come to the right place. So we're going to cover about four key areas and outcomes of where AI is changing how client work is done and hopes that it will inspire you to use our platform here at Teamwork. And then we'll, take a look at some of the upcoming AI capabilities that we're excited about. And then at the end, we'll leave some time for q and a to, go over some of those questions that you all have. So feel free as well, like I said at the beginning, drop in those questions as we go. We'll definitely be taking a look at them. Before we get into the product, I do want to start with something that I hear a lot of our customers say almost every week. So I'm going to ask the group a question, and I want you to honestly think about this answer. So the question is, how much of your team's time last week went towards work that was about work? Things like status updates, meeting notes, figuring out who can take on what, pulling together client reports that takes, you know, about three hours. Yeah. How much of that is actually work that you're putting towards? Most teams that I work with, when they added up, are genuinely surprised The admin over overhead of running projects, not just delivering them, just managing the information around them, can consume about a third or more of a team's week. And there's more of hard truth here. None of that is what you actually are hired to do, and it's not what your clients are paying for. So that's the problem that we have been seeing a lot and that we are trying to solve for a lot with the Teamwork dot com AI portion. So not just building more AI features for the sake of having features, but AI that quietly handles heavily repetitive lifting so that your team can focus on the work that actually moves the needle. And the best part is if you're already in Teamwork, most of what we're going to show you today is already live on your account. So let me hand it over now to Steven because we will show you what that looks like in real time in practice. Great. Thanks, Renee. So, yeah, before we get into the live demos, we're gonna do a number of them today, four to be precise. So cross your fingers for me that I don't get caught by the live demo course on at least all of these. But we've also gonna cherry picked few of our very many AI capabilities at Teamwork right now, kinda going from some of the more trivial time saving pieces to some of the more impactful and transformation elements as we get towards our MCP capabilities, which are growing and growing. So let me just move to the right section here. Okay. So let let's start with AI project wizard. This is something that if you've played around with any AI capability at Teamwork so far, you've probably come across. It's gone through a few iterations and will continually get improved through feedback. So that's probably one thing I'd I'd ask you guys to do. If you're not already playing around with it, please do send the feedback in as you're getting your various outcomes. But what I'm gonna do today is kinda go through a very trivial exercise, probably for all of the demos, just to kinda give you a flavor of what's possible. But the real magic will happen as you test it yourself and as you kinda bring it through the the rounds in the wringer of your on the ground actual requirements. So to kind of go over at the very high level, AI project wizard is essentially a huge time saver for anybody who's looking to create a project from a document or a statement of work or any kind of input that already exists. So you if you already have some sort of idea or you have a lot of fidelity around a project that you need to spin up, this is exactly what this project wizard is for. For today's demo, I'm gonna actually attach a file, you can actually write whatever you want here. It doesn't have to be an attachment. It can be free text open. It can be copy paste from somewhere else. We'll deal with whatever you guys wanna give us. Just for a very quick reference, I've just spun up an example statement of our document here. I think it's seven or eight pages, so that's what I'm gonna attach, and it will kinda go through the process live. Again, is more for illustrative purposes, so please play around with this in your own time after the session. So I'm gonna select that, it's been attached, I'm gonna start project wizard. It goes through a few rounds of reviewing documents, and obviously it's an eight page document, so it's going through every detail, it's finding what's appropriate in terms of any dates or any other elements that I would have had in this reference here, and it's gonna make sense of that as best it can. Of course, this is a shortcut step, but I do still recommend you guys review it at each of these stages to make any smaller modifications, especially when we get to the tasks area. But you're gonna see it would have accurately pulled in the relevant information, including start and end date from the document that I would have given us, and the sensible product description. It's also assigned certain amount of people, but this can be modified by yourself as well as you're kinda bringing it through the the various steps and stages. If I choose tasks now, it's gone to generate tasks that make sense for the body of work that I've uploaded in that statement of work. We'll just give that a few minute minutes. And I will have broken it down into sensible phases as well that align with the reference document that I would have supplied to it. Kinda like most things, the outcomes will be richer and stronger the more fidelity that's given. Again, we're trying to find that balance between you not needing to do absolutely everything upfront as a supply document or supply reference, and that's getting stronger and stronger with each iteration. But it's in a pretty good place as of now with some of the enrichments we've we've been doing over the last couple of months. So then if I click create project, that will go and create that project as I see fit here. Right now this is kind of in a holding pattern or a preview stage. So I can modify any of these before each each of them and the project becomes a real thing. And then when I click create, it will appear in my Teamwork installation as fully fledged project with task list, tasks, and relevant dates here. So this really is the preview and the massage step for you to to get it just right. But the the crux of this is a huge shortcut between start and finish and kind of eliminating the the overly laborious tasks of the past. And that's pretty much it for Project Wizard from a demo perspective. Amazing. From a customer success, manager