Harvard Grad Sees Robots Running Self-Storage Sites.

THE SELF STORAGE REPORT — FULL TRANSCRIPT Guest: Peter Smythe (Co-Founder, White Label Storage — third-party self-storage management; ~300 stores across 43 states) Host: Chris Berg (Abernathey Development — "The AI Storage Guy") Topic: Peter's thesis on how AI transforms self-storage; White Label's new "AI workflow lead" hire; the shift of property management from service to software-enabled service; tertiary-market consolidation; PSA–Welltower; robotics; cost structure Recorded: July 2026 (aired ~July 14, 2026) Background: Princeton undergrad, Harvard Business School; came from real estate development / private equity. White Label was cited in the SSA's July magazine article "Executive Decision" on high-tech talent migrating into self-storage. Source: original recording transcript (speaker-labeled) ======================================================== Chris Berg (00:00): In the July edition of the SSA — the Self-Storage Association monthly magazine — they've got an incredible article called "Executive Decision." It's all about how high-tech talent is now migrating not to more startups, but actually to self-storage. I'm really excited about our guest today. His firm was actually mentioned in this article — they just hired a new head of AI workflow. Not even sure what that means; he's going to talk about that. But more importantly, his thesis around how he sees AI transforming our great industry of self-storage. Welcome to the Self Storage Report. I'm Chris Berg, head of business development at Abernathey Development, also known as your AI Storage Guy. Speaking of AI and storage — be sure and check out next week's video. We just cut it, but I built a really powerful zoning agent that's doing some robust stuff: looking at parcels, what they're zoned for, and what the buildable envelope is. It blew me away to see what Claude and the tech is doing these days. To get that right away, subscribe to this channel so you're the first to get that content. All right, let's jump into today's conversation. Joining us today, Peter Smythe, co-founder of White Label Storage. Peter, welcome to the show. Peter Smythe (01:21): Thanks for having me, Chris. I gotta check that zoning app out — it actually sounds pretty interesting. Chris Berg (01:24): Man, I was blown away. I've got Claude now going into its own Google Chrome browsers, using Zoneomics and Reonomy. And then it goes into a CRM I built and says, bottom line, yes or no — you can build. Here's the buildable envelope. Hopefully you'll enjoy it and maybe have some insight. Peter Smythe (01:45): I mean, you layer that in with land values and rents, and you find yourself a good market for adaptive-reuse self-storage. I love it. Chris Berg (01:54): We'll talk offline about some marketable opportunities there. But let's get into your thesis. Before we do — I love your story. You went to Princeton, then Harvard Business School. I'm trying to imagine you sitting in an HBS classroom going, "You know what, I'm gonna build the next unicorn in self-storage." Was that what you were doing, or how did you end up in self-storage? Peter Smythe (02:14): That's maybe giving us a little too much credit. It's the world that I knew. I came from real estate development and real estate private equity before school, then became fascinated with the self-storage asset class. Through several cycles — even most recently — a down cycle for storage is basically flat rents. Green Street was what turned me onto that; everybody had Green Street in our office. So I sat there thinking, as a guy with no capital, how do we get into self-storage? I had a partner who understood living in tight spaces, having worked and lived in New York City, and we started trying to build this "local locker" light-industrial business model. That led to a core competency of self-storage management, and then to true pure-play third-party management in self-storage. I think it's just the world you come from — the world you know is where you end up trying to solve problems. Chris Berg (03:17): So let's jump into your thesis. Give us some background on White Label if you want, but also your thesis before we get into this new hire — because if you're allocating capital into this position, I presume you've sat down and thought, here's how I see AI transforming this industry, and this is why we've got to allocate capital here. Peter Smythe (03:35): Sure. What's fundamentally different between us and who you'd typically think of when you talk about AI in any industry is that we are not building and shipping product as a pure SaaS business. We are a service provider. Property management is a service business with a human-capital component, and that is the majority of the business. So what AI has done is supercharge our ability to deliver value to our clients. What's unique about us relative to your average property management firm is that we have an engineering team — we are building software, but we're building software to be a better property manager, not to be a software provider to the industry. This most recent hire is a good example of that thesis. We're not just looking at how to do self-storage management better — we're looking at every business unit and trying to figure out where we can get leverage from the advancements in AI. We call this guy an AI workflow lead. The best way to think about it: he's our own internal forward-deployed engineer. There's this concept of forward-deployed engineering — Anthropic just started a company with Blackstone on the thesis of competing