Kaizen Casas · Field Notes

What a short-term rental can run on autopilot

Most owners run their rental out of three inboxes, a cleaner’s text thread and memory. Every capability on this page runs today, in production, on a real 4-bedroom coastal rental. Some of it your booking software already does and nobody switched on. The rest is what we build on top. It is listed here because most owners have never seen either half.

4.9★across 80 reviews on every channel
+19%occupancy vs 108 comparable 4BR homes
the market’s average stay length
~20%above market rate on like-for-length stays

Occupancy, stay length and rate vs 108 four-bedroom comparables in the same market, trailing 365 days. Rating averaged across all channels. Live system, August 2026.

What it adds up to: mornings start with a brief, not a scavenger hunt. A guest question at 11pm is read, classified and drafted before you wake up. Nothing depends on anyone remembering anything. Every decision, from tonight’s price to next season’s minimum stay, is made against the operation’s own data instead of a hunch. And 7 in 10 guests write, unprompted, that they want to come back.

Everything this operation runs on Start with what your booking software already includes. Everything after that is built on top.
01
The foundation
21 automations
  • Guest messaging
  • Filling the calendar
  • Getting paid, and protected
  • Operations
  • Reviews
  • Books & admin
  • Direct bookings
02
Guest messaging
4 capabilities
  • One inbox for every channel
  • Automatic message triage
  • Drafted replies from a single source of truth
  • Guest profiles that remember
03
Bookings & arrivals
6 capabilities
  • Instant booking alerts
  • Door codes that issue themselves
  • Pre-arrival verification, twice
  • Damage-protection check on every booking
  • Same-day bookings, same checks
  • A morning operations brief
04
Issues & tasks
4 capabilities
  • Guest reports become tickets
  • Damage triage with evidence
  • Vendor email becomes structured alerts
  • A searchable incident log
05
Turnovers & property care
2 capabilities
  • Photo-verified turnovers
  • Pool chemistry watch
06
Reviews & reputation
1 capabilities
  • Review radar
07
Pricing & revenue
4 capabilities
  • Nightly market capture
  • Comp watching
  • Booking-pace early warning
  • Stay-length engineering
08
Repeat & direct bookings
5 capabilities
  • A real guest database
  • Signal mining
  • Campaigns written from the guest’s own words
  • Attribution to actual bookings
  • Guardrails on all of it
09
Money & reporting
4 capabilities
  • Expenses that file themselves
  • Tax liabilities tracked as they accrue
  • A real monthly report
  • A live property page, rebuilt daily
10
The system underneath
8 capabilities
  • Ask the operation anything
  • A strategy partner, not just a status board
  • Documents as memory
  • Nothing acts alone
  • Plugs into anything with an API
  • Workflows built to fit, in days
  • Your data, in databases you own
  • A private market dataset that compounds
01 · The foundation

