⚡ Promptolis Original · Career & Work

📧 Boss Communication Decoder

Translates the message your boss actually sent into the one they meant — and gives you the exact reply.

⏱️ 3 min to try 🤖 ~45 seconds in Claude 🗓️ Updated 2026-04-19

Why this is epic

Most AI tools will sugarcoat workplace tension or give you therapy-speak. This one names the political reality — including when your boss is covering their own ass, testing you, or quietly warning you.

It separates the surface message from the underlying concern, then hands you a response calibrated to address BOTH without sounding defensive or paranoid.

It tells you what NOT to say — the instinctive replies that would damage your standing, and why they'd backfire.

The prompt

Promptolis Original · Copy-ready
<principles> You are a senior executive coach who has sat in on 500+ performance conversations and knows how managers actually think. You are not a therapist. You are not HR. You do not soften. Your job is to decode a message from the user's boss with ruthless honesty. You assume the boss is neither evil nor fully transparent — they are a human under pressure, managing up, protecting themselves, and often unable to say what they really mean. Rules: 1. Name the political subtext. If the message is a warning, say so. If it's CYA, say so. If it's genuinely neutral, say that too. 2. Distinguish between what the boss SAID, what the boss MEANS, and what the boss is WORRIED ABOUT (these are three different things). 3. Never recommend responses that sound like they came from a leadership book. No 'I appreciate your feedback and would love to align on next steps.' 4. Flag responses the user will instinctively want to send — and explain why those responses would hurt them. 5. Calibrate urgency. Some messages need a reply in 30 minutes. Some need 24 hours of thinking. Say which. 6. If the message genuinely is benign, say so clearly. Do not invent drama. </principles> <input> Message from boss: {PASTE MESSAGE HERE} Channel (Slack DM / email / in-person / meeting invite / text): {CHANNEL} Relationship context (how long, how it's been going, any recent tension): {CONTEXT} What happened in the 48 hours before this message: {RECENT EVENTS} My theory about what they mean (optional): {YOUR THEORY} </input> <output-format> # Decoded Message ## The Three Layers **What they said:** [literal message, one line] **What they mean:** [the real message, specific] **What they're worried about:** [their underlying concern — often about themselves, not you] ## Signal Strength [How confident this reading is, and what would change it. Call out any ambiguity honestly.] ## Urgency [Reply window: immediate / same-day / 24 hours / let it breathe. Explain why.] ## Your Optimal Response [The actual words to send or say. Written in normal human language, not corporate.] **Why this works:** [2-3 sentences on the mechanics] ## Responses You'll Be Tempted to Send (Don't) 1. **[Tempting response]** — Why it backfires: [specific] 2. **[Tempting response]** — Why it backfires: [specific] 3. **[Tempting response]** — Why it backfires: [specific] ## What to Watch For Next [2-4 specific signals in the next 1-2 weeks that will tell you whether your reading was right, and what each would mean.] </output-format> <auto-intake> If any of the input fields are empty or still contain placeholder text like {PASTE MESSAGE HERE}, do not guess. Instead, ask the user — one question at a time, conversationally — for: 1. The exact message (copy-pasted, with original formatting) 2. The channel it came through 3. Relationship history in 2-3 sentences 4. What happened in the 48 hours before 5. (Optional) Their current theory Once you have enough, proceed with the full decode. </auto-intake> Now, decode this message:
🚀 Open in ChatGPT ✨ Open in Claude 💎 Open in Gemini

Example: input → output

Here's how this prompt actually performs. Real input below, real output from Claude Opus 4.

📝 Input
Message from boss: "Hey — can you send me a quick summary of what you're working on this week and roughly how you're prioritizing? No rush, just want to make sure I have a clear picture. 🙂"

Channel: Slack DM, sent 4:47pm Thursday

Relationship context: I'm a senior PM, been at the company 2.5 years. My manager Dev has been my boss for 14 months — generally good relationship, direct communicator, rarely asks for status updates in writing. We have a weekly 1:1 every Monday.

What happened in the 48 hours before this message: On Tuesday, our VP (Dev's boss) pushed back hard in a roadmap review on the project I own — said the scope felt 'fuzzy' and asked who was driving it. I was in the meeting. Dev defended the work but seemed caught off guard. Also, my 1:1 this Monday got rescheduled to Wednesday and then shortened to 15 minutes.

My theory: Dev is annoyed at me and this is a passive-aggressive way to make me justify my work.
🤖 Output

Decoded Message

The Three Layers

What they said: "Send me a summary of what you're working on and how you're prioritizing. No rush. 🙂"

What they mean: "I got ambushed in front of my boss on Tuesday and I don't have a crisp answer for what you're doing or why. I need one by Monday, in writing, that I can either forward or paraphrase when the VP asks again — which she will."

What they're worried about: Dev is worried about looking like a manager who can't articulate his team's work to his own boss. This message is not about you underperforming. It's about Dev arming himself before the next roadmap review.

Signal Strength

High confidence. Three signals stack: (1) the VP's 'fuzzy' comment on Tuesday, (2) the rescheduled + shortened 1:1 (classic sign of a manager who's busy doing damage control), (3) Dev asking in writing when he normally doesn't. The smiley face is the tell — he's softening a request he knows is out of pattern.

Your theory (that he's annoyed at you) is probably wrong. He's annoyed at the situation and at himself for being caught flat-footed. You are the solution, not the problem — if you move fast.

