Communicating without undermining people

How can you use AI to craft a sensitive, factual message tailored to the people concerned?

1960 € per session, excl. VAT

In short

This 7-hour day, excluding breaks, enables participants to build a sensitive communications pack using an approved generative AI tool. From establishing the factual framework to simulating difficult questions, participants structure their messages, check wording and organise human approval before any release.

Why this training?

A sensitive announcement can lose credibility through premature certainty, a commitment that cannot be honoured or discrepancies between audiences. Distinguishing established facts, unknowns and decisions prevents the pursuit of a reassuring tone from obscuring essential information. AI helps compare wording and anticipate objections, but it can also invent an explanation or downplay a risk. The training provides safeguards for this use through identified sources, explicit prohibitions and a critical review that challenges the content. Continuing case study: The fictitious cooperative Ormevia Services temporarily closes a local reception point after safety defects are discovered. The duration remains unknown. Employees, service users and partners expect explanations and service continuity arrangements, while a rumour of permanent closure is circulating. Transfer to the workplace: For an upcoming announcement, reuse the facts matrix and approval process before drafting. After release, compare the questions received with the objections anticipated and document misunderstandings to revise the FAQ.

Objectives of this training

  • Classify information into verified facts, assumptions and unknowns.
  • Map expectations and risks of misinterpretation for each audience.
  • Draft a main message and two consistent adaptations using AI.
  • Develop answers to difficult questions without speculation.
  • Assess the pack and organise its approval before release.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Internal or external communications managers responsible for sensitive announcements.
  • Managers and HR leads preparing to address issues with significant human implications.
  • Service managers coordinating messages, responses and approvals.

Prior skills

  • Be able to formulate and refine a request in a generative AI assistant.
  • Know the audiences and usual approval processes for your communications.

What you will work on

  1. Define what can be announced

    Objective: Distinguish available information from information that may be communicated. Sources, confidentiality, uncertainties and the limits of AI assistance.

  2. Analyse the situation for each audience

    Objective: Prioritise information needs. Practical impacts, legitimate expectations, rumours and information asymmetries.

  3. Develop the reference message

    Objective: Write a clear announcement without false reassurance. Facts, acknowledgement of impacts, immediate measures, unknowns and the next update.

  4. Adapt without contradictions

    Objective: Adapt the announcement for different uses. Briefing note for managers, message for service users, consistency of commitments and accessible vocabulary.

  5. Face difficult questions

    Objective: Respond accurately under pressure. Empathy, limits on answers, refusal to speculate and escalation to the appropriate person in charge.

  6. Check and organise the release

    Objective: Submit a verifiable pack for approval. Review by subject specialists, HR or legal staff depending on the issue; timeline and monitoring of misunderstandings.

What you take away

  • Matrix of facts, unknowns and information that may be communicated.
  • Messaging pack: main announcement, briefing note for managers and message for service users.
  • FAQ on sensitive issues with answers ready for approval and escalation points.
  • Checklist and release plan submitted for approval.

Method and assessment

  • Short teaching inputs followed by immediate practice using the continuing case study.
  • Individual work followed by work in pairs, with peer review of the outputs.
  • Guided group discussion: what works, what needs correcting and what remains a human decision.
  • Preparation: Check access to a tool approved by the organisation and the applicable confidentiality rules. Prepare a fictitious or anonymised example of a sensitive announcement, without personal data or confidential information.

Observable criteria: Every factual statement refers to a source in the pack. ; Unknowns are explicit and no unauthorised commitments appear. ; All three messages retain the same facts and specify the actions available. ; The FAQ addresses priority objections without minimisation or speculation. ; Each piece of content has an approver, a channel and a position in the release sequence..

Frequently asked questions

Can a real situation be used?

Only after anonymisation and authorisation; the exercises can remain entirely fictitious.

Can AI approve the message?

No. It assists with drafting; authorised decision-makers retain responsibility for approval.

What should you do if a fact remains uncertain?

Flag the unknown, identify who needs to confirm it and avoid any premature conclusions.

How can quality be measured before release?

Check sources, contradictions, commitments and understanding during the simulation.

Last updated: