At Armory, we have a strong track record of backing restaurant tech entrepreneurs, but restaurant experience has a tendency to pop up across other areas of our portfolio as well. Ahead of the announcement of their Seed 2 Round, which Armory led, Besty cofounder Sam Dundas sat down with me to discuss his entrepreneurial journey, the state of software in the short-term rental vertical, and how his experience working as an expeditor in a restaurant kitchen as a teenager shaped his perspective on streamlining workflows and running a company.

Anthony Santaro (AS): Sam, let’s start with your origin story. Before founding Besty, you were an Airbnb Superhost and worked at Planned, a Series B startup based in Montreal. Can you walk us through your path to founding Besty?

Sam Dundas (SD): It really all started with COVID. I was the first employee at Planned, joining right out of university in 2018, and my plan was to ride it out with them until the end. Then COVID hit, and like a lot of businesses, it added another five to ten years to the horizon of what an event company could look like. I’d always known I wanted to start a company, so I went to the founders and said: I need to find a co-founder and go do something on my own, even though I didn’t have an idea, a business, or a co-founder yet. That was July 2020. I spent the rest of the year interviewing potential co-founders on AngelList. Aarlo was something like the ninth person I met, right at the end of 2020, and we instantly hit it off.

AS: That’s a great segue. Can you give a brief background on Aarlo and how the two of you divide responsibilities at Besty?

SD: One of the cooler things about us as a company is that we’re born out of a pivot. I officially quit Planned to work with Aarlo full time in February 2021, and he was the first person to introduce me to generative AI, well before the big ChatGPT moment. He explained it so simply: you have two AI brains working together, a generator and a discriminator. Imagine you’re tasked with painting a rose, and I’m the critic who says, “that doesn’t look like a rose, try again.” We go back and forth thousands of times in a very short period, and boom, you have a beautiful photo of a rose. He has this gift for making sophisticated topics digestible.

He’d been following the space long before it was cool, has been programming since he was nine or ten, and founded several startups before my time. He brought the deep technical experience I was looking for, while I bring the relentless sales and go-to-market side.

AS: On the sales side, what’s the typical eureka moment for a Besty customer?

SD: A lot of our customers are heritage businesses (sometimes second- or third-generation vacation rental operators) who understand AI is a big part of the future but are rightfully skeptical about where it fits into a business that’s fundamentally about human-touch hospitality.

For them, it’s all about meeting them where they are: can we agree there are things your team does today that they don’t like doing? Let’s start there.

But the biggest moment, by far, is when customers go live and flip on autopilot, and it does exactly what we said it would do. Not for somebody else, but for their properties: talking to their guests, resolving issues, generating tasks and work orders, charging cards, updating systems. They’ve been oversold a lot of AI. When they see it working on their own business, they’re just like, “holy smokes.” I always tell them: I can’t wait for you to be on the other side of this, to see it and not just hear about it.

AS: You’ve been deliberate about not becoming a property management system (PMS) yourself, and instead integrating with a ton of different PMS’s. What was the rationale?

SD: Before we built this, I was already using a PMS for the properties I managed and owned with my brothers, so I had a good sense of how hard it would be to switch. People joke that switching PMS’s is like open-heart surgery. The PMS is baseline utility software; we didn’t need to reinvent the wheel. And if we’re already asking customers to do one hard thing — adopt AI — we can’t also ask them to switch PMS’s.

Jensen Huang says: “Do as much as needed, as little as possible.” What that unlocks is that we can grow within the whole ecosystem without picking favorites. We pay revenue share to each PMS, they see it growing, and they think, “this is clearly working.” That’s been a great way to establish credibility and grow faster.

AS: You’re becoming a thought leader in the space. Can you talk about the conferences you’ve been speaking at?

SD: At the Track hospitality show in Austin this year, we had ten people from our team and did a keynote-style presentation on the main stage: these systems are as bad as they’ll ever be today, and here’s what they can already do in your business. This time last year we were working with a couple of you; now there are hundreds or thousands of properties running through Track on Besty.

Many of these operators have been around fifteen or twenty years, and in the startup world companies live and die in three to five years, so customers are as interested in where the founders are taking the product as in the product itself. Showing up consistently has built real equity in the brand. People say, “I saw you the first time, I saw you the second time. I can tell you’re investing in this.”

AS: Going back to the product: it’s always learning, and it can be quite bespoke to a company’s standard operating procedures in a scalable way. Can you talk about how Besty learns over time with specific customers?