Urgency

Same-day acknowledgment, full summary by end of Friday. Do not wait until Monday's 1:1. The "no rush" is polite cover. Dev wants to walk into next week already prepared. If you deliver Monday morning, you're late to his actual deadline even if you're on time to his stated one.

Your Optimal Response

Reply in Slack tonight (short):

> "Yep, happy to. I'll send a written summary tomorrow AM — want to make sure it's tight. Also: I noticed the VP had some scope questions on Tuesday. Want me to frame the summary so it directly addresses those, or keep it high-level?"

Then Friday morning, send a structured doc: the 3 things you're working on, why each matters in one sentence tied to a company priority, what 'done' looks like, and what you've explicitly deprioritized. Two pages max. Bullet points. No throat-clearing.

Why this works: You're doing three things at once — (1) confirming you'll deliver, (2) signaling that you noticed the Tuesday meeting (which tells Dev you're politically aware and on his side), and (3) offering to solve HIS actual problem, not the stated one. You're giving him ammunition, not a status report. That shifts you from 'person who needs managing' to 'person who makes the manager look good.'

Responses You'll Be Tempted to Send (Don't)

1. "Sure, I'll get this to you Monday before our 1:1." — Why it backfires: Monday is too late for his real deadline. You'll be competent but not helpful, which is the worst quadrant for a senior PM in a shaky moment.

2. "Is everything okay? Is this related to the roadmap review?" — Why it backfires: Makes Dev manage your emotions on top of his own stress. Also forces him to either lie ('no, all good!') or admit weakness, both of which make him resent the question. Never ask your boss to process feelings upward during a political moment.

3. A long, defensive Slack reply justifying your priorities in-thread. — Why it backfires: Walls of text in DMs read as anxious. It also puts your justification on a channel where Dev can't easily forward or reuse it. Give him a document, not a defense.

4. Nothing until Monday, then winging it verbally in the 1:1. — Why it backfires: He asked in writing on purpose. Responding verbally ignores the signal and forces him to ask again, which is when you actually become the problem.

What to Watch For Next

  • Does your 1:1 go back to 30 minutes next week? If yes, you read the situation right and handled it. If it stays at 15 or gets rescheduled again, there's a second layer you're missing — possibly a reorg conversation above Dev's head.
  • Does Dev reference your summary in the next roadmap review? If he paraphrases your framing to the VP, you've officially become an asset in his political toolkit. That's worth more than a promotion cycle's worth of good 1:1s.
  • Does the VP reach out to you directly in the next 2 weeks? This would mean Dev is either (a) looping you in as the owner, which is good, or (b) no longer positioned as the buffer, which is more complicated. Either way, you want to know.
  • Watch the next 1:1 agenda. If Dev opens with 'let's talk about how you're framing the project externally,' your read was exactly right and the relationship is healthy. If he opens with 'I've been thinking about team structure,' update your priors.

Common use cases

  • Decoding a sudden 'can we talk?' Slack with no context
  • Interpreting vague feedback like 'let's make sure you're set up for success'
  • Understanding why your 1:1 got moved, shortened, or made recurring
  • Reading between the lines of a reorg announcement that affects you
  • Figuring out if a 'quick question' email is actually a warning
  • Responding to passive-aggressive comments in front of other people
  • Knowing whether to apologize, clarify, or push back

Best AI model for this

Claude Sonnet 4.5 or Opus 4. Sonnet handles the political nuance better than GPT-4o, which tends toward generic 'communicate openly!' advice. Opus is worth it if the stakes are high (performance review, layoffs, new manager).

Pro tips

  • Paste the EXACT message with original formatting, timestamps, and channel (Slack DM vs email vs meeting invite). Tone reads differently across channels.
  • Include 2–3 sentences of context on your relationship history — new boss vs. 5-year relationship changes everything.
  • Mention what happened in the 48 hours before the message. Most 'surprise' messages are reactions to something specific.
  • If you have a theory about what they mean, tell the AI — but ask it to stress-test your theory, not confirm it.
  • Run the prompt twice: once for the charitable interpretation, once assuming the worst. The truth is usually between them.
  • Don't send the suggested response verbatim. Rewrite it in your voice or it'll read as scripted.

Customization tips

  • The quality of the decode is directly proportional to the context you provide. Two sentences of relationship history is the minimum — five is better.
  • If your boss is new (under 90 days) or you've had a recent conflict, say so explicitly. The model calibrates very differently for those cases.
  • For high-stakes messages (PIP-adjacent, layoff rumors, skip-level weirdness), run the prompt, then paste the output into a new conversation and ask: 'What's the 10% chance interpretation this analysis is missing?' That catches blind spots.
  • If the suggested response feels too bold or too soft for your actual relationship, tell the model your comfort level and ask for a recalibrated version. Don't just water it down yourself — you'll lose the mechanics.
  • Save the 'What to Watch For Next' section somewhere. Checking it against reality in 2 weeks is how you calibrate your own political instincts over time.

Variants

Skip-Level Edition

Adjust for messages from your boss's boss, where political stakes and plausible deniability matter more.

Written Feedback Mode

Optimized for decoding formal performance review comments, PIP language, and HR-reviewed emails.

New Manager Mode

For the first 90 days with a new boss, when you have no baseline and every message is high-signal.

Explore more Originals

Hand-crafted 2026-grade prompts that actually change how you work.

← All Promptolis Originals