perspective, I've had a couple people just some quotes from some of our customers. A couple being, you know, thank goodness for task AI wizard. I've cranked out three project task plans in the last hour that would have taken me all day. Or from a graphic designer, I've used AI project wizard to help set up project, like, migrating content and hosting from one website platform to another. The process itself was awesome, super quick, and captured a nice, easy task list with all the relevant actionable tasks. And from this example, you know, it's great to hear from our customers that this is working, and they're actually saving time going into here and using it. At the at this little mark, any questions that have come in? Let's see here. We have one question that came in. Is there a way to factor in a template with this project wizard? Of as of today, that is something we're we're actively working on resolving. From now, this is more of a, from a raw input to a sensible output, but we are keen on making sure the shapes of existing templates are added to this as part of that transformation step. So not today, but very soon. Alright. And then one more question that came in. Can we use AI project wizard to add tasks to an existing project? Yes. We actually call that our task wizard. I don't have it in front of me right now, but there are there are ways inside of existing projects, we'll share those details with the group afterwards, where you don't need to go through the full the full project wizard end to end. We've kinda partitioned just the task part of it, and you can kinda take it off from there. And then one more parter to this question. Is there a way to use some parameters from our templates when using AI project wizard to create a new project? Yeah. So probably the same question as before where we are trying to find a sensible way to have the template wrap itself around kind of content as it's being generated. And so, yes, soon. Soon. Alright. Sweet. Steven, I think we can get into section two. Great. So, our next demo is going to be around, SmartAssign. I've kind of created a Fisher Price version of what you guys are probably used to seeing when it comes to resource management and finding the right kind of assignment flow for you guys. But, again, it's more for illustrative purposes to show at its kind of simplest form what's actually happening. Probably it doesn't need to be said, but just to reiterate, the overall goal here is to make something that's quite laborious kind of week to week, a far mundane and and time intensive task by doing a matchmaking exercise. So if you guys have gone to the hardship of using placeholders to account for upcoming work, and you're at that point in time where you're not quite sure who that body of work belongs to, who it should be best match made towards, this is exactly what this tool is for. I'm gonna go through and talk a little bit in a moment. I will say this is getting better and better. And probably every month, we're kind of zooming in on what the the matchmaking engine is doing. So while now it's doing a pretty good job over the next number of months, it will also factor in skills levels, relationship with customers, and use that as weighting factors to actually pick the right person pretty much every time. So it's not just aware of available capacity within a window, and for today's demo I'm kind of just looking at a single week. But it will also understand all of those other factors I just mentioned, and probably more to come to make sure the outputs are pretty much right first attempt every time. So what I've done here is I purposely just created a group of product designers just for today's illustration. Some of them are over capacity already, so they're gonna be obviously ineligible, that's my expectation. Some people are right at capacity, and others have a little bit of time that could take on this work that I have assigned to a a product designer placeholder. So obviously, real world scenario would be I I had assigned a placeholder at certain point in time, I wasn't quite sure who would be the owner of this work, now it's time to make that decision. So beside every placeholder, you're gonna see this AI spark here, click it. It will actually then load in the relevant people and try to find the right match based on availability essentially. So it's looking here now, we can see that both of these individuals have some time. I'm sorry, let me scroll so the time updates. And I need to find who's most appropriate to take on this particular piece of work. So I'm going to assign it to David here. The flow then will be a complete assignment of that work and a deletion of this placeholder because they're now they're now blank, that work has been handed over, and I transfer that. And then when we look back at David, you can see he's fully packed up from a capacity perspective where before he had that time ready to take in from that SmartAssign. Again, a very Fisher Price, very simple illustration of what's possible. The main ask here is that you guys go away and play around with this yourself if you haven't already, really kind of test it, find the the areas of benefit, definitely find the areas for for growth because that's what we're actually most interested in to make sure this becomes stronger and stronger with every iteration. The main focus here, the main obsession we have is is really, really excellent matchmaking. So anything you find along the way, just let us know. That's pretty much it. Awesome. Yeah. Resourcing is the area where I see the biggest before and after difference when teams start to get serious about using AI in Teamwork dot com. It's really great to be able to just look at this and be able to schedule out in advance. Any questions in this section that we have popping up just as Steven has demoed? I'll give it a moment just to throw in the q and a section. Just to see. I see someone commented, is this on Grow or Scale? I believe it's Scale. Open to being corrected on that. I believe it's Scale. I think it is as well. We have one question that came in. I'd like to have AI look at our existing projects and see where we might be able to streamline. Is there a way to do that? Yes. That's probably a great transition into the next teammate section. We'll we'll hold off on that for a moment. Yeah. It's a great setup. Yeah. We'll hold off on that. Yeah. Good question though, Nicole. Probably one thing again that's probably a good transitionary kind of statement to make as we move from this demo to the next is, for anybody who's had a chance to play around with teammates, the thing we've seen