Stop re-explaining your organisation to AI every session

If your team uses an assistant for work, every new chat starts from nothing. Somebody pastes the same background in again, and the answer still reads like it was written for another organisation.

Four files fix that, and you write them once.

→ The mandate. What you do, who you answer to, and the subjects you never comment on publicly. → The tone of voice. Not adjectives — the rules your revisers already apply by hand. → The file you're working on this week, so you stop pasting the background into every chat. → Three texts that were already approved, exactly as they went out.

The second one is where most of the difference is. "Authoritative but accessible" gives a model nothing it can act on. "Spell out every acronym on first use, cite the act by number, never shorten the name of the institution" gives it something.

I put the whole thing on one page: what goes into each file, and what to leave out.

Get it here: https://white-rabb.hubspotpagebuilder.com/four-files

🤖 This video uses an AI-generated avatar and synthetic voice (EU AI Act, Art. 50).

#AItools #comms #Brussels