A Tuesday afternoon
It is 2:15 on a Tuesday afternoon. Your director has sent you a Slack message:
Heard from comms peers that everyone is using AI for press releases now. Can we get on this? Pick a tool, get started, and let's see something by next week.
You have a free hour this afternoon and a half-day later in the week. You know your organisation has no AI policy. You know your IT team is friendly but slow. You know your sector carries real exposure if AI involvement is not handled carefully. What is your first move? Open each option for the reasoning.
Plausible, and many teams default to it. It produces a draft fast. It also leaves you with no documented reason for choosing ChatGPT over Claude, Gemini, Copilot, or any of the other options. Without a documented reason, you choose the same tool again next week with the same lack of clarity, and again the week after. The output is the easy part. The reasoning is the part you need for your own practice to feel sturdy.
Two or three tools you would actually use, a one-line risk-profile rationale for each, one tool you would explicitly not use, and a half-page note on what kind of work you would use AI for and where you would not. This is the move the course trains. The director asked for a press release, but the underlying question is "what is my AI position." The free hour produces both. The draft becomes easier because you have already decided where AI fits and where it does not. You are not making that call mid-paragraph.
Cautious from a risk-minimisation angle, but it carries no analysis your director can act on. It forecloses the work without doing the work to know whether it should be foreclosed. A request for a policy without a draft of what you are proposing is also slower.
Reasonable instinct, but you are handing comms judgement to a function whose role is to evaluate tools against IT criteria, not against the work AI will be used for. The shortlist exercise puts the comms judgement first. Any later conversation with IT can build on it from there.