Five questions per person.
Each staff member receives daily prompts for their role, setting and current stay. Staff answer through ordinary service and observation. Guests do not receive another questionnaire.
GuestMemoryOS is an active workflow throughout the stay. It helps staff discover useful details, record who they concern and use them in the next service. The team supplies the context; AI helps structure the evidence and suggest what to do next.

“Could we have gluten-free bread?” may not say who needs it. The person serving the party can establish whose request it is and record the context.
AI can organise information already present in a message. Staff can supply a missing detail through service. GuestMemoryOS keeps the individual, source and context attached, with observations distinct from confirmed preferences.
Why the source of a memory matters →Start with the guest experience, then follow the service knowledge that makes it possible. This scripted example shows the public workflow.
The video could not load. Download the MP4 or use the full text version below.
At another participating hotel, Alex finds the room set to 21°C, with familiar refreshments ready. The previous team’s recorded preferences helped this team prepare. Illustrative welcome, subject to agreed sharing and availability.
Illustrative service journey. Your team decides and delivers the service.
Download the film · MP4 ↓Technology helps the people delivering hospitality. Five daily prompts focus their attention; a separate QuickNote captures anything else worth remembering.
Each staff member receives daily prompts for their role, setting and current stay. Staff answer through ordinary service and observation. Guests do not receive another questionnaire.
Staff speak or type a QuickNote while the detail is fresh. Additional observations are unlimited. The system asks for useful missing detail in the staff member’s language.
Managers see the combined answers. Departments receive relevant briefs and recommendations. Staff decide how to serve, then record the guest’s response.
AFG reports an average of six seconds per QuickNote: about two minutes of entry for 20 notes. This is note-entry time, not the total daily workload. Read the staff adoption evidence →
A staff member can submit a note without opening the full guest profile. The manager sees the combined answers and has separate guest-experience prompts. Staff receive the information relevant to their role.
Capture permission and profile access are separate. Ambiguous identities, conflicting observations and sensitive needs need the appropriate review. See memory governance →
Food and beverage, the chef, housekeeping and guest services receive guidance relevant to their work. Staff act on useful suggestions, then record the response to help improve what happens next.
Illustrative service examples. Your people review recommendations and deliver the service.
Staff record an early departure and the guest’s preference for black coffee before breakfast.
Offer an earlier breakfast, with the preferred coffee ready before the guest leaves.
The team records whether the timing worked. That feedback helps shape the next morning’s service.
After dinner, the guest tells a team member that the meal felt too rich. Staff record that feedback alongside what the guest enjoyed.
Suggest a lighter preparation for the next meal, using the guest’s recorded likes and meal feedback.
The chef reviews the suggestion and adapts the meal. Feedback after that meal helps refine the next recommendation.
Staff record the scent used most, towel preferences, air-conditioning settings, preferred water and turndown times.
Bring those details into the housekeeping brief so the team can prepare the room around the guest’s known preferences.
Staff check changed requests and record what worked. The next shift receives the current room preferences.
The guest tells staff they enjoyed a coastal walk and would like a quiet evening. The team records those comments.
Suggest a relaxed evening and ideas for the next day that fit what the guest said they enjoyed.
The team records what the guest chooses and how it went, giving the next recommendation a better starting point.
GuestMemoryOS calculates an estimated likelihood of rebooking from guest comments and feedback recorded by staff. Alongside concerns and service responses, it helps the manager see how the stay is going and where attention is needed.
The questions operators ask before a demonstration.
Yes. Available booking and guest-shared context helps the team prepare. What staff learn from day one can improve another department’s service during that same visit.
No. Each employee receives five prompts tailored to their role and the current stay, to answer through ordinary service and observation. Separate QuickNotes capture unlimited additional discoveries.
Their recorded guest knowledge remains available. The colleague covering receives the relevant brief, so service need not depend on the original employee being present.
Relevant everyday preferences help the next team prepare, including at another participating property or ship within the agreed sharing scope. Current wishes guide service and sensitive needs are reconfirmed.
It brings together evidence while there is still time to act. Guest statements, staff observations and estimates stay distinct. A predicted rebooking percentage is not an actual booking or a guest’s own rating.
No. It can operate standalone or connect through an agreed integration scope. Existing systems retain their booking, billing and customer-relationship responsibilities.
See the workflow used in live residence stays, then explore how it could support your teams and guests. AFG is seeking a cruise line or large hotel trial partner.
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 →