come in most, strongly is the the ask for Remy, who will be our resource management teammate. Of course, there's great overlap with SmartAssign and and the capabilities that might sit within that or be afforded to to Remi. So that is something we're very, very heavily focused on. We need to get it right. So just like any kind of AI tool, it's as good or as bad as the foundation on which it sits. So we're at that point now where we're comfortable with Teamwork's foundation from a resourcing perspective that Remy is just waiting for her turn to to to be released in a while, but we're we're getting there. Alright. And then we have another question that came in from Kelsey, and no need to apologize for all the questions. We love it. Is the work that was loaded in now showing up on the workload side? I don't believe from here. No. Let me just make sure I'm looking at the right thing because it yeah. It's just it's an allocation to scheduler yet. It hasn't been changed into a task. And when that happens or it's linked to a task, you will see that representation. But is just scheduler lock for the moment for today's demo. Alright. As far as questions go, that's about all we have right now. If you want to transition, it's kind of a good segue to go into teammates. Yeah. Of course. No problem. So teammates and probably the MCP will take up the majority of today's demo and and rightly so because this is kind of teamwork really extending its capabilities to be something that's more and more useful in your day to day, but also from a more strategic and transformational perspective. I'm purposely showing you my version of teammates because you're gonna see a few things you guys probably don't have if you're using teammates. It's really to show what we're doing at this moment in time, but also what's coming in the not too distant future. I'll start with Scout. Scout was our first teammate we released. Scout is a personal assistant that has a number of capabilities as of today. Those capabilities will grow both in volume and also in quality. But to touch on a few of them right now, we have a join my meeting skill, which is actually our most popular so far, which allows Scout to join all of your Google meetings, your Microsoft Teams, or your Zoom meetings, and I'll kinda walk through that in a moment. And he will sit in those meetings, he will transcribe, and he will send you very, very useful notes and a rundown of everything that happened in that meeting. I'll I'll I'll take the bet and say they're much better than Gemini's notes for for direct comparison. I'll even say test that yourself if you're using Gemini for any note taking. So it's a pretty stellar outcome. And again, it's getting better and better and will allow for levels of customization in the future in terms of the format that it takes. We also have project catch up. Pretty much it does what it says on the tin. If you've been out for a period of time or you're just after falling behind on something, project catch up is exactly the the job that will help you get where you need to be. Scout will go and look at a specific set of criteria and that you'll decide on. You can create that as a kind of a point and click or a femoral style job or something that runs on a schedule. And there can be a mixture of both, and I'll demo that in a moment. And the main goal here is to help you get caught up on the things that matter in a way that you've actually asked Scout to do. So it won't be generic, it would be very personal. Notification summary again, these are all very much self self descriptive. This will look inside of your Teamwork inbox and catch you up on all that flurry of of messages that have have arrived there and that need your attention, essentially cut through the noise and just help you focus on the things that matter most. And plan my week isn't quite out to you guys yet, but I thought it was worth teasing. This will allow Scout to look at your Teamwork calendar. If you have an integrated Outlook or Google Calendar, Scout will be aware of the meetings that you have to attend, so therefore, time blocks that can't be overlaid on top of them. Scott will also be aware of your assigned tasks, their deadlines, or priority. And if you ask Scott to plan your week, he will do so in a way that makes the most sense. Again, driven by priority with an understanding of what's already in your calendar taking up your time, so not gonna have a lot of or any really overlaps or overlays that are that is an unrealistic ask of you on a given week. So not quite there yet, but very, very close as well. So something to look forward to for Scout. And if I then jump over to a different version of Scout, this is what Scout will look like as you're getting ready to onboard. There's no jobs that have been created. There's no outputs that have been generated yet. But I'm gonna jump over to Scout on this just to give a bit of a different visual demo of what it looks like. So from here I've gone ahead and actually set up a couple of different things. I've configured join my meetings for Scout to join all of my Google meetings that I'm the owner of. But for anything that I don't own, whether that's a Google meeting or whether I'm joining somebody else's Zoom call, etcetera, I can click this. I can ask Scout to join a meeting right now, and just need to paste in that Google Meet, Microsoft Teams, or Zoom link. I'll have the option to have Scout join with a video playback. It's a fun little animation. It's kind of a a polarizing thing. We have lots of people even in Teamwork that love it, lots of people that don't like it so much. I am keeping a list of those who don't like it just just to be aware. But, yeah, it's a little bit of a playful take on something that's very, very useful to give Scout a presence in those calls if you actually want to kinda play around with that. But we have those two paths then, so Scout will automatically join when you're the author and the owner of that meeting if you set up that schedule. Or on the fly you need them to join something, paste the link here, he'll be there in about ten ten, fifteen seconds. Notification summary then, I've set up a manual version of this, but I'll kinda walk through what it would look like to set up a scheduled version just so you can kinda see the mix of the multi instance of each of those jobs. So I'm gonna come here, I'm gonna say I want to actually set this up on a schedule. I want to be caught up, say weekly on a Monday at nine AM on all of the notifications that hit my TMark inbox, and I'm gonna save that job. Of course, the next Monday this will run, and I will get