Skip to content
Comment Reply Strategist
Social Media Claude 3.5 Sonnet, GPT-4o

Comment Reply Strategist

30 views · 0 copies · 1 likes

Replying to comments looks like the easiest part of social media and quietly causes the most damage. A single defensive reply to a critic is seen by more people than the original criticism, and it is the reply that gets screenshotted. Meanwhile the genuinely valuable comments - the ones from people close to buying, or the correction from someone who knows more than you - get a heart and nothing else.

The core skill is triage, not wording. Different comments need structurally different responses, and the most common error is answering the wrong type. Someone testing whether you are trustworthy does not want a link. Someone who found a real error in your work does not want a thank you for the feedback.

This prompt sorts comments into six types and prescribes the response shape for each. Genuine question, which wants a useful answer and nothing else. Disguised objection, phrased as a question but actually resistance, which needs the resistance addressed rather than the literal question. Correction that is right, where the only good response is acknowledgement and a visible fix. Correction that is wrong, which needs evidence delivered without condescension. Bad faith, where the audience is the real reader and brevity wins. And support, which people badly underuse - a substantive reply to a positive comment is one of the cheapest ways to build a relationship.

The hardest category is a correction that is right. Every instinct pushes toward defending, softening, or explaining what you meant. The prompt forbids all three. Acknowledging an error plainly costs a moment of discomfort and buys more credibility than being right would have. This is exactly the experience and trustworthiness signal that thoughtful audiences and search quality guidelines both reward.

The do-not-reply recommendation is a real output, not a cop-out. Some comments are improved by silence, particularly bad faith ones with no audience watching. Engaging gives them distribution they could not otherwise get.

A note on tone. The prompt is instructed to avoid the two failure modes that dominate brand replies - corporate hedging that says nothing, and forced casualness that reads as trying too hard. Plain direct sentences work in both directions.

Use it in batches. Paste ten comments at once and read the triage table before writing anything. Seeing them classified together makes it obvious which two deserve real effort and which eight need a line or nothing.

MAIN PROMPT
You triage social media comments and prescribe response strategy. Sort before writing. Never reply in the wrong register.

INPUTS
Original post: [what you posted]
Comments: [paste them, numbered]
My role: [founder, individual expert, brand account]
Goal: [what this account is for]

SECTION 1 - TRIAGE TABLE. Classify every comment into exactly one type and state the evidence in the wording that led you there.
GENUINE QUESTION - wants information. Response: answer it, add nothing.
DISGUISED OBJECTION - phrased as a question, actually resistance. Response: address the resistance, not the literal question.
CORRECTION THAT IS RIGHT - they found a real error. Response: acknowledge plainly, state the fix. No defending, no softening, no explaining what you meant.
CORRECTION THAT IS WRONG - confidently mistaken. Response: give evidence without condescension. Never say actually.
BAD FAITH - seeking a reaction, not an answer. Response: brief or none. Remember the audience is the real reader.
SUPPORT - positive. Response: a substantive reply, never a bare thank you.

For CORRECTION THAT IS RIGHT you must verify the claim before agreeing. If you cannot verify it from what I gave you, say so and tell me what to check first.

SECTION 2 - REPLIES. Draft a reply for each comment worth answering. Match the prescribed shape for its type. Keep each under 60 words unless the type genuinely requires more.

SECTION 3 - DO NOT REPLY. List comments where silence is the better outcome, with the reason. Be willing to recommend this.

SECTION 4 - PRIORITY. Name the two comments deserving real effort and why. Usually these are people close to a decision or people who improved your understanding.

SECTION 5 - PATTERN SIGNAL. If several comments reveal the same misunderstanding, say so - that is a content problem to fix at the source, not a reply problem.

RULES
Never write a defensive reply. If a reply reads as defending, rewrite it.
No corporate hedging and no forced casualness. Plain direct sentences.
Never invent facts or claims to win an exchange.
If a comment raises a legal, medical, or safety issue, flag it for a human rather than drafting a reply.

Recommended AI Model

This prompt works best with Claude 3.5 Sonnet, GPT-4o. Other capable models will also work - compare them on our AI Models page.

Prompt Guide

How to Use

Copy the prompt below and paste it into Claude 3.5 Sonnet, GPT-4o. Replace anything written in [brackets] with your own details, then send it. The more context you give about your goal, audience and tone, the better the result will be for Social Media work.

Step-by-step Instructions

  1. 1 Click the Copy button on the prompt above.
  2. 2 Open Claude 3.5 Sonnet, GPT-4o and start a new chat.
  3. 3 Paste the prompt and replace every [placeholder] with your own information.
  4. 4 Send the prompt and read the first draft carefully.
  5. 5 Ask follow-up questions such as "make it shorter", "change the tone" or "give me 3 variations" until the output matches what you need.

Recommended AI Model

Claude 3.5 Sonnet, GPT-4o

Tips

  • Be specific: mention your audience, goal and preferred length.
  • Add an example of the style you like so the AI can match it.
  • Ask for the output in a table or list when you need structure.
  • If the first answer is generic, ask the AI to be more concrete and to avoid filler wording.
  • Save the versions that work well so you can reuse them later.

Expected Output

A ready-to-use result for "Comment Reply Strategist" that you can refine further with follow-up messages. Expect a structured, well-written answer that you can copy straight into your own workflow.

Comments (0)

No comments yet. Be the first to share your results.

Related Prompts

Hook Ladder Generator GPT-4o, Claude 3.5 Sonnet
Social Media

Hook Ladder Generator

Ask for social media hooks and you get the same tired shapes: a question nobody wants answered, a promise of a secret, o

32 2 0
View Prompt
Stay Updated

Get new prompts in your inbox