Four cabins, running like a portfolio.
A short-stay cabin operator with four properties, four booking channels, and a phone that never stopped. We embedded Otto to run the guest communication and channel operations, so the owner could run the business.
The work that keeps a business running is not the work that grew it.
The cabins are good. The locations work. The bookings come in. Behind them sits the reality every short-stay operator knows: a guest emails at 9pm asking for the WiFi password. Another asks about pets. A booking lands on one channel and the calendar has to update on the others before someone double-books.
Two of the four cabins were not even live on a major booking channel. Nobody had the time to set them up properly. Listing copy drifted out of step across platforms. A contractor handled guest messages, but coverage was patchy, and anything after hours fell back to the owner.
The owner was not failing at any of this. He was carrying it. And the carrying was the problem: every hour spent answering “what time is check-in” was an hour not spent on growth.
Around the system of record. Not next to it.
The decision was not to hire another VA, plug in another SaaS product, or build a chatbot. It was to embed an AI operations layer: a single agent, Otto, inside the existing stack, running the recurring, communication-heavy, channel-management work that the business needs but the owner does not need to do.
The property management system stays the system of record. Booking, calendar, and listing data lives where it has always lived. Otto reads, drafts, sends, and updates against it: not next to it, not in parallel, not as a replacement. Around it.
Phase 1 was scoped to two modules: the guest concierge and multi-channel distribution.
Two modules. The whole guest journey.
Runs 24/7. When a message lands (direct booking, channel guest, or email), Otto reads it, pulls the reservation context, classifies the intent, and answers. FAQs, pre-arrival questions, in-stay requests, local recommendations, review requests: handled in seconds from a per-cabin knowledge base that knows the WiFi password, the access code, the pet policy, and the two restaurants worth recommending. Every response is logged and bounded by a confidence threshold.
When confidence drops, sentiment turns, or a guest mentions cancellation, safety, or anything urgent, a human gets it, fast. The team member first, the owner if no response. Where booking platforms restrict automated sending, Otto works in draft-assist mode: it writes, a person clicks send.
Confirmation the moment a reservation lands. Instructions seven days out. A reminder forty-eight hours before check-in. A welcome at 8am on arrival day. A thank-you and review request twenty-four hours after checkout. None of it requires a human. All of it lands on time, every time.
Otto configured and validated the connections from the system of record out to every booking channel, cabin by cabin. When a booking lands on any channel, availability blocks on every other channel within seconds. Double bookings stop being a thing.
Listing content (titles, descriptions, amenities, house rules) is drafted from the same per-cabin knowledge base and formatted to each platform’s requirements. The owner reviews before anything goes live. One update propagates to every active channel.
What changed.
Before
- The 9pm WiFi question went to the owner.
- Calendars raced against double bookings.
- Two cabins sat dormant on a major channel.
- Listing copy drifted across platforms.
- After-hours coverage was patchy at best.
After
- Routine messages are answered inside two minutes, any hour.
- A booking on one channel blocks every other within seconds.
- All four cabins are live on every channel.
- One update propagates everywhere.
- The owner hears about what needs him. Nothing else.
Four entries from the record.
The owner stopped being the after-hours support desk. Mornings are no longer built around triaging guest messages. The routine ones are already answered, and the rest arrive with context attached.
Pre-arrival instructions, check-in reminders, welcome messages, review requests: the milestones run themselves. The cabin’s information reaches the guest at the right point in their journey without anyone having to remember.
Channels that sat dormant for want of setup time came live. The cabins are now visible where the bookers are actually looking. Distribution stopped being a constraint.
Bookings that were not happening start happening.
Otto does not run the business. The owner does. Otto does not negotiate with land partners, choose the next cabin location, or decide what the guest experience stands for. That is the owner’s work. That is where the value is.
One embedded agent, running the operations so the operator can operate.
The numbers will be the run log’s, not ours.
The measure for this work is concrete: nights filled that would not have been filled, bookings on channels that were not previously active, and the hours no longer spent answering the same questions. For a four-cabin portfolio, one additional booked week per month covers the ongoing work comfortably.
We publish verified numbers or none. Figures from this engagement are available on request once the measurement period closes.