Email drafts that learn my voice

Request: „Write a reply to the team — as a draft.“

1. The AI reads the mail thread and the relevant notes in the Business Brain
2. It writes to fixed style rules: informal address, prose, short sentences, result first, font and format as usual
3. The draft lands in the mail client, not in the outbox
4. After sending, the AI compares draft and sent version and adopts the differences as new style rules

In any organisation, AI-written text sounds like AI: generic, too long, too preachy — the rework eats the time saved.

Checking the content, adjusting, sending it myself.

A feedback mail to a colleague was too preachy in draft — deadlines, lists, side criticism. The differences to the sent version now sit in the brain as rules for feedback mails: appreciative, set the direction, leave the how to the colleague.

Anyone with a lot of correspondence. A personal voice is learnable, from the delta between proposal and reality.

Most AI-written text sounds like AI. Here the Business Brain learns my own voice from practice: after every send, the AI compares the draft with the mail actually sent, and every difference becomes a rule. A feedback mail that came out too preachy turned into concrete guidelines — appreciative, set the direction, leave the how to the colleague. A human always sends; the draft lands in the mail client, never the outbox. The learning comes from comparing with the real result; more prompting does not deliver it.