Trust

AI that answers to people.

AI Otel builds hotel software with AI inside. Each hotel’s data is kept apart; a person confirms every sensitive AI step.

Request a briefing Where data lives

The hotel team’s AI assistant: 207 actions

Builtworking, ahead of launchCounted June 2026

A person confirms

Confirmed again

Illustrative
107 Run directly

Shown

The hotel’s system

90 Need a person

Changed

10 Need two confirmations

Canceled

Run an action yourself.

Read the figure
  1. The hotel team’s AI assistant: 207 actions, counted June 2026Builtworking, ahead of launch
  2. 107 run directly, for example “Show today’s arrivals”
  3. 90 need a person to confirm, for example “Change a room’s price”
  4. 10 need two confirmations, for example “Cancel a booking”
  5. The hotel’s systemBuilt
02 / 05Every action

207 actions.
None improvised.

Everything the hotel team’s AI assistant can do is listed in advance, from the front desk to accounting.

Counted from the assistant’s action list, June 2026

What the assistant does today
Read the figure
  1. Builtworking, ahead of launch The hotel team’s AI assistant: 207 actions, counted June 2026. 107 run directly, 90 need a person, 10 need two confirmations.
  2. Front desk: 13 actions. 5 run directly, 7 need a person, 1 needs two confirmations. For example: “Reopen a checked-out stay”.
  3. Reservations: 14 actions. 6 run directly, 6 need a person, 2 need two confirmations. For example: “Cancel a booking”.
  4. Rooms: 7 actions. 3 run directly, 3 need a person, 1 needs two confirmations. For example: “Take a room off sale”.
  5. Housekeeping: 9 actions. 4 run directly, 5 need a person. For example: “Mark a room clean”.
  6. Guests: 12 actions. 7 run directly, 3 need a person, 2 need two confirmations. For example: “Remove a guest profile”.
  7. Guest messages: 11 actions. 6 run directly, 5 need a person. For example: “Send a guest a message”.
  8. Prices: 9 actions. 4 run directly, 5 need a person. For example: “Change a room’s price”.
  9. Booking sites: 9 actions. 5 run directly, 3 need a person, 1 needs two confirmations. For example: “Stop selling a date”.
  10. Website: 11 actions. 4 run directly, 6 need a person, 1 needs two confirmations. For example: “Delete a website page”.
  11. Events: 9 actions. 5 run directly, 3 need a person, 1 needs two confirmations. For example: “Cancel an event”.
  12. Billing: 15 actions. 6 run directly, 8 need a person, 1 needs two confirmations. For example: “Void a charge on the bill”.
  13. Accounting: 17 actions. 12 run directly, 5 need a person. For example: “Reverse an accounting entry”.
  14. Finance: 12 actions. 9 run directly, 3 need a person. For example: “Set an exchange rate by hand”.
  15. Restaurant: 9 actions. 6 run directly, 3 need a person. For example: “Take payment for an order”.
  16. Sales: 10 actions. 6 run directly, 4 need a person. For example: “Send a sales proposal”.
  17. Marketing: 12 actions. 6 run directly, 6 need a person. For example: “Send an email campaign”.
  18. Staff: 15 actions. 7 run directly, 8 need a person. For example: “Approve a leave request”.
  19. Stock: 8 actions. 4 run directly, 4 need a person. For example: “Issue stock to a department”.
  20. Settings: 5 actions. 2 run directly, 3 need a person. For example: “Change e-invoice settings”.
03 / 05The ledger

Bounded. Logged. Reversible.

Changes the AI proposes follow three rules.

Where this sits in the ecosystem

  1. Small changes apply automatically and are logged.

  2. Larger ones wait for a person.

  3. Every change can be undone, except a message already sent to a guest.

Builtin our working demo

Illustrative
The hotel’s limit Held at the limit Ledger

Undo the applied change, or approve the one held at the limit.

Applied
Proposed
Undone
Approved

Read the figure
  1. The hotel’s limit: changes inside it apply automatically
  2. A small change: applied, and written to the ledger
  3. A larger change: held at the limit until a person approves it
  4. The ledger: every change, and every undo, is added as a new entry Builtin our working demo
04 / 05Data

Every hotel’s data, kept apart.

Built so each hotel sees only its own records.

Partners
Published rooms and prices
This hotel KVKK built in
  • Consent, asked at booking
  • Export, on request
  • Retention period for guest data
Another hotel
No path between hotels

One operating systemBuilt

Illustrative
Read the figure
  1. This hotel: KVKK built into the data model, with consent at booking, export on request and a retention period for guest data Built
  2. Published rooms and prices: the only part partners reach Built
  3. Another hotel on the same system, built so none of its records show here
  4. One operating system underneath every hotel Built

KBS, KVKK, e-Fatura and e-Arşiv are built in; by default, e-invoices wait for a person’s review.

The hotel system
05 / 05Where data lives

Planned, not claimed.

Moving hosting and AI processing into Türkiye is on our roadmap. Until it is done, we do not claim it.

Today

Hosting and AI processing outside Türkiye

Our architecture decides where data may go, for each hotel’s setup.

Roadmapplanned

Hosting and AI processing in Türkiye

Illustrative
Read the figure
  1. Today: hosting and AI processing outside Türkiye, a solid line that stops at a closed gate
  2. The gate opens when the move is done Roadmap
  3. Hosting and AI processing in Türkiye, drawn dashed Roadmap
Ask about data and hosting Privacy and KVKK