The same information needs to be completely reframed depending on who receives it. What works for engineers confuses executives. What excites investors bores developers.
AI is exceptionally good at this translation — if you define both ends of the bridge clearly.
02 Weak vs. Strong
EX 01Engineering sprint → CEO update
Translate this engineering sprint summary into a 5-sentence executive update for our CEO. She's non-technical but reads fast and asks good business questions.
Her priorities: Is the product improving? Is the team on track? Does she need to decide or unblock anything?
Translation rules:
- Replace ALL technical terms: no "API", "refactor", "latency", "deploy", "postgres", "cache" — replace each with a business outcome
- Convert metrics to impact: "reduced API p99 by 180ms" → "the dashboard now loads under 2 seconds for all users"
- If something is a risk, state the business consequence, not the technical cause
- End with EXACTLY ONE ask — what does she need to do? If nothing, write "No action needed from you this week."
Sprint summary to translate:
[paste your engineering update here]
→ Why it works
Each translation rule is specific with an example. Decision-forcing ending.
EX 02Tech spec → developer changelog
Translate this internal spec into a developer-facing changelog entry for our API docs.
Audience: developers who use our API in production and have been frustrated by unexpected 429 errors without clear feedback.
Translation rules:
- Lead with the problem it solves FOR THEM, not what we built
- Write in second person ("you can now..." not "we have added...")
- Replace internal system names with what developers see in API responses
- Include: what changed, new response headers or fields they'll see, any action required, what happens if they don't update their integration
- Length: under 200 words
- Tone: like a senior developer wrote it, not a marketing team
Internal spec:
[paste the internal spec here]
→ Why it works
Customer frustration as the starting point. Second-person rule. Internal-to-external translation explicit.
03 Key Points
01Define source AND target audience explicitly
02State what the target audience CARES ABOUT — not just their title
03'Remove X, add Y' for targeted translation vs. full rewrites
04For exec communication: replace process with outcomes, data with decisions
05For customer communication: replace feature names with problem-solution pairs
04 Model-Specific Notes
Claude is excellent at register translation. 'The reader cares about X, not Y' framing consistently produces audience-appropriate output.
05 For Your Role
Before sharing anything, ask: 'Rewrite this for someone who [their role] and cares most about [their goal].' The translation is usually surprising in a good way.