Two chats, one question, the same AI. You judge the result.
Open a new chat. Don't add your file. Paste Question 1 below. Copy the answer somewhere you can see it.
Open a second new chat. Attach your Align file, or paste its text, with this line:
This file is my own words about my own work, including phrasings I rejected. Draft in this voice. Do not smooth it. If something is not in the file, ask me rather than invent it.
Then paste the same Question 1.
Put the two answers side by side. Read them out loud. Ask yourself: which one could I send as it is? Which one would I have to rewrite?
Don't regenerate, and don't edit. Asking again until you like the result turns a test into a search.
A chat remembers everything earlier in it. For the next question, open two new chats and start again from Step 1.
Write a short reply to a client who has asked what makes my work different.
Write the opening paragraph of a proposal for a client who has asked me to scope an engagement.
Write the opening paragraph of a note to my board before next week's meeting, asking for their view on a strategic decision I have to make.
Write an email to a long-standing client turning down a request that falls outside what I do, and suggest what I'd do instead.
Write a one-page note to my team explaining a decision they won't like: what I decided, why, what it costs us, and what I need from them.
Or use your own. Anything you'd normally write will work.
Delete the test chats. In most AI apps: find the chat in the sidebar, open its menu (the three dots), choose Delete.
If you can hear yourself in the second answer, you've seen it. Keep your file close, and give it to every AI that writes for you.