SD: As we grew past 100 customers, each paying mid five-figures (and many breaching six), they were low-maintenance but had a lot of bespoke feature requests. Instead of building incremental features per customer, Aarlo built a product within our product: Besty’s platform is now aware of what it can and cannot do. Every feature request just becomes a workflow, written in natural language directly into Besty: “if Anthony is staying with his family and adds family members in the check-in form, add them to the reservation and send them automated messages.” It generates the workflow, and we approve it. The number of custom features we had to build per customer has gone from five or six down to basically zero. We’ve taught customers how to fish.

The AI knows it can respond to guests and generate tasks, who the cleaners and maintenance people are, who the owners are. If you have a cleaner who doesn’t want to use an app, Besty can text her whenever somebody books, wait for a yes or no, and reply accordingly. It mirrors exactly how people actually work.

AS: One feature that’s really cool to call out is the revenue upselling engine (e.g. autonomously selling orphan nights, as they’re called in the industry) and the success-based component that adds to the business model alongside the subscription.

SD: It’s been just over a year since we released the v2 product. In June of last year, our first batch of users did about $35,000 in upsell revenue in the first month. This May, I think we did half a million.

That’s the AI handling orphan nights, early check-ins, selling pool heat. A guest says, “my flight gets in at 1 PM,” and the AI recognizes they might want an early check-in, sends the message, fulfills the order, captures the payment, and updates the PMS. And instead of creating a task for someone to go set the Aqualink pool heat to 78 degrees at 4 PM, our system now spins up a browser and does that last piece automatically too. It can communicate with any third-party tool, no API needed. It can use a browser, make phone calls, send texts.

Closing the loop end to end is what makes people willing to pay a percentage of revenue. If you just send a message saying “I filled a gap night for you,” are you going to pay five, six, seven percent for that? Maybe not. But if it’s done entirely end to end, people are a lot more likely. Call it agentic upselling.

Fun side note: Stripe reached out a couple of weeks ago. One of the founders pulled up an internal leaderboard of the highest agentic usage of their new agentic API, and Besty was number two. We got a shout-out at their all-hands (apparently one of the Collison brothers mentioned us). The funny thing is we’d been doing millions in revenue and never even had a rep there until then.

AS: We were honored to lead the seed two round earlier this year and have been so excited about the partnership with you, Aarlo, and the team. Can you talk about why you came to market and what you wanted to do with the funding?

SD: This was a sleepy enough industry that it wasn’t the obvious place for AI to land, compared to massive monolithic markets like healthcare or insurance. Within real estate, we’re in hospitality; within hospitality, short-term rentals; within short-term rentals, professionally managed properties. It’s a really discrete place to start.

And the beauty of this industry, versus restaurants or other hospitality fields, is that the service is delivered through an inanimate object that doesn’t need to be staffed. As long as the property is clean, accessible, and what you paid for, it can in theory be run almost completely autonomously. Roughly half of these properties still aren’t professionally managed today. It’s a fast-growing, still very overlooked market.

AS: On scaling: you’re increasingly signing six-figure contracts and moving upmarket to larger enterprise customers. What has that been like?

SD: The larger folks are slower to move and have every incentive to wait. In this new category of AI automation for short-term rentals, probably over a hundred vendors have popped up. So the larger players said: let’s watch the knife fight. Let’s see who gets traction, who shows up at the conferences, who our peers and competitors are using.

Another dynamic: a lot of large private-equity-backed roll-ups are buying our existing customers. When they take over, they ask, “why are we buying this company? It’s so well run — what software are they using?” And the answer is Besty. Now they’ve underwritten it firsthand. It’s not really feature-bound. It’s the Mr. Jones effect, listening to what your neighbors are doing. All these lines in the water from two years of showing up consistently are now converging at once.

Rapid Fire Questions

Favorite restaurant in New York City? The 86.

Favorite neighborhood? Greenpoint.

Favorite Knick? Brunson.

Best Ironman time? My best ones are all my worst ones. Nice had a seventeen-hour cutoff and Texas had a sixteen-hour cutoff, and I made both by a couple of minutes.

How many Ironmans and marathons have you done? Two Ironmans and two marathons — not counting the marathons inside the Ironmans, of course.

Who or what has been your greatest inspiration? My dad, for sure.

A business leader you look up to? Travis Kalanick.

If you couldn’t work in technology at all, what would you do? I’d be an expediter in a restaurant. It was my first serious job. You’re basically the quarterback between the kitchen and the servers. Peanut allergy, side salad, the right plates to the right tables. It’s the triage point, really the heart of the kitchen. In high-end kitchens it’s usually the head chef doing it, but lower-end restaurants let scrubs like me do it. I put myself through school as an expediter, and I joke that when all this is over, however it ends, I might just go back and do it again.

Favorite book or podcast? The Acquired podcast. It’s just so entertaining. My favorite book would be one I’ve read a few times: The Upside of Stress.