Practical guide
AI fluency for executive assistants: prepare a decision agenda
Turn fictional meeting requests into a feasible agenda with decision owners, pre-read dependencies and time checks.
An executive assistant can use AI to prepare a meeting agenda that protects decision time and exposes missing prerequisites. This original exercise uses a fictional forty-minute leadership meeting. Its core artifact is a feasible decision agenda, not a summary of meeting notes or a generic list of administrative AI uses.
Separate decisions from updates
Identify what participants must resolve together before allocating time.
The fictional requests are: approve a pilot scope, hear a ten-minute project update, choose an owner for a customer handoff and discuss next quarter's priorities. The pilot sponsor has supplied a two-page brief, but the priorities request has no decision question. Do not give each request equal space simply because it appears in the inbox.
Ask the requester what outcome they need when the wording is vague. An update may be suitable for a pre-read, while an unresolved owner needs discussion. These are suggested choices for this case, not a universal rule against verbal updates or exploratory conversations.
Check the prerequisites for each decision
An agenda item is not ready merely because it has a title.
For the pilot decision, identify the approver, scope brief and unresolved resource question. For the customer handoff, identify the teams whose ownership needs agreement. If the required person cannot attend, state whether the item can produce a recommendation or must wait for the actual decision.
AI can organize dependencies, but it should not assign authority from seniority or invent an alternate approver. Keep the difference between a proposed owner and a confirmed owner explicit. This helps the chair understand what the meeting can realistically accomplish.
Build a schedule that adds up
Use arithmetic to test the agenda before asking anyone to accept it.
For this practice meeting, allocate five minutes to opening and confirmation, fifteen to the pilot decision, ten to handoff ownership and five to recording actions. That totals thirty-five minutes and leaves five minutes within the forty-minute window. These are illustrative allocations, not recommended durations for all meetings.
If the ten-minute update remains a spoken item, the plan becomes forty-five minutes. Resolve that conflict by changing scope, format or duration with the chair. Do not ask AI to make the agenda look compact while keeping an impossible set of time commitments.
Prepare the chair's prompts
Each decision item should have a question and a place to record the outcome.
For the pilot, write the requested decision, options and remaining condition. For the handoff, write the ownership question and the interfaces the selected owner must coordinate. The chair should not have to infer the purpose of an item from a broad label such as alignment or discussion.
Keep a separate parking area for questions outside the meeting's scope. Moving an issue there does not resolve it; assign a follow-up owner when the meeting provides one. The assistant can draft prompts, but the executive assistant checks them against the request and pre-read.
Respond to a late change
Test whether the agenda preserves its purpose when a dependency changes.
Introduce a fictional update that the pilot approver will arrive fifteen minutes late. Reorder or reframe the agenda without pretending another attendee can authorize the pilot. Recalculate the time plan and make the changed decision status clear to the chair.
The final handoff includes the agenda, linked pre-reads, attendance dependencies and a decision-record template. Review feasibility, authority handling and responsiveness to change. This exercise does not establish broader executive support performance; it supplies concrete evidence about preparing one bounded meeting that can produce usable decisions.
Sources and scope
The references below are background reading, not evidence that this fictional exercise has been validated or endorsed.
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