with the McKinseys of the world by getting embedded in companies and building unique solutions that give them leverage. Palantir does this too. That's how we're looking at property management, and it allows us to be a better property manager at a lower cost structure. Chris Berg (05:26): I want to hear how you chose this particular gentleman — I think his name is Austin Rockman — and what your pitch to him was. There was a guy who went from NVIDIA to 10 Federal. And you bring in someone who worked at Apple, Amazon, YouTube. Everyone's so enamored with the AI tech space right now. What was your pitch to get him to leave that and come do self-storage? Peter Smythe (05:51): First off, we know the 10 Federal guys — they're owners of assets; we're purely third-party managers. We were on a panel with them recently and admire what they've done; they pioneered the remote-management model in many ways. The pitch, honestly, was framed around that buzzword — forward-deployed engineering. You're going to come into our business as an operational leader rather than someone on our software team. You're going to sit with our ops team and ask: where are the bottlenecks? Where can we create value on non-storage workflows? Looking at our CRM to understand client sentiment, doing a better job communicating with clients and owners, generating better financial reports — these aren't specific to storage, they're core business practices in any industry. What got him excited: he'd been doing automation on Apple products, he was looking for a leadership role, and it was a good fit. Chris Berg (07:11): So what's your ROI expectation on the capital allocation? Peter Smythe (07:16): If you think about a property management business — or any service business — you're probably somewhere north of 10% EBITDA margin, and approaching 50% would be really aggressive for a non-SaaS business. We think a service provider with the right application of AI and software can start to approach high-30s, 40% EBITDA margin. Sequoia wrote a good long post about how SaaS and service are starting to blend. The bigger TAM has always been service, but it's been less attractive because it's lower margin. Now a lot of AI workflows have companies focused on outcomes, not solutions — not "if X then Y," but how do we deliver the entire value prop in an industry better and cheaper? So this one hire sits within an entire roadmap of hiring aimed at getting us from a ~20% EBITDA margin to closer to 40%. Chris Berg (08:47): Walk me through that. How do you go from 20 to 40, and what's happened so far? I know he's only a month or two in — any use cases yet where you were doing X and now you're doing X times Y? Peter Smythe (09:02): Service delivery as a property manager has two buckets of COGS. There's customer service at the tenant level — essentially running a call center and doing triage at the property. That's about 30% of our P&L right now, and it's the best use case for AI. AI has most quickly been able to attack customer-service calls — not all of them. Our philosophy — and this is what separates us from a Lumio or a Unity, who are trying to build a service to be "the AI" for any property manager — is that we already run our own call center, and we're looking at which calls we can do better with AI. Things like: what's your gate code, take a payment, track delinquent tenants to capture their card. AI does those better than a human. Closing a sale requires nuance — understanding what the potential tenant is interested in. So if customer service is 30%, the low-hanging fruit is maybe half of those calls. Then the other side of the business is distributing tasks at the property level — marketing automations, distributing tasks to boots-on-ground or on-site managers. And the third bucket is interacting with the client. That's where this hire is focused: measuring sentiment, understanding whether what we provided was helpful, did they open the report? If not, should we say something different when we deliver the monthly or weekly reporting package? That's something we probably couldn't have done before — we were always QA-ing that with people. Now we're building solutions to flag "here's how we could do a better job," and train our property managers to communicate faster and better so clients have a better experience. Chris Berg (11:43): A few things. Are you customizing the reports for each individual client and how they want the information, or is it the same UI? Peter Smythe (12:02): The UI is the same — I don't know that there's a reason for it to be different. We tailor to the realities at the facility. Traditionally you'd just use the FMS as the reporting package — it's the system of record, has the rent roll and management reports. We built our own data warehouse that pulls from the FMS, from our call center — we report on conversions, number of calls, pickup rate, every call-center SLA — plus Google Analytics for SEO/SEM performance, plus feedback from the ground. There will always be an element of service you can't eliminate: "Hey, remember you weren't happy with curbside appeal? Here's what I did this week." We're always trying to do something either cheaper or better. On this client-service side, he's more focused on better. On margin, our engineering team — led by a guy named Ben Gross — is attacking more aggressively on the low-hanging fruit, which is customer service. Chris Berg (14:12): Let's stay on the asset owners for a minute. What's the number one problem you're solving for them right now? Peter Smythe (14:19): The disparity among owner types and