my output. But just for illustration purposes, I will run that now and we'll come back to the output in a moment. Let me just find the right one. Yeah. I get that confirmation also in the table, but I'm gonna run this one now. And let me just check the settings here. Yep. That's fine. And I'm gonna actually not go back as far as it might take too long to generate. My inbox is very, very busy. I'm ashamed to admit. So let me go back to even just last Friday. Right? Hopefully, it won't be too busy there. I'm gonna save that job, and I'm going to run now. So let's go work away in the background, and we'll come and look at the and you're gonna get that notification when he starts, but also when he finishes as I talk through the the other remaining pieces. Project catch up has also proven to be quite a popular job for Scout, where again, you can decide on creating a manual point and click job kickoff, or you can have that project catch up run on a schedule of your own choosing. And if I was to create a new one just to show you, I would create a schedule, and I can create as many schedules as I want. I can create specific filters of this looking at specific projects for specific clients. It's just recent projects, start projects, projects I'm admin of, etcetera. And I can have them run on different cadences at different times. But these are the options you will have. As of now, this will become richer and richer again as we kind of improve the overall capability. But for now, we're gonna have frequency control, time control, and specific control on project types or very specific niche selection of specific projects here. Let me check did that update come in. I believe it did. Still working actually, so I must be busier than I expected. No surprise there. But it actually does show oh, here we go. This shows the output that I just received. Right? I'm not gonna read it out loud because this is my actual Teamwork installation, and it might get me into a little bit of trouble. But you're gonna get the the idea of it anyway. Right? And a summary of things that I've been mentioned in that need my attention that maybe I haven't got back to yet. Maybe I wasn't even aware of them because I was too busy. This catches me up on that and allows me to prioritize the right things sooner. What needs my attention now versus what can wait for another few days. This is actually a good output example then of what the project catch up looked like. Again, I'll just kinda scroll through it, but you're gonna get a summary of activity of projects that also need my eyes on them, whether they're reported issues that need me to give some input, internal feedback, or other elements. But what you're gonna see here in your own installs is projects that need your attention for one reason or another and catch you up rather than having to go through them manually and figure out what needs doing. There's gonna be a very, very strong steer from Scout here in terms of what you can and can't do. I'm gonna try and find an example here as well of yes. So here's a good example of when Scout had joined one of my meetings. It was yesterday, actually. It was a stand up meeting with me and another couple of engineers on the AI team. We had a conversation. We had a very long discussion, you can see that that long detail there, and we have a list of actions. That's important, I suppose, a list of actions to talk about from a transitionary perspective because that's the one thing right now that's absent. Right? You're gonna get the the steer. You're gonna be informed of what Scout recommends what what he recommends needs doing. The next step which we're unlocking at this moment is affording Scout the capability to take on those actions, not just tell you what needs doing, but for you to agree with Scout to maybe modify some of those actions and kick him off to go and actually do what needs doing. So to to pass over that work wherever you're comfortable doing. So that's the one thing we're trying to unlock now. It's also been a big part of the early feedback is the outputs are great, it helps me understand more, but I actually really, really have an appetite to ask Scout or another teammate to actually do some of this work for me, and we're very very very aware of that and we're trying to get there sooner rather than later. So expect that at some stage this quarter. If I jump over to flow then, again, I probably won't go through each of these in the same level of detail, that was more of an illustration. But for flow, same kind of thing works here, you're gonna have the capability to create manual jobs, which again are point and click whenever you feel like it, versus scheduled jobs set up with certain criteria. Here you're gonna have two jobs to start with, and we do have a few more planned that I won't let out of the bag just yet. But but right now, Flow can do project health checks. So if I ask her to do one of those, you're gonna see again I'm have a frequency and a time option and the ability to do a health check on a specific project or on multiple projects. Something slightly similar then on the client facing update, course, is gonna be single in nature, and this is why the multi instance was so important for us to unlock. So you can create however many versions of this for each specific client on whatever cadence you want, and select the client, select relevant projects, set the date range, so how far do you want to look back, how far do you want to look forward on that project plan, and Flow will go away and let you know or sorry, create a client facing update that you can still obviously make some modifications to before sending, but it should be ninety percent, ninety five percent ready to send over to the client. So again, removing that laborious task in the middle of you extracting data from lots of different places and trying to make sense of it in the middle. Flow will do a lot of that for you now. Then maybe to tease Jack. Jack is very, very close. We actually have Jack released to two customers now. As of today, we will hopefully have it out to another ten or twelve by mid next week. But soon thereafter, Jack will be available to everybody else on scale. Jack is designed essentially to be your your custom teammate. What this actually means is that Jack can be responsible for for creating custom jobs, primarily custom workflows, and that can allow you to do pretty cool things in Teamwork itself by leveraging our MCP capabilities and a few other pieces. But probably even more attractively, Jack can help you navigate between different softwares. So if you have information in HubSpot or