GuestMemoryOS grew from a standing start through months of daily operation at Mandarin Beach Villa, Koh Samui. Stays of 5–18 nights gave the team time to learn, act, adapt and retain useful memories. That live experience has shaped the product in use today.
Explore AFG’s reported 40-family trial setting, Mandarin Beach Villa’s public reviews and what person-specific memory could mean for five cruise operating models.
Company-reported family cohort · Public review context · Proposed cruise evaluations
Read the multi-generational guest identity paper →A live luxury residence at £1,350 per night, including Airbnb fees, with real service teams and a public guest review record. Airbnb lists 40 reviews, an overall rating of 5.0 out of 5, and every review in the five-star category.
The residence’s public review record. Checked on Airbnb on 21 September 2026. Read the Airbnb reviews . The reviews reflect the complete guest experience at the residence where GuestMemoryOS is used.
AFG reports the development history, live use, 5–18-night stays and nightly price including Airbnb fees. The review record is evidence of guest experiences at the residence; it does not isolate the effect of the software from the team, property and wider service.
The trial tested whether the team would adopt GuestMemoryOS and continue using it after the training manager left. AFG reports that staff kept contributing, with additional observations increasing after the trainer’s departure.
The setting was Mandarin Beach Villa, Koh Samui, an Airbnb Luxe villa priced at £1,350 per night including Airbnb fees in AFG’s trial account. AFG reports that the residence began the trial with no previous Airbnb reviews. All six trial staff were Thai nationals; five did not speak English and one spoke both Thai and English. Staff used the system in their own language.
AFG initially allowed two weeks for training. The reported training period was three days, estimated at 2.2 hours per day. The trainer brought senior sales-training experience from the UK telecoms sector. One-button QuickNote entry and native-language use were designed to keep the staff workflow simple.
The trainer left on 22 July 2026. As of 21 September, AFG reports no further trainer contact, with note volumes and quality maintained without further trainer input. AFG also reports an increase in recorded observations and facts after the trainer left: 11–18 additional observations per staff member per day, alongside the five assigned daily questions.
In AFG’s account of a follow-up visit, staff said the system made their jobs easier and less of a struggle. They also reported substantially higher tips at the end of stays. This is staff-reported feedback; the trial account does not quantify the change in tips or isolate its cause.
When a note lacks useful detail, QuickNote pauses the entry, explains why in the staff member’s own language and identifies what is missing. Staff can improve the note at that moment. The system also monitors staff contributions, giving the manager a view of participation.
What the 86–97% scores measure. The system’s internal staff contribution score combines completion of the five mandatory daily questions, the number of notes recorded each day, and the quality and relevance of those notes to the staff member’s role. AFG reports a lowest score of 86% and a highest score of 97% across the trial team. These are staff contribution scores, not guest satisfaction ratings or a measure of prediction accuracy.
Source: AFG’s account of the live residence staff trial, supplied 21 September 2026. Training and contribution figures are company-reported and have not been independently audited. These results describe this team and setting; they are distinct from the dated activity totals below and from future hotel or cruise trials.
The operating figures below cover activity logged from 14 July to 21 September 2026, as supplied by AFG. This is the measurement window for the activity snapshot. The 40 Airbnb reviews above are a separate, property-level total checked on 21 September 2026.
Source: AFG-supplied research readout, 21 September 2026. Company-reported activity, not independently audited. Categories may overlap and must not be added together. An opportunity, a recorded action and a recorded response are different measures; a response is not necessarily a verified outcome.
Read how AFG’s reported experience of repeated failures during live development shaped the GuestMemoryOS approach—and why identity, context, corrections and service follow-through matter.
Why AI alone is not guest memory →Learning happened during service: before arrival, through the first day, across a longer stay and into the next visit. The product developed alongside the people using it.
Frontline observations become individual guest knowledge. Departments receive relevant context and suggestions while there is still time to improve the experience.
Service actions and guest responses inform the next decision. Months of real operation have helped the team refine how the product supports everyday service.
Useful preferences are retained for handovers and return visits. Sensitive or changed requirements are checked again when needed.
Extensive internal cruise simulation testing has been completed using room-by-room virtual cruise environments populated with simulated guests.
Virtual cruise environments give the service model a much larger setting than the live residence.
Simulated guest activity exercises the system across the reconstructed environment.
The full simulation testing record is available privately to verified cruise-line representatives. The next step is a focused workflow trial with real teams and guests.
Requests must come from a bona fide cruise-line company email address. AFG verifies the organisation and mailbox before issuing private access details.
AFG has also completed a simulation of a 600-room hotel in Dubai. This extends the internal simulation work beyond cruise environments. A live larger-hotel trial with an operator’s real teams and guests is the next step.
Company-reported simulation work, separate from the live Mandarin Beach Villa deployment. This is not a claim of a live 600-room hotel installation.
Original analysis of documented airline service models, connected to AFG’s guest-memory evidence and operating methods.
We are seeking a cruise line or large hotel for a live trial. Start with the product already used at the residence, examine the simulation work, and agree a focused deployment within your operation.
Follow a guest detail into a service suggestion, a delivered action, a response and retained knowledge. Examine the live residence experience and relevant test scenarios.
Agree the teams, guests, workflow and success measures together: useful suggestions, staff effort, service delivery, guest response and continuity into the next visit.