Latest / Future of Work Tech with Fexingo: Remote Tools, AI Productivity, and Workplace Software / How AI Is Designing Your Workspace Layout
Transcript
- Lucas: Luna, have you ever walked into an office and thought, 'this layout makes no sense' — like the conference rooms are all on one floor and the quiet zones are next to the break room? Luna: Literally every time I visit a client site. It's like someone just drew rectangles on a grid and called it a day. Lucas: Right. Well, that's starting to change. Companies like Walmart and JPMorgan are now using AI to redesign their floor plans — not based on what some architect guessed, but on actual data about how people move through the space. Luna: So the AI is watching where we sit and then telling the facilities team to move the coffee machine closer to the sales team? Lucas: That's exactly the idea, but it's more granular. Walmart, for instance, rolled out a pilot at their Bentonville headquarters where they put sensors under desks and above meeting rooms. Over three months, the system learned that the finance team barely used their assigned section — they were always camped out near the data science pod. Luna: So the AI basically said, 'just merge those two teams physically.' That's a pretty big intervention for a company with a hundred thousand square feet of office space. Lucas: Exactly. And that's not a one-off. Gartner put out a survey earlier this year — June 2026 — saying that 38% of large enterprises plan to experiment with ai driven space planning by 2027. The driving force is simple: after the pandemic, companies have less office space per employee, and they need to use it more efficiently. Luna: I can see the cost savings. If you can shrink your real estate footprint by 15% because you know exactly how many desks you actually need, that's millions in rent saved. But what about the human element? Like, do people actually want to be optimized? Lucas: That's the tension. The technology comes from companies like VergeSense and Density — they sell sensor platforms that track occupancy. They can tell you which desks are hot, which conference rooms sit empty all week, and even which floor plans encourage more cross-team interaction. But employees can feel surveilled. Luna: Yeah, I remember when Microsoft tried something similar with their workplace analytics a few years back — there was pushback because people felt Big Brother was watching their coffee breaks. Lucas: Right, and that's why the best implementations are opt-in. At JPMorgan, they introduced a system in their Manhattan offices that generates 'neighborhood' layouts — clusters of teams that the algorithm predicts will collaborate most. But employees can override it. They can say, 'I want to sit with the marketing team even if the AI thinks I belong with risk.' Luna: And does that actually happen? Do people override the AI? Lucas: In the first quarter, about 12% of employees moved themselves. But after three months, only 4% were overriding. The system got smarter. It learned that the legal team needs proximity to compliance, not to sales. So the layout evolved. Luna: That's the thing — the AI isn't static. It updates every week based on new data. That's a huge shift from the traditional approach where you redesign the office once every five years. Lucas: Right. And the savings are real. A company called EY did an internal study where they used AI space planning across three of their European offices. They reduced their real estate costs by 18% and saw a 7% increase in employee satisfaction scores — because people ended up sitting closer to the colleagues they actually worked with. Luna: Seven percent isn't huge, but when you consider that most office redesigns tank satisfaction for months, that's a win. But let's talk about the elephant in the room — privacy. The sensors are tracking where I sit, how long I stay, who I meet with. Where does all that data go? Lucas: That's the critical question. Most vendors aggregate the data — they don't show facility managers 'Luna was at desk 3A from 9:15 to 11:22.' They show 'desk 3A was occupied 72% of the morning.' But the data pipeline itself can be vulnerable. There's a startup called Envio that encrypts the sensor data at the device level and only releases aggregate statistics. They've been getting a lot of attention from companies worried about GDPR and employee trust. Luna: So the technology is there to protect privacy, but it's an extra cost. And when companies are trying to save money on real estate, they might not spring for the privacy upgrades. Lucas: That's the tension. But I think we're seeing a market solution — if a company rolls out a creepy sensor system, employees will just work from home more. That's the ultimate check. The AI only works if people actually come into the office and use the space. Luna: Good point. So AI layout design isn't just about efficiency — it's about making the office a place people want to be. If you get the layout right, maybe the commute feels worth it. Lucas: Exactly. And I think we're still early. Right now most of the patterns are about adjacency — who sits near whom. But the next frontier is dynamic layout: desks that move on robotic casters, walls that slide based on the day's needs. There's a company called Ori that makes robotic furniture for small apartments, and they're starting to pitch it to offices. Luna: So in a few years, your office could reconfigure itself every morning based on the calendar invites? Lucas: That's the vision. If 80% of your team is remote on Tuesday, the AI shrinks the collaboration zone and expands the quiet area. On Wednesday, when everyone is in, the walls slide back. It's wild. Luna: Before we go there — and I want to hear more about those walls — I just want to mention something about this show. We don't run ads, and that's deliberate. We want these conversations to be as direct as possible, no commercial breaks. Lucas: Yeah, and that's possible because some of you who listen choose to support us directly. If you've gotten something useful out of today's tech deep-dive and want to keep it ad-free, the simplest way is buy me a coffee dot com slash fexingo. No pressure, just a way to keep this sustainable. Luna: And it genuinely helps us keep the focus on the tech itself, not on selling you something. So, back to those robotic walls — Lucas, how do they actually work in a real office? Lucas: So Ori's system uses a motorized partition that runs on a ceiling track. The AI scheduler — connected to your calendar system — decides at 6 AM whether the wall should be open or closed based on the day's meeting density. One client, a law firm in Chicago, uses it to create temporary deposition rooms that vanish at 5 PM. Luna: That is so much more elegant than having a conference room sit empty for 20 hours a day. So what's the catch? Cost? Reliability? Lucas: Cost is still high — about $50,000 per motorized wall segment. And there's a maintenance concern. But the bigger issue is that the AI can make mistakes. If a partner books a confidential client meeting and the wall decides to open because it 'predicted' a collaboration session, that's a breach. Luna: So you need human override for that too. I'm sensing a pattern — AI as advisor, not dictator. Lucas: Exactly. The best implementations give the facility manager a dashboard where they can approve or reject the AI's layout for the next day. The AI learns from those decisions. So it's a partnership. Luna: I like that framing. AI helps design the space, but humans still decide. That feels like a smarter version of the open-office debates we had ten years ago. Lucas: Yeah, the open office was a one-size-fits-all solution. This is dynamic and personalized. The next step is integrating individual preferences — like, the AI knows I prefer a standing desk near a window, so it reserves that spot for me unless someone more senior claims it. Luna: That's going to create some interesting power dynamics. But I think it's better than the free-for-all hot desking that leaves people hunting for a seat every morning. Lucas: Absolutely. And that's where we'll see the next wave of innovation — not just optimizing square footage, but optimizing the experience. The office as a responsive environment. Luna: Alright, so if someone wants to start looking into this for their own company, where should they begin? Lucas: Start small. Pick one floor or one pod. Install passive sensors — the kind that detect Bluetooth signals from phones, not cameras. Run the experiment for a month, then let the AI generate a proposal. The key is to involve employees from the start. Tell them, 'we're trying to make the space work better for you,' not 'we're tracking your every move.' Luna: Communication is everything. And maybe offer a coffee machine relocation as a peace offering. Lucas: Exactly. Put the coffee near the people who actually talk to each other. The AI can figure that out.