Five questions before you move on
Answer each one in your head, or on paper, before you reveal the response. This is a self-check, not a test. If two or more answers surprise you, re-read the section they point back to.
1. What are the five categories of AI tool used in comms work?
- A. Chatbots, image generators, schedulers, dashboards, and plugins.
- B. Conversational, content creation, productivity and meeting, intelligence and monitoring, and workflow and automation.
- C. Writing, editing, designing, analysing, and publishing.
- D. Free, paid, enterprise, open-source, and custom tools.
Answer: B. The categories are organised by the job the tool does, not by brand, price, or format. That is what keeps them stable while specific tools change. Most practitioners actively use two or three.
2. A colleague says: "Why learn any of this? The tools will all be different in a year." What is wrong with that?
- A. Nothing; it is fair, so the course should focus only on this year's tools.
- B. The tools will not really change, so the worry is misplaced.
- C. The tools change, but the five categories are stable, so you can swap tools without rebuilding your practice.
- D. Newer tools are always better, so waiting carries no risk.
Answer: C. The tools will be different; the categories will not. Understanding the categories lets you replace a tool without starting over. Only knowing specific tools is what forces a rebuild each time.
3. Name the three failure modes that show up across all five categories. Which one is hardest to notice?
- A. Hallucinations, confidentiality slips, and the composite voice; the composite voice is hardest to notice.
- B. Slow output, high cost, and a steep learning curve; cost is hardest to notice.
- C. Hallucinations, bias, and copyright risk; bias is hardest to notice.
- D. Downtime, data loss, and vendor lock-in; lock-in is hardest to notice.
Answer: A. The composite voice is noticed last because it does not look like an error. It reads as adequate writing while eroding the recognisable, contextually appropriate voice you are paid to produce.
4. Every row of your shortlist ends with a one-line rationale. What must that line connect to before it counts?
- A. The tool's headline features and benchmark scores.
- B. A specific element of your risk profile, not the tool's features.
- C. The price of the tool and its return on investment.
- D. Whether the tool leads its category.
Answer: B. "Best-in-class writing" is a feature. "Configured to retain no inputs, appropriate for a high-scrutiny sector" is a rationale. If a row reads like marketing copy, that is the one to rewrite.
5. It is 2:15 on a Tuesday. Your director wants an AI-drafted press release by next week and you have one free hour. Why is a draft shortlist stronger than opening a chatbot and drafting?
- A. A shortlist simply takes less time than a draft.
- B. The director secretly wants the shortlist, not the release.
- C. The shortlist answers the underlying question, your AI position, so the draft comes easier; drafting first leaves you re-deciding it next week.
- D. You should never draft a press release with AI under time pressure.
Answer: C. The director asked for a press release, but the real question is your AI position. The shortlist hour gives a documented reason for your tool and makes the draft easier. Drafting first produces fast output and leaves the same undecided call for next week.