# Fictional source packet: a launch decision across meetings

This is an authored teaching example, not customer data, an OpenMax run, or a transcription accuracy test. It contains five short permitted excerpts, not full recordings. Names, authority assignments, timestamps, and correction records are fictional. There is no original audio available for independent verification. Do not upload real meeting content merely to reproduce the example.

## Corpus and authority

Eight submitted files refer to six meeting IDs. F02 duplicates F01. F04 is an obsolete transcript version replaced by F05. M06 is outside the analysis permission scope; only its inventory metadata is present. The selected corpus consists of F01, F03, F05, F06, and F07: five meetings.

For this fictional project, the chair named Lee can approve the launch target and assign the test-plan task. Acceptance by the assigned person is recorded separately. Rina and Omar are identified within their named excerpts. The anonymous label Speaker 2 in M04 has no verified mapping to either person. Meeting order does not by itself establish decision authority.

## M01 / v1 / F01: a proposal

Event time: 2026-08-25T09:00:00+09:00. UTC: 2026-08-25T00:00:00Z.

- M01-L01 — Rina: “Could we aim for September 14?”
- M01-L02 — Lee: “That is a proposal. We have not approved the scope or the launch date.”

F02 is an exact duplicate export of this fictional file. It adds no meeting and no independent corroboration. Theme: launch_date.

## M02 / v1 / F03: a conditional offer

Event time: 2026-08-24T18:00:00-07:00. UTC: 2026-08-25T01:00:00Z.

- M02-L01 — Rina: “If scope is approved, I can prepare the test plan.”
- M02-L02 — Lee: “The launch date is still undecided. We will confirm the scope and owner at the next review.”

This event is one hour after M01 despite its earlier local calendar date. Rina has made a conditional offer, not accepted an assigned task. Themes: launch_date, test_plan.

## M03 / v2 / F05: approval and acceptance

Event time: 2026-08-26T09:00:00+01:00. UTC: 2026-08-26T08:00:00Z.

- M03-L01 — Lee: “Scope is approved. I approve September 21 as the launch target.”
- M03-L02 — Lee: “Omar, please prepare the test plan by August 28.”
- M03-L03 — Omar: “I accept the test-plan task, due August 28.”

Correction record COR-01, entered 2026-08-26T10:00:00Z: the fixture defines M03 v1 line L01 as an erroneous transcription of the target date, “September 12.” The authorized corrected text is M03 v2 L01, “September 21.” F04 contains that obsolete v1 text and is excluded from the selected corpus. This authored correction record establishes the intended case answer; it is not evidence that anyone verified a real recording. The meeting event time is unchanged. The correction is not a second business decision moving a previously approved launch.

Themes: launch_date, test_plan.

## M04 / v1 / F06: a later proposal is declined

Event time: 2026-08-27T09:00:00Z. UTC: 2026-08-27T09:00:00Z.

- M04-L01 — Speaker 2: “Can we return to September 14?”
- M04-L02 — Lee: “That change is not approved. September 21 remains the approved target.”

No evidence identifies Speaker 2 as Rina. Theme: launch_date.

## M05 / v1 / F07: draft readiness is not completed execution

Event time: 2026-08-28T09:00:00Z. UTC: 2026-08-28T09:00:00Z.

- M05-L01 — Omar: “The test-plan draft is ready for review.”
- M05-L02 — QA reviewer: “We have not accepted the plan yet, and testing has not been executed.”

This supports a reported draft-ready state. It does not establish acceptance, independent delivery verification, or completed testing. Themes: test_plan, test_execution. No launch-date discussion is present in this excerpt.

## M06: inventory metadata only

Meeting ID M06; file F08; event time 2026-08-28T12:00:00Z; access status restricted; selection status excluded_access. No excerpt, speaker list, topic, or substantive meeting conclusion is supplied. Its absence must remain visible in the coverage statement.

## Review key and counting rules

Selected files: F01, F03, F05, F06, F07. Distinct submitted meeting IDs: M01–M06. Coverage: 5 / 6 = approximately 83.3%, relative to this inventory only. This does not measure all meetings that may have existed in a real organization.

The launch_date theme appears in M01, M02, M03, and M04: 4 / 5 = 80% of included meetings. It does not mean that 80% of people favor a date. Duplicates and obsolete transcript versions never enter the denominator again.

Eight deliberately authored candidate claims are supplied in the companion CSV. C02, C05, and C08 are supported as written; C01, C03, C04, C06, and C07 are not. The resulting 3 / 8 = 37.5% describes this constructed challenge set, not an AI system's accuracy, a customer outcome, or comparative product performance.