Philippines staffing research · Published October 22, 2026

A Fidelity Test for Meeting Notes and Action Items

Measure whether a note preserves the decision, the owner, the date, and the open question without adding invented certainty.

On this page

Methodology

Bounded method review with 4 authoritative sources, 6 analysis sections, and 0 service performance claims.

Key Stats

4: authoritative sources
6: analysis sections
0: service performance claims

Key Takeaways

Define the reference record
Score four elements
Test attribution and chronology

Define the reference record

Before the meeting, agree what the reference will be: a recording, a written agenda with owner notes, or a post-meeting confirmation from the chair. The reference should be authoritative enough to settle a disagreement. If no reference exists, the test can still check internal consistency, but it cannot measure fidelity to the discussion. Record the meeting type, the sensitivity, the attendees, and the note version. Remove or restrict personal and protected content before review. The test should not create a second uncontrolled copy of a sensitive meeting.

Score four elements

For each decision, check whether the note states the decision, the rationale, the owner, and the effective date. For each action, check the specific output, the owner, the due date, and the dependency. For each open question, check that the note says it is unresolved and names the next step. Score omissions and additions separately. An omission can leave a team without a required action. An addition can create a commitment that was never made. Both matter. A note that is concise but invents a decision is less faithful than a longer note that labels uncertainty.

Test attribution and chronology

Check that the note attributes statements to the right person or to the group. A comment from a customer advisor is not a decision by the executive. Check the order of events when the meeting changed direction. A note that presents the final decision without the earlier constraint can hide why the choice was made. Look for ambiguous pronouns and passive voice that hide the owner. Rewrite examples such as it was agreed or the team will follow up into a named owner and a specific action. The reviewer should be able to act on the note without asking who meant what.

Reporting

Keep the reference record, the note, the scoring sheet, and the owner's confirmation. Report omissions and additions separately, and show the meeting type and sensitivity beside the result. A note that is accurate for a routine stand-up may not be sufficient for a board meeting with protected content. Share the summary with the note owner before using it in a performance review. The owner may have context that changes the interpretation. The purpose is a better decision record, not a score that punishes the assistant for a meeting that lacked a clear close.

Limitations

The reference record may itself be incomplete or biased. A recording captures words but not side conversations or prior context. The note may be accurate for the declared purpose but insufficient for a different audience. The test measures fidelity to the stated standard, not universal completeness. Sensitive meetings may not be eligible for recording or review. In those cases, use a chair confirmation and a shorter note. Do not trade privacy for a more precise score.

Decision use

Use the result to improve the template, the meeting owner's instructions, or the review step. If decisions are consistently missed, the note template may not prompt for them. If actions lack owners, the meeting may be ending without a clear close. Retest after a template change or a new meeting type. Keep the reference record and the reviewed note under the approved retention rule. The goal is a decision record that the team can trust, not a perfect transcript. Include the meeting owner in the review, because a note can be faithful to the words and still miss the decision the chair intended. That conversation is part of the standard, not an exception to it.

Sources checked October 22, 2026. Method protocol only; no provider performance claim.

Sources

Google Site Reliability Engineering Workbook, Postmortem Culture: Public guidance supporting the protocol.
W3C Web Content Accessibility Guidelines 2.2: Public guidance supporting the protocol.
U.S. GAO, Assessing Data Reliability: Public guidance supporting the protocol.
NIST Privacy Framework: Public guidance supporting the protocol.

FAQs

Plan the next step

Translate the method into a bounded role brief with named review points.

Review virtual assistant services

Read the daily blog guides · Explore service workflows · Plan your staffing routine

Use the method to define the population, evidence, and reviewer before testing.