Paying for software you use a fraction of

Everything below is already included in what you pay each month

Guest messaging
  • The whole message arc, written once Booking confirmation, pre-arrival details, check-in day, a mid-stay nudge, checkout instructions, post-stay thank you.
  • Inquiries answered before you see them Someone asking about dates at 9pm gets an immediate, useful reply instead of silence.
  • A follow-up after they settle in A short message a few hours after check-in asking whether everything is as expected.
Filling the calendar
  • Gap nights that fill themselves Rules that automatically drop the minimum stay and adjust the rate on orphan nights nobody can book.
  • Extension offers on open nights When the night after a guest’s checkout is unsold, they get offered it before they leave.
  • Minimum stays that move with the season Longer minimums when demand is strong, shorter ones as dates approach and unsold nights start to spoil.
  • Dynamic pricing connected and tuned Your pricing tool wired to your calendar and actually configured: base rate, seasonal floors, last-minute behaviour, orphan-day handling.
Getting paid, and protected
  • Payments that collect themselves Deposit at booking, balance auto-charged on schedule, failed cards retried and flagged.
  • Rental agreement signed before arrival An e-signed agreement triggered by the booking, chased automatically until it is done.
  • Deposits and holds, handled Security deposit or card hold placed automatically on the schedule you set, released after checkout.
  • Pet fees and waivers, automatically Pet fees applied by the booking rules rather than negotiated in the thread, and damage waivers added to every stay without anyone opening a portal.
  • Upsells the software offers for you Early check-in, late checkout, extra guests: offered automatically when the calendar allows it, priced in advance.
Operations
  • Tasks created by the calendar New booking, cancellation, date change: the tasks that follow are created and assigned automatically rather than remembered.
  • Cleaner scheduling driven by the calendar New bookings, cancellations and date changes notify the crew automatically.
Reviews
  • Review requests, timed to land The request goes out when a guest is most likely to write, not whenever someone got round to it.
  • Host reviews on autopilot, with a brake Guests are reviewed automatically, with a hold rule so a problem stay never receives an automatic five stars.
Books & admin
  • Tax set up per jurisdiction Lodging tax configured on the booking so it is collected correctly from the start, rather than reconstructed at filing time.
  • Owner statements that generate themselves Per-property statements produced from the bookings and expenses already in the system.
  • Your books, connected Your booking software wired to your accounting so revenue and fees land in the right place without re-keying anything.
Direct bookings
  • A direct booking site, switched on The booking page and widgets your software already includes, live on your own domain and wired to your real calendar and rates.
  • WiFi guest capture Everyone who joins the property WiFi, including the companions who never appear on the booking, becomes a properly opted-in marketing contact.
Twenty-one automations your PMS already includes. None of this needs us to run it — it needs someone to sit down and configure it once. Everything after this point is what gets built on top.
02 · Guest messaging

Messages missed, and the bookings that go with them

A slow reply costs you the booking and the ranking that would have brought the next one

📥One inbox for every channel

Airbnb, Vrbo and Booking.com conversations flow into a single place within about two minutes of arriving. A reply typed once lands back inside the guest’s own app.

checking three apps and hoping nothing slipped

🔀Automatic message triage

Every incoming message is read, classified, and — when it reports a real problem — filed as a ticket with a category, a priority and the guest’s own words, linked to their booking. Routine questions surface with a suggested answer instead. Nothing sits unseen overnight.

✍️Drafted replies from a single source of truth

Each property has one canonical fact sheet: checkout time, parking, pet policy, quiet hours. Drafts are written from that sheet, never from guesswork, and a person approves every send.

retyping the same six answers forever, occasionally wrong

🧠Guest profiles that remember

Every stay, message and review builds a profile: who traveled with kids, who asked about fishing, who mentioned an anniversary, who said they would be back.

M
Michelle
Aug 8 – Aug 12 · 4 nights · 6 guests
Airbnb
M
Michelle · 2:14 PM
There are a few wasp nests in the outdoor bathroom and I just saw one flying around. We have the kids with us. Also, where’s the switch for the outdoor light?
read and classified · 2:14 PM

↓ filed as a ticket automatically — no one forwarded anything

MaintenanceHigh

Wasp nests in outdoor bathroom

Guest reports a few wasp nests in the outdoor bathroom and saw a wasp flying around, with children present.

Confidence 98%Linked to the bookingNow: Repaired

A real message and the real ticket it produced, guest name changed. Note the second half of that message — the light switch question — was answered as a question, not filed as a fault.

Guest · 11:04 pm Hey! Quick question, is there a coffee maker in the house? And what time do we need to be out on Sunday?
Drafted 90 seconds later, from the property fact sheet Yes, there’s a drip coffee maker on the counter and filters in the cabinet above it. Checkout is 11am Sunday, and I’ll send the checkout steps that morning so you don’t have to hunt for them. Waiting for your approval

Every draft is built only from that property’s verified fact sheet, never from guesswork, and nothing reaches the guest until a person approves it.

4.9
★★★★★
80 reviews across Airbnb, Vrbo and Booking.com
Every stay is reviewed, and every review gets a reply. 4.9 average across every channel
03 · Bookings & arrivals

Unsigned agreements, failed cards, and guests who cannot get in

An unsigned agreement means no card on file, which means no damage protection at all

🔔Instant booking alerts

Every new booking lands on the phone in seconds, including bookings that start as pending and confirm later, which is exactly the kind most alert setups silently miss.

🔐Door codes that issue themselves

A unique code per guest, active at check-in, dead at checkout. Nobody texts codes, nobody forgets to change them.