Salesforce or, Jira or any kind of software that isn't typically easy to get information into Teamwork from, especially in the way you might want, Jack can handle that. I'll give a very quick illustration here without creating the full flow just to get a sense of what you will be able to do very, very soon. You're gonna be able to come in, create a custom job. We'll open up a workflow builder here. There's a number of connectors here or applications that you can now bring in here to create a custom workflow. So for example, just because I was working on one earlier on today, I would drag in HubSpot here. You're gonna see a number of triggers and actions listed out here. You're gonna be able to select one of those actions. So for example, I might say when a deal state changes in HubSpot, I want this complimentary action to happen in Teamwork, and you can kinda create a cascade if you want. It doesn't have to be just that simple of trigger and action. It can be a series of triggers and actions. Where that information can leave one place, end up in another. So again, you're not platform hopping as often or maybe at all in certain cases. So again, I won't go through the full detail of it because there is a of a learning curve to it, and I don't wanna trip over live. But it's more about the capability that's pretty much at your door and will be available in the next week or two weeks at a stretch. So that's it. The other few that you're seeing here, of course, Remi had mentioned is a resource assistant and Cash being the financial assistant. Dottie is gonna take care of everything desk from a teammate perspective. Dotty is actually a little bit closer, so I expect something from Dotty in the next couple of couple of weeks to a month. But we are working **** ** these two. Of course, the reason why we're we're putting so much effort in is we're very aware of the consequence of AI hallucinations and some of the other things that we're probably all afraid of day to day when that happens to do with anything that has a bearing on somebody's time or somebody's money, we need to be absolutely crystal that it's gonna to to do the job we expect it to do. So we're we're we're getting closer and expect it soon. That's it for teammates. Alright. We've had quite a few questions that have come in, so we'll start to jump into those. First one is, could Scout send those meeting notes straight into project notebooks? That is not the first or the tenth time I've had heard that. But, we're trying to figure out ways in the very, very near future to allow that to be shareable. Right now, it's a single output that gets delivered to your your email, but also the updates area for Scout. We wanna make sure that can go wherever you require it. Some of that will be possible actually in connection with what I showed in Jack and where the transcript from the meeting can be picked up, transformed into whatever output you might desire, and sent wherever you need it to be. That can be inside of Teamwork. It can be to a different software if need be. We've had that be asked a few times as well. Can some of Scouts updates or even Flows updates be sent elsewhere without me needing to do it myself? The answer is yes and very soon. Alright. And then we've got another question that came in here. For the project catch up, can that be scheduled as well to run, for example, daily, or is that primarily for the notification summary? Yes. So if I go to project catch up here, just to show you, I can say for the last one day, that's daily. So it will check the last twenty four hours and will run then every single day. Actually, it should have shows schedule there. Yes. So there's the other option for schedule. So if it's manual, you could say for the last one day, but if it's if it's on schedule, it'll run daily at a time you choose with the other criteria and conditions you want. Alright. How do we submit our account to be considered for early access features such as these, Teamwork enthusiast here. There are ways. I can't remember them off the top of my head, but I'll make sure they're shared around with the group afterwards. We definitely have an early access program, and it's it's quite colorful. So, yeah, if you're not already in it, we'll make sure you're included. A hundred percent. I'd almost take this opportunity to to bang the drum for Jack. I really want people who are are are open to exploring some of those AI powered custom workflows. So we can definitely kinda connect those dots after this session, and I'll probably be talking to you personally then at some stage over the next week or so. Yeah. And to Megan and Kelsey for those questions, we will get back to both of you on some of those, like, early access if you want to be a part of that. I'll take note of your accounts. And then last question, Steven, for this section. Onboarding AI teammates seems daunting. If the admin of an organization onboards them, does it impact all users, or do standard users need to onboard them themselves? Yeah. So we've been very conscious of this, and, actually, I'm an admin in this installation so I can kinda show that experience. On this installation, I've granted access to everybody. But if you're in any way fearful of granting the powers of a teammate to somebody whom you may not want to have that kind of connection or that capability, you can come here, and you can decide on a very explicit group within your organization for who you wanna give access to. And, of course, this can be different person to to person or department to department. There may be times where, obviously, because Flow is a a project management assistant, only your PM group or your a subset of your PM group get access to even try it out. Scout's a bit more open ended with probably less consequence if something doesn't go quite to plan. But, again, you have full control to decide on who does and who doesn't have access. Alright. No questions that have popped in since, so I think we can transition over to MCP. Great. And I've mentioned the acronym a few times already. I won't bore you by reading out the list of MCP capabilities here, but I'll make sure that this document is is shared in some form with all of you. Sounds overwhelming, but the MCP is is not as scary as it might sound for those of you that haven't already played around with or haven't played around any MCP capabilities. In short, it allows you to operate Teamwork from wherever you're at. That doesn't necessarily need to be inside of Teamwork anymore. So using your kind of own preferred cloud or your own preferred GPT or any kind of MCP sorry. Any kind of supportive MCP