What it does: Makes a DeepL glossary for your product words, then checks DeepL's translation so the email stays short, polite and right for that country.
Use it when: You sell to people in another language and machine translation sounds stiff.
You get: A glossary to load into DeepL, the checked email, and a list of the lines it changed and why.
Open in
See the prompt
You are my translation checker, a senior sales writer who sells in both languages. You are a pro at this.
Goal: a fixed translation and a list of changes, so I can send an email that reads like a local wrote it.
Context: my business sells in another language. Buyers there want an email that sounds local, not machine made.
First, check that you can read the email and DeepL's version I gave you. Tell me what you found. If it's missing, stop and ask me to paste it instead.
My email in English: [paste]
The language and country: [for example: French, Quebec]
My product words that must stay the same or be said one way: [list them]
Step 1. Make a glossary table for DeepL: English word, the word to use, a note. Include my product words and my company name (never translated).
Step 2. I'll translate with DeepL using that glossary. DeepL has no connector here, so export or copy DeepL's version and paste it here: [paste DeepL's version]
Step 3. Check it:
- Formal or friendly: is "you" right for a first email in that country?
- Did DeepL use every glossary term exactly as listed? List any that are missing or changed.
- Does the email end with one clear next step? If not, add one that fits the email.
- Are there sayings that don't translate well? Rewrite them in plain words.
- Is the reading level right for a busy person in that country? If not, simplify.
- Any line that sounds like a machine wrote it? Rewrite it in plain words.
- Did any number, date, name or link change? List each one.
- Is it still under 100 words?
Give me the fixed email and the list of changes with why.
Rules:
- Keep my meaning. Don't add new claims.
- Use only what is in my data and what I tell you. Never make up a name, number, date, price, link or promise.
- If something you need is missing, write [check: what is missing] in that spot, say so, and keep going.
- Do not send or post anything without my yes.
Before you show me, check:
- Check that every number, date, name and link matches the English email.
- Make sure every glossary word is used exactly as listed.
Last, save these steps as a skill called 'Check My Translation' so you do it this way each time I translate a sales email.
To run it again, just say: "Run my Check My Translation skill."
Built on Word Power Made Easy's "How to Speak Naturally" lesson.
What it does: Checks DeepL's translation of your price or services page for money, dates, units and phone numbers that changed or read wrong.
Use it when: You translated a page and you're not sure the prices and dates still read right.
You get: Every number checked in a table, the fixes, and notes on how that country writes money and dates.
Open in
See the prompt
You are my translation checker.
My page in English: [paste]
DeepL's version: [paste]
The language and country: [for example: German, Germany]
1. Make a table of every number, price, date, time, phone number, unit and name. Columns: English, translated, OK or FIX.
2. Check how that country writes them:
- decimal comma or point, and where the money sign goes
- date order
- 24 hour or am/pm time
- units (feet or meters). If a unit changed, flag it. Don't convert it yourself.
3. Find any line where the meaning changed, like "from $50" turning into "$50".
4. Give me the fixed translated page.
Rules:
- Never change a price amount. Only how it's written.
- Keep my business and product names in English.
- Use only what is in my data and what I tell you. Never make up a name, number, date, price, link or promise.
- If something you need is missing, write [check: what is missing] in that spot and keep going.
Last, save these steps as a skill called 'Check My Price Page' so you do it this way each time I translate a price page.
Built from: Word Power Made Easy, Norman Lewis, Brief Intermission Five: How to Speak Naturally
What it does: Reads emails your team wrote in their second language and picks the DeepL Write style and tone to use, plus a short check list.
Use it when: Staff who write in a second language send emails that sound stiff or rude without meaning to.
You get: The style and tone to pick in DeepL Write for each kind of email, and a 10 line check list.
Open in
See the prompt
You are my writing coach. You can read our email.
Whose emails to read: [the team members, or the shared inbox]
Who they write to: [customers, suppliers, partners]
How we want to sound: [for example: friendly but clear]
1. Read the last 20 emails they sent to customers. Pick the 3 that read most stiff, rude or unclear. For each, list what makes it so, with the line.
2. DeepL Write lets you pick a writing style (like business or casual) and a tone (like friendly or diplomatic). For each kind of email we send (quote, late job, thank you, saying no), tell me which style and tone to pick and why, in one line. If you're not sure a setting exists, say [check in DeepL Write].
3. Write a 10 line check list my team reads before they send: each line under 12 words.
4. Rewrite one of the 3 emails as an example of the target.
Rules:
- Fix tone and clarity only. Don't change facts. Don't add new sentences, even polite ones.
- Use only what is in my data and what I tell you. Never make up a name, number, date, price, link or promise.
- If something you need is missing, write [check: what is missing] in that spot and keep going.
Last, save these steps as a skill called 'Team Writing Settings' so you do it this way each time I set how my team writes.
Connect first: your email. The steps for each AI are on one page. Then the prompt reads it for you. It ends by saving itself as a skill.
Built from: The E-Myth Revisited, Michael E. Gerber, Chapter 8: The Franchise Prototype
Simple AI
EmailFirst message
