What makes a guest preference trustworthy?
A trustworthy guest-memory record identifies the person, preserves the source, states when the information applies and shows whether it remains suitable for the proposed service decision. Confidence without that context is difficult for another employee to evaluate. An accurately transcribed request can still be misused if it concerns the wrong person or the wrong occasion.
This paper proposes an operator review model. Its fields and states are evaluation criteria, not a disclosure of GuestMemoryOS’s internal algorithms or a promise that every deployment exposes an identical screen. Start with the existing explanation of observations and confirmation for the product context. [1]
A reviewable memory needs seven answers
W3C’s PROV overview describes provenance in terms of the entities, activities and people involved in producing information. It supports inspecting origin and history; it cannot establish that a recorded observation is true. [2] AFG proposes the following hospitality review questions.
- Who does it concern?
- The individual, or an explicitly identified party arrangement. “Room 204 likes early breakfast” is ambiguous when several people share the room.
- What was actually learned?
- The attributed request or observation, with interpretation distinguishable from the original account. A polished summary should not quietly add certainty.
- Where did it come from?
- The reporting role, interaction and time, accessible to an authorised reviewer. Receiving staff may need only a limited service summary.
- When does it apply?
- This service, this stay or a possible future preference. A late arrival today does not establish a permanent dining habit.
- What is its current status?
- Usable, awaiting clarification, superseded, under review or no longer eligible for the relevant use.
- Who may use it, and for what?
- The agreed service purpose and receiving roles or locations. Availability in a database is not sufficient justification for every possible use.
- What changed afterwards?
- Corrections, reconfirmations and relevant results, with enough history to explain the current decision under the applicable retention policy.
A practical lifecycle for review
| State | Meaning | Service treatment |
|---|---|---|
| Captured | A report exists, with attribution or scope still to assess. | Clarify missing context before relying on it. |
| Current for this stay | The request applies to an identified visit or service window. | Use within that scope; review at departure. |
| Retained for review | The information may help a future visit. | Check relevance, permission and changed circumstances. |
| Confirmation needed | Uncertainty or a conflict affects the proposed use. | Ask an appropriate question; avoid treating uncertainty as certainty. |
| Superseded | New information changes the working understanding. | Update dependent briefs and pending actions. |
| Expired or removed | The record is no longer eligible for that use. | Stop recall and check downstream copies under the agreed policy. |
These states are not a one-way ladder. A retained preference may return to clarification. A new request may be limited to one evening rather than superseding an enduring preference. Review the relationship between the two statements before changing the record.
Correction must reach the service workflow
In a fictional example, Alex previously requested sparkling water. Before today’s dinner, Alex asks for still water. The receiving team needs today’s instruction before preparation. A reviewer can then establish whether this is a one-meal choice or a broader change. Merely appending the new note leaves the kitchen at risk of seeing the older brief.
An operator test should therefore inspect the record, the current summary and any pending task affected by the correction. Confirm which version the receiving employee saw. Where a task has already been completed, record that fact and arrange an appropriate response rather than rewriting the history.
The UK Information Commissioner’s accuracy guidance distinguishes opinions from facts and emphasises recording sources and relevant changes. It is UK data-protection guidance, cited here for that distinction; the requirements applying to a particular deployment depend on its jurisdiction and processing. [3]
How to test memory quality before wider use
Use fictional records containing a direct request, a second-hand report, a temporary arrangement, a conflicting update and two people with similar names. Ask a reviewer to explain the permitted next action from the record alone. Include a case where the correct action is to seek clarification.
Check that a translation retains the request’s limits. “Today” must not disappear; “might prefer” must not become “always prefers”. Have a competent speaker review the meaning where language affects the decision. Capture the error type so a fluent but inaccurate sentence does not pass on presentation alone.
Measure whether a correction reaches the responsible team before the next decision using the correction measure. Use the handover protocol to test what the receiving colleague actually understands. Review retention and access with the operator through the governance process.
A guest-memory system earns trust when the team can explain both why a memory is useful and when it should stop guiding service.
Sources and evidence notes
- [1] AFG Holdings. Guest memory: sources, observations and confirmation.
- Existing AFG explanation of the product proposition; related reading, not independent corroboration. Source reviewed 25 September 2026.
- [2] W3C (2013). PROV-Overview: An Overview of the PROV Family of Documents.
- Working Group Note, 30 April 2013; abstract and introduction. General provenance concepts. Source reviewed 25 September 2026.
- [3] Information Commissioner’s Office. Principle (d): Accuracy.
- Official UK guidance; sections on opinions, sources and keeping information current. Jurisdiction-specific guidance. Source reviewed 25 September 2026.
Cite this paper
AFG Holdings Ltd. (2026, September 25). The Guest Memory Lifecycle: Sources, Corrections and Useful Recall. AFG Guest Memory Reference Series, paper 2, version 1.0. Permanent article URL
Use the relevant section link when citing a specific definition or measure. Read the editorial and correction policy.
Extend this evidence and method.
Original analysis of documented airline service models, connected to AFG’s guest-memory evidence and operating methods.
AI needs a method.
Guest memory needs evidence.
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 →