AI space planning in Claude: from prompt to floor plan

Hello, everybody. Thank you for joining us today. We're just gonna give people another minute here before we get started. Alright. I think we have a pretty good audience to get going here. So thank you so much everybody for joining us today. My name is Austin Blass. I'm the director of corporate marketing here at Robin. We're super excited for this webinar. We've been looking forward to it for the past couple of weeks. I think the things that you can do with Robin and Claude are so impressive, and we're really excited to share that with you today. So joining us today, who will be walking through everything is Ciara Peter, who's our SVP of product and Milo McPhail, who is our senior product manager. They've both been working on this for the past few months, and we have some great demos and information to share with you today. Just a couple of quick housekeeping tips before we get going. So we will share the recording and slides and some other resources with you later today, tomorrow morning at the latest. We have a Q and A panel, so please be sure to drop any questions that you have in that section. And as much time as we have at the end, we'll get through as many questions as we can. And just a quick note, everything that you see today is live in Robin and Claude. So that's something you can go in and test for yourself after you set it up. So we have a quick poll we wanted to just launch and get some feedback from everybody. We're just curious, how long did your last space plan take from request to move day? We'll give a minute or two for you all to respond and we'll pop up the results in a couple minutes, but I will pass it over to you, C. R. Thanks, Austin, and thank you so much everybody for joining. So we're all here because space planning is so important. It really shapes how the office is used and all of your investments. And as you know, real estate is usually the second largest cost for a company, so every small decision really matters. Also, when people join and leave an organization and organizations are changing, it is very tedious to try to make space plan edits every week, every time somebody joins. And we're here to help with kind of doing that at scale. But also, offices are really different. In the past, base plans have been based on sort of a one to one desk ratio, expecting everybody to come in three, four, five days a week. And in a hybrid world, there are a lot of ways to just make the office more efficient depending on the real behavior of your employees. The problem that we really set out to solve is that space planning can be still very manual and slow, and a lot of the information that you need is missing. And so you might have questions like whether you can actually consolidate your space, what do we do with the new organization that is joining this building, maybe an acquisition. And these questions come up all the time. And typically in a spreadsheet, this can be a very disconnected process. You have to have kind of, there are a lot of politics that go into the decisions that are made without data. And an IWMS is very rigid. So you have maybe two to three different attributes that you can decide to organize the plan by, and it just doesn't work in 2026. I think I saw a poll come up. Three months. Yeah. So three months is, you know, sometimes the decision making actually takes that long, but, typically, it's kind of a result of a lot... You know, process opportunities. And so we are here to help with that. Alright. So today, we're going to talk about space planning in Claude and MCP. We decided to build our space planning directly in the tools that you use so you have a choice. You have a choice to do this in context of the rest of your work. If you're using Claude and, you know, or Copilot or ChatGPT, or you even have your own custom connector, or you can do this within Robin. But the flexibility is really important. And what we provide is pretty much every data point from your usage. So that is the amount of floors and desks you have, bookings and check ins and what's actually happening in the space, and really some of those more advanced analytics that go into your decision making process. And the process can be completely customized. So what we'll do is start this with a question, and you can really sort of customize this to your organization's needs. And the Cloud Connector will process this in combination with your data. And as you'll see in this example, we're looking to organize around the anchor day peak. So this org has different anchor days for different parts of the organization, and, it's found a way to split the space and really optimize it. From there, you can review the draft in Robin and go ahead and publish it when you're ready. And the great thing about this is you can really ask all these creative questions, but depending on which LLM you're in, you can also combine Robin's data with external systems. So you can imagine an HR system that lets you know when new people are going to join the organization in advance. You can look for who collaborates together if you want to organize by different teams, from Jira, from your sales CRM kind of system, the the decisions are up to you. So we'll walk you through how to draft the right plan, get that into Robin, collaborate on it, and publish it. So here, I'll pass it over to Milo who will do a demo. Alright. Hi, everybody. Thanks, Ciara, so much. As mentioned, I'm gonna drop into Claude to do this. As Ciara alluded to before, this is built on an MCP server, which means that this is available in any LLM that you use across your organization. It will ultimately be available on platform in Robin as well, you can stay right in in that context. What I've done here at the start is invoked, our