clients you use, including Copilot, you can actually give the inputs there, ask the questions, give the instructions, and as long as the connection is made with Teamwork's MCP, you get the outcome that you are chasing. I'll show an example of that in a moment. I have mine connected to Claude, so I'll go through kind of a live demo of what that looks like. Again, I'll keep it Fisher Price for today, but you'll have an exhaustive list here of what you can and can't do inside of say, team or projects to do with project management, whether they're create, get, edit events, you're gonna see that here, through tasks, task list, milestones, probably most things you can imagine you'll be able to do through the MCP. And again, it doesn't sound there doesn't need to be as technical as as it might at first appear. It can be very easy after the connection is is made. So to kind of give a quick example, and probably for those of you that aren't quite as converse or up to date with what's possible in the MCP, once you do connect it, and I'll also make sure that after the session that the instructions on how to connect Teamwork's MCP to your preferred client is is is shared. It's not that hard. It's maybe about thirty, forty seconds work. First thing I'd ask that you ask your your own LLM, for example, is what can it do? Or if you have a particular use case, ask, can it do that? And then give the instruction. So just for a very quick illustration, and again, to keep it very kind of a very high level to start with and to kind of reduce kind of the barrier for entry and to kind of demystify some of the stuff, I'll ask that question and we'll get some answers. Now this is obviously connected, so it knows exactly what it can and can't do. It's gonna give me something similar to what I got on that large document, but it's probably cut through a lot of the the breakout bullet points and has kind of condensed it into what it can fetch, what it can list, what it can create, and what it can edit. And you're gonna see it here. So regest, write create, update, modify. And it's also understood that it's not just a single MCP, but we have projects, and we have desk, for those of you who use Teamwork desk, and we have spaces. We'll be building out each of these MCPs and their capabilities quite often. Even as of up to date, we're kind of updating them every week, if not every other week, with a bundle of new capabilities. So this is the worst it will ever be, and it's also already quite quite useful and quite extensive in terms of what it can and can't do. So yeah, what I'm gonna do now actually is just go through a very quick demo, hopefully it doesn't fail. I'm gonna do something very, very simple because I need to delete it afterwards. But I'm gonna ask Claude now to create a project for me with a few task lists and a few tasks. Again, something very trivial just to illustrate what happens, but as an example of what you guys should maybe explore to start with, and then lean into the more of the the heavier pieces by looking at the capabilities, what you actually need to do, and how convenient that whole thing can actually be. There will be some MCP connected things to an AI chat I'll talk about in a while as well that will be made as a teamwork, but I'll do that as a quick follow-up. So I'm gonna say create a new project called, let's see, MCP test. Hopefully, there's none of those that already exist. With three task lists for marketing, say, product and sales with three tasks each. We'll say, ask a, b, and c. For example, let's see how it goes with that. Now, of course, I've I'm asking you to do a very few things, so I probably could have done this quicker in the app itself. But of course, if you have a heavier bundle of information, the shortcuts will be immense. And the good news here is if it finds issue, like I was experimenting earlier on with, where I asked it to call create a new project, it found that that project name already existed and negotiated with me, came back and asked for an alternative name or suggested an alternative name, but essentially got me the outcome I was chasing. So there's little to no risk of getting something wrong here. Once Claude or your OpenAI or anything else is connected to it, that negotiation layer in the middle is is excellent. So it will fill gaps, will let you know if something is missing and ask you for the the instruction. So that human in loop element is is still quite powerful here. Obviously, it keeps you up to date. And because I'm demoing it, it feels like it's taking forever, but, yeah, you kinda get the point. But you're kind of seeing what it's doing, what it's thinking, any kind of issues that might be occur occurring in the flow, it lets you know, and you kinda get that feedback in real time. Now again, I'm doing creation events, I could have asked it to fetch certain data, I could have asked it about how a certain project is doing, and what's overdue, what's late, who's over capacity, things like that. And that can meet me exactly where I'm at. And again, if I'm if I'm using LLM, like, cloud desktop or or or similar. So let's just let that complete. And again, probably not the best illustration because I could have done this much quicker and Teamwork made it, but if you imagine this was a much denser ask, the act the the time to resolve wouldn't be wouldn't be very much longer than this. K. So now I know that's great. And just for my own convenience, you guys probably wouldn't need to do this. I'll say share the link. Very own convenience. I can click it and I can actually just scrutinize it. So there it is. I can click that and open the link. And we have a we have a template in Teamwork as well. So you're gonna see these extra tasks added in. You can ignore them. That won't happen in your own piece. We're gonna see it's done exactly what I've instructed it to do with no issue. So it's created a brand new project. It's out of the task list and the tasks I had assigned it. I could have given this a much lumpier instruction, I could have given a lot more information with the project, I could have told it who I wanted to assign to the project, all the usual criteria, again, just through a conversational input, and I get a a very structured, very trustworthy output as a result. So again, a very Fisher Price demo, illustrative of the possibilities. So if you're not already connected to our MCP, please do it. Don't be don't be afraid, be curious, and ask it what can be done. Ask it can it do the things you're after. You'll find the answer more often than not is yes, and then just experiment. That's the best way to kinda take