Practical guide
Prepare an executive decision pre-read with AI
Turn fictional owner updates into a pre-read that states the decision, options, missing inputs and meeting use.
Executive assistants can use AI to assemble a decision pre-read while checking versions, ownership and unresolved inputs. This fictional exercise produces a meeting artifact, not an executive decision, delegated approval or proof that the chosen option will work.
Name the meeting decision
A pre-read is useful when readers know what they must decide.
The fictional meeting must choose whether to move a workshop by one week, keep the date with reduced scope or cancel it. Record the decision owner and constraints. Do not describe the meeting as approval already obtained.
Ask AI to turn the request into one decision sentence and three bounded options. Verify that none adds spending, staffing or external communication outside the packet.
Reconcile owner inputs by version
A newer document is not automatically the approved source.
The venue owner confirms both dates in version 2; the facilitator says only the original date in an older email; finance has not confirmed the change fee. Record source date, version and owner beside each input.
Do not silently use version 2 to overwrite the facilitator’s constraint. Ask whether it changes feasibility and preserve the missing fee as unresolved rather than zero.
Build comparable options
Each option should use the same decision dimensions.
Compare participant availability, facilitator availability, venue status, scope effect and unresolved cost. The reduced-scope option fits the known facilitator date but needs content-owner approval.
Use concise tables, but do not assign invented scores. If a field is unknown, show the owner and the latest useful response time instead of a plausible value.
Plan how the meeting uses the pre-read
The artifact should reduce meeting work without pre-deciding the choice.
Ask readers to correct factual inputs before the meeting and reserve discussion for the unresolved tradeoff. The exercise suggests five minutes for corrections and fifteen for decision discussion, clearly labelled as proposed allocation.
Do not send calendar invitations or messages. If a critical input remains missing, include defer as a valid meeting outcome rather than forcing a selection.
Deliver the controlled pre-read
The final should make version and authority visible.
Include decision, options, constraints, owner inputs, missing evidence, proposed meeting use, document version and approval status. Leave the selected option blank until the executive supplies it.
Review version fidelity and meeting boundary. The parent guide organizes agenda flow; this resource builds the decision-specific evidence packet that the agenda points to.
Run a last-minute version check
A pre-read can become misleading when an owner updates one dependency after distribution.
Introduce a fictional facilitator update that the reduced-scope option is no longer available. Mark the old pre-read superseded, create a new version and remove that option from the current comparison while preserving the reason in the change history. Do not silently edit a distributed artifact or claim that all readers saw the revision.
Ask AI for a short change note naming the affected option, unchanged facts and still-missing fee. Verify it against both versions. The meeting owner can choose to continue with two options or defer; the executive assistant controls the artifact, not the decision.
Sources and scope
NIST: AI risk management. Skills England: workplace AI foundations.
These references provide background, not validation or endorsement of this exercise. The case details, calculations and suggested review questions are original instructional material. Use them to discuss observable work, not to infer customer outcomes, professional credentials or performance in every setting. Before adapting the exercise, confirm the relevant facts, approved tools, data permissions and decision owners. If you change the case, revisit the expected answers and checks as well. These examples describe practice tasks, not a promise that a particular product includes the fictional features.
Sources: [1] [2]
Sources
- 1.AI RMF Core · NIST
- 2.AI foundation skills for work benchmark · Skills England