stack planning skill. Wastik skill is basically a program that runs, in Claude, and it returns to me, the current organization of this building or building Neptune here in, in Robin, in our in our test organization. So this is a pretty standard stack plan, format. As you can see, it both gives me the departmental breakdown over here. And one of the things that we built into this as well is a person to person understanding of who is assigned to each of these seats. So each of these seats represents a real employee that we can then, do some of the fine tuning here in the sandbox UI. As you can see, it also handles assigned... Shared assigned desks here. It can say, okay. Tunde is here on one day, of the week, and then their coworker, Stellen, is here on a different day of the week as well. So they can... You know, there's... All of those, use cases that we have in Robin are also handled here in this UI. That said, I'm gonna use this as a real planning tool for just a second. So I am going to say that we in this office are trying to save on real estate costs. So I'm gonna ask the, LLM to fit the current workforce into as few floors as possible, and I'm gonna say that it's okay to reset, the hotel or hot desks to assign to accomplish that task. We can kick it off, and now this is sort of where it does its magic. I'm gonna draw your attention while it's thinking into a couple other features that are built into this. We have the information on platform already about department, about manager, so I can see this as a breakdown of of manager if that's a way that you organize your teams instead. Neighborhood, for instance, if these are your pre, drawn neighborhoods that already exist, for instance, if you if you have those, built into your floor plan already. And we also have an understanding of job title, so we can group folks of maybe a certain level of employee or hire. So you say your executive team wants to either sit together or move those to individual, you know, solo offices in a in a floor plan. That's really, really achievable with this. As Ciara mentioned earlier as well, this can also get powered up from any external systems that can handle information that maybe we don't have on platform in Robin. That might be an HRIS system. That might be a a better understanding from maybe even Jira boards about what teams work together, most frequently so we can get that behavioral data, including even from, including from, pardon me, including even from Slack messages or Google Calendar or any of the other ways that you can determine what the best way to organize, employees for collaboration is. And that's really where this thing gets gets powerful is when you say, okay. All of the behavioral or or or, HR data that we have, make Robin make the best use of that. Further down in this as well, we also provide a really easy interface to... Oh, pardon me. I'm a go a little a little easier here. We provide a really easy interface as well to start building projections. So on the previous screen, Ciara mentioned there that we have, maybe wanna drop the the... You know, change everybody by 20%. We wanna do an expansion of our workforce. Can you show me what that would look like, what my needs would be, for instance, there if I wanted to increase the size of each department by, by 20%, and then I could move some of those things around here, for instance, as well. So then I could restack my office, my office plan here as well. As you can see... Oh, cool. This is our plan. This has said that we can remove Floor 3 here at a ballpark of $55 a square foot, which... Or a 175 square feet potentially per desk here, which is data that we've just self entered here. It also gives us a ballpark real estate savings of the amount of rent that we've saved per year. So, again, that can be powered up, you know, give us a really good understanding of what the business impact is gonna be of, any move. I like this. I said, okay. This is all... These have all been grouped together. Obviously, we can't fit all of our engineers on the same floor. That's totally fine. But I like this, so I'm gonna say, create this as a draft in Robin. And while it cooks, I'm gonna, drop right over into that to, to show you what happens when it creates it. One second, please. Excellent. It has already updated it, which is marvelous. Okay. Cool. So everyone can see this. This is our updated plan right here. It's already fired over, and this is our current, plan here. So you can see that Floor 3 has, now has has those folks on it. Floor 1 has had its folks moved here. I can have it rename the neighborhoods as well if I wanted to. I probably shoulda just done that in in the other screen here, but this is the current allotment desired for each of our, for each of our our our plans. From there, all I have to do is start the execution of the move itself. Can I do that here in Robin very easily? I can also do it in the LLM. But, to get that kind of consensus, here, I'm gonna invite Ciara to join this draft as a collaborator. A collaborator is a role that we have in Robin that is a sort of... Between a viewer and an editor, it's someone that can move desks around but not publish a field. So when you have those those teams that are like, okay. My team leader wants to do some of the final fit and finishing for where folks should sit, this is a a great way to have them brought in. So I'll invite her, and then I will, say, okay. Great. We... I can notify users of assignment changes. I can export that CSV of changes to my facilities team, and I can now go and review and publish. Let's say that we have alignment on it, and these are the floors that have the changes on them. So I will move through. Likewise, notify the users. I can add a custom message, and I can notify users whose users are