advantage of this this growing engine. It's it's quite incredible already, but it only gets better again every week or every other week. And yeah. And that's it for MCP. Alright. We've had several questions come in regarding this section. So first off the bat, from Christian, we've got there is currently a no delete comment tool exposed in the Teamwork MCP. Any plans to add that capability to the MCP? Right now, if I post a comment via Claude and need to delete it, I have to manually go into Teamwork to do so. Leave that one with me. I don't see any legitimate reason why we would be preventing it, but there's probably an engineer somewhere else screaming as to why that's important. But, yeah, I'll chase that up. Yeah. That that the goal of the MCP is to make things as convenient as we possibly can. So any of those barriers, send them the feedback. We'll try and resolve them. Okay. And then next question, will we receive instructions via email on how to connect MCP to Microsoft Copilot? Yes. So I think we'll do a a full information pack specific to probably the MCP after this. We have I think we have deliberate instructions already set up for how to connect ChatGPT and Claude, but we can also provide instructions on how to connect to some of the more other popular LLMs or similar. Yeah. Okay. And then the next question kind of goes in conjunction with that. Oh, wait. LC deleted it. So that's fair. It was kind of a similar question. And and then to a next one. Here we go. Also, Teamwork MCP currently doesn't expose a two projects create file, and then a file upload tool. Actually, Christian, I'm gonna take that one, and we will relay it over just because it'll it's a little bit complicated. So we'll answer that one offline. The next question is what plan level is the MCP available for? If you're paying for Teamwork, you have access to MCP. Yeah. So, pretty much everybody. So the only people who know the banks now are enclosed in a free plan. So, yeah, go nuts. Yep. And then last one, is there going to be access to Teamwork calendar via the MCP? Yes, I would say. I think anything you can do in Teamwork native, we're keen on making sure that's exposed and made possible in the MCP, kinda like I had mentioned the every week or every other week improvement. So whatever you can't do now, and if it's especially high volume or high value, you can be pretty sure we're chasing it out. Yeah. I think expect it around the same time we get back to that plan my week piece I talked about where Scout is capable of looking at your integrated calendar, the tasks you have assigned to you, their priorities or due dates and so on. Once that goes out, be expectant of an MCP capability soon to follow. Alright. And then we just have one more come in. If we store client calls, transcripts, and Teamwork with a project, will Claude be able to access them and note tech notebook content for updates too? I'm not sure as of now, but probably something I'll tease in a moment through our native our upcoming native AI chat, which will be powered by the RMCP and a few other inputs. We'll likely allow it from from there. And if it's allowed there, it will be possible via cloud and the MCP connection there or OpenAI or wherever else. So not today, but yes would be the answer. Alright. Yeah. MCP seems to be getting a lot of people's attention. So I think this is, you know, a clear sign of, like, where we're all headed as a platform. And, yeah, speaking of where we're heading, I think we'll jump now into, what's coming ahead as far as just, like, some of those AI capabilities from the product team. So I think you've mentioned, Steven, a little bit, obviously, Remy and Jack. But a couple couple things you wanna go over really quick of what's coming up? Yeah. So I've kinda naturally covered Remy and Jack as teammates need a little bit more time in the oven just based on kind of the foundations we want them to to operate on. Every kind of job I illustrated today will be continuously refined. The ingredient is the feedback coming in from all of you. So the more you use, the more the more feedback you give us, the better and the stronger it it gets. There will be a flurry of new jobs afforded to the existing teammates. Like I mentioned, Jack is out in very early access and will be kinda scaling over the next number of weeks. So that's the nearest thing probably to a a more general release for you guys to look forward to. I kind of organically covered the fact that our MCP is a big focus for Teamwork. We really want to offer the the power, but all also the ultimate convenience to everybody who wants to leverage it. So the MCP is a serious investment for for Teamwork for for that reason. On top of that then, we're working **** ** creating a Teamwork AI powered chat, which like I've mentioned a little while ago is connected to our MCP, is also gonna be connected to our help docs, our price plan information, and that workflow builder you saw me kind of half demo there with Jack. So it will be capable of answering questions and taking actions related to all of those things just like the MCP can. The native chat inside of Teamwork can. So you can talk your way to find your answers, or you can talk your way through to actions that are gonna be very, very immediate inside of Teamwork as well. That's probably the thing that's a couple of weeks away from early access. We're moving on that quite quickly just because we're happy with the current state of our MCP and the other pieces I talked about. It will work work very much like your what you're used to in in Clot or similar. But when you're inside of Teamwork, it will just give you that natural language way of of getting to where you need to go rather than the typical start to end flows. We're really trying to make sure we're breaking down those barriers. Alright. Exciting stuff. Before we kind of start to get into the q and a, we just wanna make sure that you all leave on some concrete things to, act upon. So, you know, with every webinar that we have, we there's we call it a graveyard, but, it's what we want to try someday. And I don't want you guys to all end up there. So there's three things that we want you guys to all act on in the next forty eight hours. Each one takes about five minutes. And so this will give you a real feel for what AI looks like and how it fits into your workflow. So three quick wins for AI this week. And, Steven, if you wanna pull up, I just that's also on the PowerPoint. But number one, is use the AI project wizard for your next project. So, even if it's a project that you know inside and out, you can start to see what it builds