asset types is the hardest part of our business. We're scaling across the bottom of the pyramid — smaller, non-institutional assets in secondary and tertiary markets — though we manage large stuff too. We came up quickly because we can manage small assets remotely cheaper than anybody else. Those assets come in every shape and size. A nicer, larger asset in a core market tends to present similarly to other core assets, and the owners are similar — institutional financial players. In other markets we're dealing with very different owners: someone who has another job and is the boots-on-ground at the property, or a young guy spinning out of a private equity firm trying to roll these up. They all have a different approach to what they want and what their goals are, which makes it hard to deliver a consistent service. That's where we spend a lot of time. Chris Berg (15:47): Is there a way to interview those owners and customize? A legacy-asset owner is solving for something much different than the PE guy who wants to flip in five to seven years. Is that what this new hire will look at? Peter Smythe (16:01): Yes. You start learning about your client at the sales process, when you're pitching them. You learn more during transitions — the first handoff from the person who closed the deal to the transitions team, and there's risk there. By the end, we're trying to bucket these — not just by client persona but by asset type — into tiers, where one property manager manages facilities we believe will behave most similarly based on market type and owner type. How we categorize these things is where we need to do a lot of work. Chris Berg (16:59): Because what we're moving toward is almost a futuristic projection — as we build our CRM and get data, it starts telling us "you should be looking here, not there." Is the next thing you're trying to solve for clients: "we see the market's at a five-and-a-half cap and there are a lot of transactions — you should sell now"? Peter Smythe (17:27): We're collecting the data to be able to do that. We manage about 300 stores across 43 states right now, so we're pretty spread out. We have density in some markets, but it's hard to extrapolate across the very different assets we have. We can "sniper-shot" markets where we have a strong presence. As we scale, we'll have powerful data — that's a goal, to empower clients to make better decisions about which markets to enter and maybe anticipate where rates are going. I've always thought Storable — we operate on top of their FMS, Store Edge and SiteLink — nobody's got better data in the industry. Every once in a while they put out a report and I think, man, if we could only leverage all of your data, we'd understand everything about this market. Chris Berg (18:36): Fascinating. Let's go back to the tenants calling into your call centers. Are you recording those calls? How are you utilizing AI there to increase ROI? Peter Smythe (18:49): A lot of where we spend our time now is resolution rate using AI. If you can keep the call and do a good job within your agentic version of the call center, you're doing it at a fraction of a fraction of the cost of a person. But it's not build-it-and-test-it-once — it's, why did somebody escalate their way out of what was ultimately a simple solution, and how do we tweak that? Resolution rate is huge. We're not quite at 10,000 calls a week — we are around the first of the month — but we get close to 20,000 calls a month. Little swings in percentages are huge swings in cost reduction. Sometimes we should spend more time perfecting the use cases we've already built — our agents already take payment, but a lot of times someone still goes to a human agent to pay. Why couldn't we close that? Perfect that one use case before moving to others. Chris Berg (20:30): Are you analyzing the recordings for the number-one objection, and training salespeople to overcome it? Peter Smythe (20:41): Yes, we record calls, and you treat it the same way you always would have without AI — except the powerful thing is you can QA every call instead of two. We used to QA two sales and two service calls per agent per week, grade them, give feedback. Now we can do sentiment analysis across all calls. That's why you can improve faster. But people still know it's AI and will escalate their way out of it. I like talking to the agents — I try to get them to hallucinate just to see how good the AI is. We deal with people in markets that don't want to talk to AI. Chris Berg (21:45): Are you finding a demographic breakdown — older demos want nothing to do with it, younger demos work with it easily? Peter Smythe (21:59): I don't know that we've extrapolated that well. The most extreme case: our local-locker business in New York, DC, and Boston, where we convert retail storefronts into storage — we won't even take calls at those facilities. It's mostly a younger, yuppie demographic and they'll figure out how to book online; they will not be stopped by us not picking up the phone. So it's a better business decision to force them to solve their own problem and just take the service calls when there are problems. So yes, some demos clearly impact that, but we aren't yet at "here's a senior demo in this market, don't even send them to the AI agent." We'd have to get pretty far down the path to make projections like that. Chris Berg (22:57): Walk me through five years out — what do you see with self-storage through your AI thesis lens? Peter Smythe (23:07): First, the call center is just going to be mostly AI — it'll keep getting better. Stepping above