canceled. Likewise, send a CSV and then publish the message maybe to my, to my, facilities team. So once that's done, I'm gonna publish those two floors. There will not be an error published in the draft, but I believe it just goes updated. Pardon me. And, and the floor would get published and our floor will be updated. So it's pretty simple to, yeah, just to use the flexibility that your business needs and and get it done in Claude and Robin, on Robin. We'll think. Thanks very much. Thank you so much, Milo. So what we have just seen is the ability to plan your space in ideally, you know, going from months to minutes. We also talked about kind of the longer term planning, but depending on what system you use, in Claude, you can create a weekly task to look for new employees, to assign them automatically to desks, to remove them from remove people who have left the organization from their desks and get a report. Space planning, as we mentioned, is not just something that you do annually. It's not something that you just do when you have a big move. It's something that really needs to happen all the time based on what's taking place in your business. And, if you are looking for inspiration, we also have launched the Robin Prompt library. We will send this link to you in the follow-up email, But, we have several scheduled tasks that you can set up, as well as about 30 different prompts. So you saw the stack planning prompt today, but we also have, several prompts on how to create a hybrid policy and how employees can really personalize and book their schedules based on their lives and what's important to them in choosing days to come to the office. Team planner for team events, and you'll be able to just input your information and copy and paste that into the LLM of your choice. As mentioned, this is available today for all Robin customers, and so there are just a few simple steps to getting started. The first step is your Claude owner will add the Robin app. So you can either ask the owner of Claude or an admin of Claude directly. Excuse me. It is the owner. You'll ask them directly to enable Robin for your entire organization, or you can just click over to our listing and request it, and your Cloud owner will also receive that request and be able to enable it from there. Then as a user, each person will connect to their Robin account through SSO. We also mentioned MCP. So for the different systems, if you have a custom LLM for your organization or you can't wait to get your hands on Copilot or ChatGPT, we have instructions to set that up, and they're fairly similar to this. You also have the ability to turn this off and on. So we hope that you all leave this on, try it out because it's a great experience. But our external AI assistant access will follow your policies. So if you decide that you do not want people using Robin with external data, you can turn that off within Robin. And that's it. The easiest way to learn this is to connect it and just go ahead and try it. Go ahead and start asking questions. Like Milo mentioned, nothing automatically publishes. It is safe for you to get in there and test different scenarios and experience it for yourself. So at this point, I think we'll move on to questions. Yeah. So we have a few questions here. So you touched on this a little bit already, CR, but someone asked about where they can find Robin and Claude. And you can find that just by going to connectors and Claude. I will share the link to our listing in the chat as well. But if you have any issues finding it, you can also reach out to our support team or your CSM. And then can... Sierra or Milo, could you talk a little bit about, the reporting capabilities that we can do in Claude? I can I can talk to that? So if you, have advanced analytics, you can report on any data point in Robin. So if you had imagined a dashboard in the past that has, you know, some kind of very customized capabilities, now you can just type that into Claude, and it will, build the report for you. So I... So that could be anything from, existing usage. You can do predictive analytics. You can mash up Robin data with an external system if you want to understand how different kind of parts of the organization work together, if collaboration data is really important to you you truly can kind of create whatever report you like. And if you want a report sent to you, you can also create a weekly task that will run that report and deliver it to your inbox, Slack teams, whatever you prefer. Cool. Another question that came in. Can you talk about how Claude works with our user permissions in Robin? Yes. So Cloud respects the same user permissions. So if a... Dependent on the role and the permissions associated with that role. So if somebody with a limited access role attempts to create a scenario plan, our tools, the first thing they do is look at what you have access to, and they would not be able to do that. Similarly, the visibility matches. So in Robin, if somebody wants to work with their team and the team has publicly reserved desks, that person will be able to use that same kind of publicly or at least within the organization public information to plan team days, but they would not be able to see, you know, more personal information about maybe the team's past usage or anything that is, you know, meant to be private. So this connector matches and reviews all permissions that are already set in Robin. Okay. There were a couple other questions that came in that our team will follow-up with those individually. So I think that about does it for today. Thank you so much, Tiara and Milo, for walking us through this, and thank you everybody for joining us today. We hope that this was useful for you, and we're excited to hear how you're using Robin and Claude. Thank