I run two rental houses in the north of Okinawa. “pono 118” sits in a quiet neighborhood in Nago; “pono 02” is near the water in Motobu. I have no employees. With the exception of my cleaning partners and some help with guest messages, the bookings, the pricing, the guest register required by law — all of it is me.

There’s one more member of the team, though, and it’s an unusual one: an AI. Not “a tool I adopted” so much as a colleague who quietly handles the overnight messages and the check-in paperwork. How that happened — and why it was never about efficiency, but about the experience on both sides of the door — is what I want to write about.

The mess

Running a small rental is more granular than it looks from outside. Messages arrive whenever they arrive: how to check in, how to find the place, how the appliances work, what to do on a rainy day. Each one takes a couple of minutes. But a day sprinkled with unpredictable two-minute tasks is a day in which you can’t do anything that requires a whole mind.

“Breaking point” sounds dramatic. What I actually remember is this.

Every night, I stayed up until the guest had finished checking in. When a flight ran late and arrival slid past midnight, I couldn’t sleep properly — not while waiting, and somehow not afterwards either. Guests who had said “we’ll be fine” would find themselves at the door, unsure how to get in. I’d sent the code and the instructions in advance — but of course they hadn’t studied them; who reads the manual carefully in the middle of a trip? I wouldn’t. Then, the moment they were inside, the questions would arrive in a burst. How do we get hot water. Which switch turns this light off.

I want to be clear: I like talking with guests. It might be my favorite part of this work. But when it never ends, something tightens quietly inside you. And when you’re short on sleep, the notification sound — a sound that should mean something good — starts to feel like a threat. There were nights I caught myself thinking please, no more messages tonight. About the part of the job I love most.

This is not a complaint about guests. Nobody did anything wrong. It’s simply what I experienced — and everything that follows came out of it. When the system underneath doesn’t hold you up, even the work you love can start to frighten you.

There’s a second layer, too. In Japan, lodging is regulated: we’re required to maintain a guest register, and to verify passports for guests from overseas. Every booking quietly adds paperwork. Pricing is alive as well — season, day of week, how the calendar ahead is filling. All of it doable. And precisely because it’s all doable, all of it lands on one person.

I had three options: hire someone, shrink the business, or change the system. Hiring, at this scale, eats the profit. Shrinking felt like a loss I didn’t want to accept. So: the third. With one rule I set from the start — never optimize for efficiency itself. Only adopt what makes both sides better: the guest’s stay, and the operator’s day. An “automation” that works by hollowing out one of those two isn’t worth having.

Getting the upstream in order

From then on, whatever free time I had went into building the system. I’m not a professional programmer. But we live in a time when AI can write code — so my real work was to put each task of running an inn into words, and sort out where judgment was genuinely required and where it wasn’t.

Handing things over one at a time, the AI stopped feeling like software I’d installed and started feeling like a teammate quietly carrying the fine-grained work. (Guest messages now have human help behind them too, from a partner — same philosophy, applied to people as well as systems.)

What I learned in the process: the work of an inn splits into upstream and downstream. Upstream is information and preparation — knowing the bookings, keeping the register, writing the guidance, having a reason behind the price. Downstream is the guest’s actual experience. When upstream is a mess, it eats the attention that downstream deserves. Get the upstream in order, and the downstream starts to flow on its own.

My mornings now look like this. I get up and look at one screen: today’s arrivals and departures, what came in overnight, and whether anything needs me. If nothing does, that’s the end of it. The questions that arrived during the night are already sorted, and only the ones that need a decision come to me. That isn’t neglect — it’s an entrance I built so that what reaches me is the part only I can answer.

The check-in paperwork now happens on the guest’s own phone, in six languages: Japanese, English, Korean, Traditional Chinese, Spanish, German. The register fills itself in and I confirm it. Passport verification was rebuilt so it doesn’t leave the guest standing there waiting.

Pricing comes to me each day as a proposal with its reasoning attached. I decide. I think that part stays human to the end — but having someone lay out the material to think with, every morning, changes the quality of the decision entirely.

Where the time went

Honestly: the freed-up time didn’t all become life. I spent it sharpening this system, and it turned into a company. But one thing did change, definitively — I got back the time to think about my guests. Boiled down, running an inn is a single cycle: prepare, welcome, send off, prepare again. I’m back at the center of that cycle.

Why release it as products

I built this for myself. But partway through, I kept seeing the faces of people with the same problem: operators running a handful of properties alone or with family, at a scale where you can’t hire an engineer, doing work as detailed as a large hotel’s.

So I’ve been carving out what I use every day into something others can use. The first piece is that one screen I look at in the morning — pono WINDOW, from $9.90 a month. I didn’t want to build an expensive tool: I’m exactly the kind of operator who will try something without hesitating at that price, and hesitate at anything more.

Alongside it: pono CHECKIN for check-in and the register, pono CONCIERGE for guest communication (in progress — early access opening soon), and pono LISTING, a health check for your listing. One rule for all of them — if I don’t use it at my own inn, it doesn’t ship.

What’s next

I’ll write the origin story of each tool: why it became necessary, how the first version failed, how I actually use it now. And I’d be glad to hear how you handle the same things at your place.

Our inn in Okinawa → stays.ponoand.com The tools → pono