storage: I feel the yield on assets over the long term is going to be higher because operating costs are going to come down. Owning an asset — that'll normalize, people will pay more, the yield will find its level and normalize to a spread on treasuries. But in the meantime, if I had a trillion dollars I'd invest in hard assets, because operating costs are only going to come down. And tertiary-market smaller assets — think about self-storage management minimums. We're probably not going to manage an asset for less than $1,000 a month, because one phone call with one property manager on salary starts to eat away at that. If we can eliminate so many of the other costs, we can lower management fees. Fees are traditionally 3 to 8% depending on asset size. If you can get a small facility down to 3%, you can get more institutional capital into tertiary markets that are currently overburdened by expenses. Right now, if you buy a 75-unit apartment building in Hampton Roads, Virginia, there aren't a lot of property managers who'll manage it — and if you burn through one, you might not have another. So AI will let people chase yield and distribute capital into smaller assets. That's the trend I'm most confident in: tertiary assets become more institutional, and there's a consolidation of institutional capital. Chris Berg (25:31): A couple questions. What I'm hearing is you're building a platform — like what Amazon did, what Costco's done — and as you bring costs down, you give that back to your asset owners. Is that what I'm hearing? Peter Smythe (25:49): Yeah, that's what we're doing. It's not a race to the bottom — we don't want to be a commodity business. But the cost advantage is the most powerful competitive advantage; it's the scale advantage. There will be consolidation, and there will be an institutional player in every asset class. We want to be the institutional player in self-storage across the non-institutional assets. That's what we're building, and I think it'll happen in every asset class. Chris Berg (26:38): With that being said, I'd love your thesis on the Public Storage–Welltower JV. Are you familiar with it? Peter Smythe (26:47): Yeah — Public Storage, Welltower, NSA, SmartStop, Argus. All examples of this consolidation starting to happen; every owner-operator wants some proprietary advantage. Chris Berg (27:12): My thesis on the PSA–Welltower thing: Welltower's been building an AI platform for a decade-plus that a lot of people weren't aware of. I know they do senior living, but I think that's where Public is going. One thing that jumped out at me at the New York SSA — I was shocked how much capital was being invested in data-science people. It wasn't about the real estate; it was "we're taking all our energy to find the best data-science minds we can." So I'm curious how you see PSA leveraging Welltower. And Public Storage stock, as of a few weeks ago, was outperforming the QQQ and S&P 500 year-to-date — I think people are buying the platform. Peter Smythe (27:58): I always think about it from a competitive standpoint — are they going to enter our market sooner than we want them to? You had Extra Space get into smaller stuff a couple years ago with Bargold, then the Jefferson Street portfolio — maybe a false start; we certainly haven't bumped into them on very small assets from the management front. Their job is to deploy capital and report on FFO. If they can make small assets attractive, one way to do that is to learn how to manage them — and to manage them, you need the technology to lower the cost structure. My sense is it's probably easier for Public Storage to partner with somebody who's done the legwork to get there faster. I think this is all to expand their market within storage — I don't think they're focused on getting outside of storage; they're focused on expanding their storage TAM. Chris Berg (29:29): Yes — like what they did with Canada. I'd love your thoughts on this: you look at high-end industrial assets, everything's robotics — what Amazon's doing. When is the day I just say "Hey Tesla, go to my storage site, grab my bike and bring it back," and a robot grabs it, puts it in my Tesla, and drives it back? Peter Smythe (29:59): I'm not the guy to make that prediction. But what I thought you were going to say was: the hardest part of remote management right now is the on-site labor. That's where we spend a lot of time with a lot of variability. How do you find somebody in a tertiary market with 3,000 people to show up, overlock units? So I thought you'd say drone, robot, phone — that's the part of remote management that isn't remote. Some downsides to remote management are security and having a good understanding of what's happening on the ground. Eventually that'll be a possibility — and I don't think it's that hard in storage. We've got robots doing laundry now. At some point something could overlock units, take video, be a moving security camera that can block-and-tackle on a couple of tasks. Chris Berg (31:03): Bro, I love that. Great conversation. Peter Smythe (31:08): I'd love that — I'd deploy it to every one of our assets, and we'd solve one of our major headaches: finding good talent in these other markets. Chris Berg (31:15): I think you're what, three to five years from that? I don't think you're that far away, do you? Peter Smythe (31:21): It's not that you can't do it now. It's: who's going to build that specific product? The cost to build it relative to the size of the market isn't there yet. The only reason it probably wouldn't happen right now is it's not one of the