Overbooking response
- Type
- Demonstration project
- Project
- Hotel demonstration project
- Industry
- Hotel
The scenario and numbers in this demonstration project are simulated data.
- Industry
- Hotel
- What was built
- Agents, skills, semantic models and write tools, built on demonstration data
- Where the knowledge came from
- The group contract, plus demonstration data from the PMS, channel, revenue management, loyalty and housekeeping systems
The gap after the night audit
- The hotel's night audit finds more confirmed bookings for a room type tomorrow than there are rooms.
- At hotels like this, the reservations manager used to start going through the list and phoning guests only after arriving in the morning.
Upgrade and contract checks
- The agent uses semantic models to check free rooms by type, guests' loyalty tiers and room readiness.
- The agent reads the group contract and confirms that held rooms not confirmed by the cut-off can be released.
The plan and its approval
- The agent makes its judgements with these skills: overbooking detection, upgrade matrix, loyalty priority, relocation cost and room readiness.
- The agent ranks affected guests by loyalty tier. Members go on the upgrade list, and a non-member single-night booking becomes the relocation candidate.
- The agent compares the cost of relocating a guest to a partner hotel with the opportunity cost of giving a suite to a non-member.
- The agent first checks that the partner hotel can still take a booking tomorrow, so the relocation does not fail on the night.
- Sindri gathers the upgrade list, the group rooms to release, the relocation arrangement and the guest notices into one decision pack.
- Before upgrades, group block changes, relocations and guest notices are written back, the agent stops.
- Only after the reservations manager approves them together in the morning does the agent write to the PMS and the loyalty system.
Continuous improvement
- At hotels like this, rate plans and group contracts will keep being added and revised.
- The agent will collect feedback on its answers from users. The feedback will go to Odin.
- Wrong answers will become tests. A change will be kept only if scores improve.
Get started
Book a two-hour workshop
AI feasibility review of a live workflow