Pre-arrival verification, twice

Three days out, and again on arrival day: rental agreement signed, card on file, door code issued, check-in information actually sent. Any missing item becomes an alert while there is still time to fix it.

discovering at 4pm that the agreement was never signed

🛡️Damage-protection check on every booking

Each reservation is checked for waiver coverage, and gaps are flagged before the guest arrives rather than after something breaks.

Same-day bookings, same checks

A guest who books at noon and arrives at four gets the full verification sequence on an accelerated clock.

☀️A morning operations brief

Every morning: today’s check-ins with their readiness verified, today’s checkouts with cleaning status, anything flagged, and tomorrow’s preview. Exceptions first.

the daily report every owner asks a VA for and never consistently gets

Tuesday, 26 August 1 check-in · 1 checkout 07:00
Needs you
Check-in today, 4:00pm · 6 guests · 4 nights
Rental agreement signed
Card on file, hold placed
Door code issued and active from 4:00pm
!Check-in details never sent Send now
All clear
Checkout completed, 10:40am
Cleaner arrived 11:15am
Turnover report received, photos verified
8 of 8 rooms captured, nothing flagged
Tomorrow

1 check-in. Agreement still unsigned, day 2 of 3 — no card on file yet, so the incidentals hold cannot be placed.

Exceptions first, always. The guest arriving at four has a door code and a signed agreement but was never sent the details telling them where to use it, and that is the entire difference between a smooth arrival and a phone call from the driveway.

04 · Issues & tasks

Problems forgotten until they become refunds

A reported issue nobody logged is a refund request, a bad review, or both

🎫Guest reports become tickets

When a message reports a real problem rather than asks a question, it becomes a ticket with a category, a priority and the guest’s own words attached, linked to their booking.

problems that live only in a text thread until someone forgets

🔍Damage triage with evidence

Anything the crew flags is paired with its photo and a severity, and routed for review while the deposit window is still open.

📧Vendor email becomes structured alerts

The pool company’s service reports, the cleaning crew’s turnover emails, statements and notices: watched automatically, parsed into data, and turned into alerts when something needs a human. An inbox nobody has to remember to check.

🗂️A searchable incident log

Every issue is filed as a record: what happened, what it cost, how it resolved. Patterns become visible; the same problem stops being a surprise twice.

Ops Log 21 tickets opened automatically from guest messages
Jul 14High

Front door lock not working

Access / EntryGuest reported
STStaff 98%Resolved
Jul 13Medium

Low water flow in hall shower

MaintenanceGuest reported
STStaff 97%Resolved
Aug 8Low

Guest questions a card charge

BillingGuest reported
MPOwner 92%Explained

Priority is assigned, not guessed at later: a broken lock outranks a billing question, and the difference decides what interrupts your evening and what waits for morning.

Turnover — damage needs attention
Serviced 22 August · guest departed that morning
Reported by the housekeeper
“Hay mucho pelo de perro en el sofá y las alfombras de la sala.”

EN There is a lot of dog hair on the sofa and the living room rugs.

Photo shows Heavy pet hair across the sofa cushions and the rug beneath it, beyond a normal vacuum pass
Assessment Looks guest-caused · confirms the housekeeper’s note: yes
Severity Medium · extra cleaning time, no lasting damage
View the photo
Possibly chargeable to the guest Flagged for your review before the security deposit is released. The system never decides what to charge — it makes sure you see it while you still can.

Illustrative alert in the real format. The housekeeper wrote in Spanish and it was translated automatically, because the crew should report in whatever language they think in. The photo is attached to the claim — an alert you cannot verify in one click is an alert you will eventually stop trusting.

05 · Turnovers & property care

Damage found after the deposit window has closed

By the time you see it yourself, the guest is gone and the claim is late

📸Photo-verified turnovers

The cleaning crew’s photos are checked by AI against a staging standard: beds made, towels placed, every room photographed. Misses are flagged with the photo attached, before the next guest walks in.

trusting that “all done!” text

💧Pool chemistry watch

The pool company’s service reports are parsed automatically, and out-of-range readings raise a same-day alert instead of waiting for a green pool.

