Practical guide
Prepare an executive issue briefing with AI
Build a fictional reputation issue briefing that separates known facts, approved language, options and the next update trigger.
Corporate communications managers can use AI to compress fast-moving information while preserving fact status and message authority. This fictional exercise produces an executive issue briefing, not a public statement, crisis advice or evidence about a real organization.
Fix the briefing question
An executive briefing should make the decision or awareness need explicit.
The fictional issue is an online claim that a scheduled service change caused an outage. Leaders need to know what is verified, what is unknown and whether to issue a holding message. The exercise provides no confirmed causal link.
Ask AI to draft the question and scope, then remove speculation. Do not repeat the online claim as fact or search for private information about its author.
Build a fact-status table
Known, reported and inferred information should not share one bullet list.
The packet confirms a service interruption and a scheduled change, but timestamps do not establish causation. Record source, time and status for each. Mark the causal claim unverified.
Keep approved internal facts separate from material cleared for public use. AI may group duplicates, but a worker should check every retained statement against the source note.
Draft bounded message options
Message choices should not outrun the approved facts.
Option one acknowledges the interruption and says investigation continues. Option two waits for a confirmed cause while preparing the same factual language. Neither option blames the scheduled change or promises a resolution time.
Label wording as draft until the named communications owner approves it. Do not publish or send from this exercise. If legal or specialist review is required by the fictional process, show it as pending.
Add the next update trigger
Fast-moving briefings become stale unless they say what changes them.
The trigger is a verified incident timeline or an owner-approved cause statement, whichever arrives first. Record the owner and update channel without inventing a deadline that the packet does not provide.
If no new evidence arrives, the executive still needs an explicit choice about whether to use the holding message. Silence should not be converted into approval.
Deliver a one-page issue brief
The final artifact should support a bounded executive choice.
Include issue, decision need, fact-status table, audience implications, two message options, approvals and update trigger. Put unverified claims in a clearly labelled section.
Review claim discipline and message authority. The parent role guide manages coordinated audience updates; this page handles one issue-specific executive decision under uncertain causation.
Test the brief against an unsupported update
A fast update can sound authoritative even when its source is second hand.
Add a fictional message saying a team member heard that the scheduled change caused the interruption. Place it in the reported-not-verified category with its source chain. Do not use it to strengthen the causal claim, include it in approved language or identify an individual beyond the role supplied in the exercise.
Ask AI to draft a revised executive summary, then compare every sentence with the fact-status table. Remove any wording that implies confirmed causation. The unsupported update may justify a verification question, but it does not change the public message options until an accountable source supplies evidence.
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