most immediate use cases — you'd focus on the largest problems, like pharmaceutical drug delivery, because that has the most upside. I don't know who's going to spend time building the remotely-managed self-storage robot. You could market-size it pretty quickly — it's 50,000 assets. Chris Berg (32:00): But you see robots now laying tile and doing all kinds of things, so it becomes a transferable skill set — whether it's putting a lock on or whatever. Great conversation. Anything else you want to add? Peter Smythe (32:23): I think people make the mistake right now of deferring too much to AI. It's really easy to sign up for a tool to outsource one workflow and eliminate some labor. Why I like our approach: I try to understand your current capabilities, your current team, and where your needs are — and whether AI can actually, in a closed loop, do the best job at the lowest cost. You should be using it to enhance and supercharge your team, not replace them at this moment in storage. I have not seen any tool do a full-resolution, great job on any individual task yet. Chris Berg (33:22): That's great. If people want to reach out, or say "I want you to manage my assets," how do they do that? Peter Smythe (33:29): On our website, whitelabelstorage.com, you can put your information in. My email is peter@whitelabelstorage.com — I say that because it's not that hard to find anyway. We love being a good partner. We'll also look at stuff if you're talking to an SBA lender and want a sense of underwriting — we'll do that without being contractually obligated. So please contact us. Chris Berg (33:58): Say that again — you'll help guys underwrite assets? Peter Smythe (34:01): We won't allow ourselves to be abused and underwrite assets where somebody's not for real, but yeah, we do it all the time. If somebody's serious and we think we have a chance at managing it, we'll take a crack at it. Chris Berg (34:17): Are you plugging it into your AI agent, or actually underwriting it? Peter Smythe (34:20): A little bit of both. You can do a lot of the work now — I wouldn't have done this pre-Claude, I'll just say that. What's most helpful for people to know: we have a cost structure and management fee that's unique to us. The broker gives you the OM and the unit mix; there's not a lot of work to do around revenue — we can scrape it using our revenue-management tool. What's helpful to a potential client is understanding our cost structure, because on small assets the biggest variable is the expense ratio. If you don't know our fee proposal and all the costs we strip out of the tech stack, you can't really underwrite that facility with a property manager in mind. It's almost like a proposal alongside an underwriting. Chris Berg (35:31): Walk me through it. Say I come to you with a 150-unit asset — how does the cost structure work? Am I paying based on profitability? Peter Smythe (35:43): Usually our minimum fee is going to hit — call it a $1,000 minimum and a 7% fee. We try to price on a cost-per-unit basis. Barring unique characteristics, a 100-unit facility in Idaho and a 100-unit facility in New York cost the same for us to manage. On a very small asset you're probably paying around $10 a unit, which sounds high — but we strip out a lot of the tech-stack costs. Our website is free (an FMS website typically runs $350–$750). The revenue-management tool — if you were using Prorize or StorTrack — ours is free. Our delinquency dialer is our own, so our costs stay low. Our reporting package is our own, included in the fee. And tenant protection is the big piece on small assets — we convey 100% of the tenant protection to the landlord; we don't take a piece, which is unique to us. Add all that up and our side-by-side take versus other property managers is way cheaper. All the line items other managers have — call center, training programs, regional marketing fees, the flat tech fee — are wiped out. There's an inversion point where, at a certain revenue, you're just at ~6% of revenue like another manager if the deal's big enough. Chris Berg (37:59): Check them out — whitelabelstorage.com. Peter, I want to share one thing with our audience. As always: go check out storagedemandscore.com. We built this tool so you can see if there's actual demand where you want to put an asset. I took it from the SSA demand study — demand equals households times 12.6% (the usage rate) times the average 125 square feet. Then you look at total supply (current plus pipeline), and subtract. A developer at the New York SSA told me that unless there's 3x demand in that market, he won't touch it — he needs that cushion in case someone else builds later. So we teamed up with TractIQ. You put an address in, put in the projected project size, hit the blue button, and it spits out a score. It's not the end-all, but it's a powerful metric — there are a lot of developers who are going to be in trouble in the next 18 to 36 months because they thought "if there's 15 square feet per capita, who cares, I'll keep building." We're trying to help protect their investors. So check it out — storagedemandscore.com. Hopefully you'll use it as you underwrite deals for your clients, Peter. Thanks so much for joining us. Peter Smythe (39:39): It's one more input — I like it. Chris Berg (39:54): Next week we're putting out a great video about a zoning agent I built — it blew me away, hopefully it'll do the same for you. Check out storagedemandscore.com. I'm Chris Berg. Thanks for joining us — we'll see you back here next time. Peter Smythe (40:03): Thanks, Chris. [END OF TRANSCRIPT]