Excerpt from the practical guide
What should stay out of the AI?
Three questions and a placeholder trick for sensitive data
A difficult request is on your desk, and the reply needs to be empathetic. This is exactly where AI is a strong support: it helps you find the warm yet clear tone you sometimes spend a long time searching for. Except that the raw text contains a name, an address, a few very personal details. Now what?
You don’t have to choose between AI support and confidentiality. A routine of a few seconds combines both – without interrupting your flow.
Three short questions before sending and a simple placeholder trick are all it takes. With them you can work on sensitive topics with AI without real data ever leaving your desk. The hesitant question “Am I allowed to enter this?” becomes a calm answer: “This is how I enter it.”
Cautious, not anxious
Data protection with AI sounds at first like a rulebook you would have to work through. In daily practice, however, it comes down to one single, very manageable principle: data minimisation. You give the AI exactly as much information as the task needs – and not one name more.
It is the same professional attitude you have long mastered. When you ask an external expert for advice, you describe the case – you don’t hand over the whole file. Nobody would call that anxious. It is simply cautious. You bring exactly this caution to your work with AI: you keep working, just smarter.
The payoff shows quickly. If you know how to keep sensitive details out, you trust yourself with more – precisely the important, personal tasks where good wording counts most. Caution is not a brake, then, but the entry ticket to exactly these topics.
Method · Caution reflex
Three calm questions for your own text, right before sending:
- 1
Personal? Names, contact details, information about individual people?
- 2
Confidential? Something that should stay internal – contracts, internal figures, agreements?
- 3
Can it be anonymised? Do I need the real detail – or will a placeholder do?
Placeholder step: replace sensitive spots with [NAME], [PLACE], [AMOUNT]. You put the real details back into the finished result yourself at the end.
Prompt template
For a sensitive letter you want drafted in anonymised form – you put the real details in yourself at the end. The template works unchanged in any AI assistant.
Help me word an empathetic letter of reply. The situation: [MATTER IN YOUR OWN WORDS, WITHOUT REAL NAMES]. The letter goes to [NAME], living in [PLACE OF RESIDENCE] – leave these placeholders exactly as they are in the draft; I will fill in the real details myself afterwards. Tone: warm, respectful, clear.
Key takeaway
Caution is a reflex, not an obstacle: three questions, a few placeholders – and you can work on sensitive topics with AI without giving real data out of your hands.
How this guide continues
The full guide takes you further
In the full practical guide, this excerpt is followed by a review box for double-checking the AI's result, follow-up questions that take you deeper, and a hands-on exercise with your own case. And in the portal, one click adds the template to your AI Workbench – so your own toolbox grows guide by guide.
After the check, the practical guide matching your skill level arrives complete and free by email.