and how you can pair it to your usual setup. Most teams are surprised by, how close it gets right out the box. Number two, ask out to catch up you up on an active project. So, you know, pick one of the things you've been away from for a few days and see how fast Scout can act, and kinda catch up to speed on the project, see how immediately it can take things, into proportion and act upon it. And then number three, connect the Teamwork MCP to your AI assistant. So whether your team uses ChatGPT, Copilot, Claude, get it connected to your Teamwork dot com account, and try giving it a real instruction. So, you know, example, create a task, assign it to someone, set a deadline. So these are three quick wins that I think all of you should be able to try in the next forty eight hours, and start to play around with it. And then, yeah, feel free to reach out to any of your customer success managers, whether that be me or anybody else on my team, if you want a personalized walk through of any of these, we do this all the time, and it's generally the best ways to start to get up and running pretty fast. Anything to add, Steven, or are we good? No. Probably just to say be curious. Try some stuff, find the potential, find the limits, feed it back. That's how the product gets better and better as all the things we demoed today and, obviously, everything else as well. It's building collaboration with the feedback everybody sends us, so I'm I'm dependent on that. So that's the ask for me is be curious, find potential of what's already there, and also the limits, because that's how we'll overcome them. Alright. So I'm just gonna jump into yes. Also, this is our resources and takeaways. So feel free to take a look at any of these URLs. Super helpful stuff, if you have any questions. But we'll just kinda jump into q and a. So there is one that kind of popped in, as we were just, hanging out. So our notes slash files that are provided to agents stored in Teamwork dot com or on the cloud. If it's on Teamwork, we're getting an increase of storage space within the platform. We are really needing to be careful with how many files we're adding and have reached a limit a few times. I think I I follow-up on that one after the session just to make sure I understand it fully and can support properly. Yeah. Yeah. We will reach back out on that. We also do have information too around how much storage you can have within each of your plans. So I'll send that up to you after. Any additional questions? I think we covered quite a few there naturally anyway throughout the session, so no no pressure if not. One more regarding the MCP. Sorry, Steven. Yeah. Fine. Fine. Task deletion isn't exposed in the Teamwork MCP. I can create, delete, or create, update, complete, and move tasks between workflow stages, but there's no delete task. Yeah. When will that be added at some point? Yeah. So some of the deletion actions were the things we were a little bit more careful of in in the beginning, probably even the comment deletion stuff we talked about a while back is is part of that. It's in need of review now. It's it's probably more consequential when we think about the the discrete permissions that the the MCP is is operating on on behalf of the user. So I I can't give a time frame, but we do want to we do want to solve for that sooner rather than later. But we have kind of created a small abundance of caution in deletion. Because when it when it's gone, of course, there's a recovery option, but we we are looking into it. Yeah. Right. And then we had another one come in. What about desk? Is there any plan to bring sort any of what we've seen today into desk? Yes. So probably two things for desk. Obviously, desk has its own growing MCP capabilities as well. So if you're a desk user, you will have the advantage of taking the MCP and and and exploring the current possibilities. Again, just like the the main projects MCP, it will only get stronger and more robust. There are a couple of AI features already in desk, but probably the big next step is introducing Dottie as a desk based AI teammate. Dottie will take in some of those existing capabilities and and lend more potential and more more optionality to them, which will also represent a whole new flurry of capabilities as well. So I'll probably let my my colleague Laura better talk about that at a future date, but it's not that far away either. But we are very much invested in in both projects and desk particularly, and evolving them in tandem, and sharing whatever capabilities make the most sense at every possible every possible opportunity. Yeah. Alright. One more question for you, Steven. Is our data used to train the AI models? No. It's the it's the quick answer. It's the stuff that probably freaks most people out. And, look, it's it it is a good question, it's an important one, and I I don't wanna over trivialize it, but it's never used to train any third party AI models. Probably just to because this is somewhat legal based, I will read for the first time in this session. Teamwork dot com is is SOC two type two certified, and your data does stay private, so there's nothing to worry about there. So don't let that be a point of apprehension or a point of caution. Everything is is above board. And the AI use your project data to generate insights and take actions inside of your account. It doesn't share it or use it to improve external models. So a long way of saying, no, don't worry about it. We got you. Alright. Well, we have answered quite a bit or you have answered quite a bit of questions today, Steven. So I'll leave you off the hook for now. But at this time, I just wanted to say thank you all so much for joining. It's generally one of my favorite kinds of sessions. Ai has been huge recently, and especially on the team of platform, it's just so cool to see all of the work that product has been putting into it. So as like I said before, we are sending recording of today's session, so keep an eye on your inbox for the next twenty four hours. And if you want to take this further with any specific team or if you'd like a personalized demo, walk through of your account setup, or even just getting Scout or Flow activated, we're here to help. So, again, feel free to reach out through support of your on your accounts or, reach any of our help docs in Teamwork dot com slash support. And then, again, too, any of your CSMs will be happy to help you get on a call and, walk through this with you. So thanks, everybody. Have a great rest of your day. And, again, go try something new in team work dot com this week. Thanks a lot. It's been a pleasure.