Beds made and staged — hand towels placed, both queens photographed
All rooms captured — 8 of 8 photographed this turnover
Bathrooms clean — no items left, surfaces clear
!Pet hair, bedroom 2 — flagged with the photo, before the next guest arrives
06 · Reviews & reputation

A rating that slips before anyone notices

Below 4.8 you lose search placement, and most hosts find out from the drop in bookings

📡Review radar

A new review triggers a same-day alert with a drafted response. Anything under five stars escalates immediately, while there is still time to respond well.

finding out about a bad review from a suspension notice

L
Linda
August 2026 · stayed 5 nights
Airbnb
★★★★★

“We could not have had a better place to stay for our family vacation. There were 4 generations in the house ranging from 1 to 83. However, we never felt cramped or had a lack of sitting, eating or sleeping space. The patio and yard area were the hit of our Zen Retreat! While the kiddos played in the cowboy pool, Granddad would grill and the other adults had a comfy couch area to relax and visit. […] I will definitely book here again and whole-heartedly recommend you do too!”

KC
Response from the host AI-drafted

Four generations under one roof and everyone found their own spot.
That is exactly what we hoped this house would do.
The cowboy pool and the grill get plenty of love, but hearing that Granddad claimed the grill made our week.
Thank you for the recommendation, and for spending your trip with us.
Come back any time.

The response is drafted automatically from what the guest actually wrote, then approved before it posts. This is the one place a future guest sees how you treat people — reviews get read, and so do the replies underneath them. Most hosts never write any.

07 · Pricing & revenue

Nights sold too cheap, and nights never sold at all

An unsold night is worth nothing tomorrow, and no discount later recovers it

🌙Nightly market capture

Prices and demand data for the market are recorded every night, building a private dataset of how the area actually books: how far out, at what rates, in which seasons.

👀Comp watching

Comparable listings’ nightly rates and fees are pulled and compared, so a pricing decision is made against the street, not against a hunch.

📉Booking-pace early warning

Bookings are compared to the same point last year, so a soft month is visible while there is still time to do something about it.

noticing the empty calendar when it is already next week

📏Stay-length engineering

Minimum-stay rules timed to how the market actually books: protecting weekend yield when demand is strong, releasing restrictions at exactly the point unsold nights start to spoil.

What a monthly brief tells you about your own property
Revenue, year over year +38%same period last year
Occupancy 47.2%market comps 39.7%
Average stay length 4.0 nightsmarket comps 2.0
Nightly rate, like-for-length stays +20%above market

Compared against 108 four-bedroom homes in the same market, trailing 365 days. Rate is stated like-for-length on purpose: a market that books two-night weekends shows a higher headline rate than one booking four-night stays, and comparing the two directly is meaningless.

08 · Repeat & direct bookings

Paying commission twice for the same guest

You already paid to acquire them once. The second stay should not cost you 15% again

🗃️A real guest database

Guest identities resolved across the masked email addresses OTAs use, with marketing consent tracked properly, evidence and all. The list is an asset the platforms cannot take away.

💎Signal mining

1,923 facts extracted from what guests actually wrote: “we’ll definitely be back”, a sister’s wedding, a dog named Biscuit, kids’ ages. Each one stored with the quote that proves it.

💌Campaigns written from the guest’s own words

A rebooking note that mentions what the guest actually said, not “Hi {first name}” boilerplate. Written by AI, reviewed on a calendar, approved by a person, sent on schedule.

never following up with the 82 guests who said they would return

🎯Attribution to actual bookings

Campaign links are tracked through to real bookings and real commission saved. Not open rates. The number an owner sees is dollars kept from the platforms.

🚧Guardrails on all of it

Any message that mentions money requires human approval, always. Consent is re-checked at the moment of sending, not just when the campaign was written. A guest with an ongoing issue is automatically excluded from cheerful marketing.

Return intent

“This may possibly be an annual weekend for our family at this home!”

82 guests on record saying they would come back

Occasion

“My son is getting married”

→ dated, so the follow-up lands the season it matters

Occasion · Spanish

“por la graduación de mi hijo mayor”

→ read in any language the guest writes in

Preference

“Walking distance to the beach and lots of restaurants”

→ the line their next message actually opens with

Real extractions, quoted as written. 1,923 facts held across eight types, each stored with the sentence that proves it — so a message can never claim something the guest did not say.

09 · Money & reporting

Not knowing what the property actually earned

Revenue is not profit, and no platform you use knows the difference

🧾Expenses that file themselves

Every card charge syncs to QuickBooks nightly, and AI turns messy processor memos into real vendor names and meaningful categories. Tax time stops being archaeology.

a shoebox of statements every April

🏛️Tax liabilities tracked as they accrue

Lodging tax obligations are recorded booking by booking, so the filing number exists before the filing deadline does.

📊A real monthly report

Revenue, occupancy, average rate, booking pace against last year, and where the property stands against comparable homes nearby. A brief that says where you stand, not a dashboard you have to interpret.

🖥️A live property page, rebuilt daily

Upcoming arrivals with readiness status, recent alerts, the review feed and this month’s numbers, on one page that is never out of date.

Booking revenue$8,400
Cleaning and turnover−$1,850
Utilities−$640
Repairs and upkeep−$705
Software and platform fees−$285
Insurance and waivers−$180
Channel commission−$1,260
What you actually kept $3,48041% margin

Illustrative figures. Your booking software shows reservations. Your channel shows payouts. Neither one knows what the property cost you to run — so neither can tell you what you actually made. Because expenses are categorised automatically as they hit the card, this exists every month without anyone building a spreadsheet.

10 · The system underneath

Losing your guest history the day you switch platforms

Every tool you rent holds data you will want when you leave it

💬Ask the operation anything

“Is Saturday booked?” “What did the cleaner flag this week?” “How did July compare to last year?” Answered in seconds, from live data, in plain English.

♟️A strategy partner, not just a status board

The same AI that answers questions works through real decisions: whether to restructure a pet fee, where the cleaning fee sits against the street, when releasing a minimum-stay rule fills nights that would otherwise spoil. Analysis grounded in the operation’s own market data, argued through in conversation, decided by the owner.

strategic decisions made on gut feel at 11pm

📚Documents as memory

Property manuals, house rules and procedures sync in automatically. Answers come from the documents, not from anyone’s recollection of them.

🤝Nothing acts alone

Drafts never send themselves. Money always escalates to a person. Every automated action is logged. The system does the watching and the writing; a human makes the calls that matter.

🔌Plugs into anything with an API

The PMS, the pricing engine, the books, the WiFi portal, the damage-waiver program, the marketing CRM, task trackers and dashboards: twenty-plus services connected and acting as one system. When a new tool earns its place, it plugs in, in days.

a stack of tools that have never met each other

🛠️Workflows built to fit, in days

When the operation needs something no product sells, a custom workflow gets built for it. The pool monitor, the turnover photo checks and the pre-arrival verification all started as “it would be great if…” and were running within days.

changing how you work to fit someone else’s software

🔒Your data, in databases you own

Every guest, booking, message, review, expense and nightly market price is stored in databases the operation owns, not trapped inside a vendor. Encrypted backups run every four hours to off-site storage.

losing your guest history the day you switch platforms

📈A private market dataset that compounds

Because prices and demand are captured nightly, the operation gets smarter every month it runs. Two years in, that history answers questions no off-the-shelf report can.

Connected to your bookings, rates and market data
Should I lower my cleaning fee? It feels high next to other listings.

Probably not, and the reason is in how it displays rather than what it costs.

Your flat cleaning fee$287.50
Spread over a 2-night stay$144 / night
Spread over your actual average stay$72 / night
Your average stay vs the market’s4.0 vs 2.0 nights

On a two-night search you look expensive. On a four-night search you do not. That fee is quietly filtering out the short bookings, which is a large part of why your average stay is double the market’s and your occupancy runs above it.

Cutting the fee would win short weekends and cost you the mix you have. More turnovers, more wear, more cleaning nights, for revenue you are already earning at a better rate.

Want me to model a tiered fee instead, so two-night searches see a lower number without changing the economics of your four-night stays?

Ask a follow-up…

A real question with a real answer, from this property’s own numbers. The point is not that the software decides anything — it is that the case for and against exists before you decide, instead of after you have already changed the price.