Create a practical copy-paste AI prompt library for my social media management work. Build prompts from the operating details below, not generic platform advice. The prompts must prepare drafts and analysis for human approval; they must not fabricate performance, customer stories, product facts, legal claims, or audience feedback. Inputs: [BRAND AND AUDIENCE] [PLATFORMS AND OBJECTIVES] [VOICE CLAIMS AND GUARDRAILS] [RECURRING WORKFLOWS] [ASSETS DATA AND APPROVAL RULES] Produce 10 distinct prompts covering the workflows I named, with a useful mix of content planning, platform-specific post drafting, creative briefing, repurposing, community-reply triage, social listening synthesis, campaign reporting, and stakeholder updates. Make every prompt ready to paste into ChatGPT or Claude. For each prompt, include: title; exact first-person prompt; 2-5 bracketed inputs; a defined output format; and a verification instruction. A post-drafting prompt must specify platform, audience, objective, asset, approved claims, CTA, and character or formatting limits. A reporting prompt must separate supplied metrics from interpretation, calculate only requested rates with stated denominators, compare periods only when comparable, and list hypotheses rather than claimed causes. A reply-triage prompt must categorize messages by urgency and escalation owner, never send a response, and flag legal, medical, safety, payment, harassment, privacy, or press issues. Include a short reminder to verify platform policies and current character limits before publishing. Before answering, check that each prompt has a distinct job and includes no unapproved claims or invented metrics. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [BRAND AND AUDIENCE] | Describe the brand, its products or services, and the specific audiences it serves. | BrightDesk makes appointment software for independent salons; audience is salon owners and front-desk managers. |
| [PLATFORMS AND OBJECTIVES] | List each social platform and the objective, audience behavior, and content role for it. | Instagram for education and community; TikTok for short operational tips; LinkedIn for partner credibility; goal is qualified demo interest. |
| [VOICE CLAIMS AND GUARDRAILS] | Provide tone rules, approved product claims, prohibited language, and sensitive topics. | Helpful, direct, never snarky; approved claim: automated appointment reminders; no revenue guarantees or customer results without approval. |
| [RECURRING WORKFLOWS] | List the daily, weekly, and campaign tasks for which you want prompts. | Weekly calendars, Reels scripts, carousel copy, comment triage, monthly reports, launch briefs, converting webinars into posts. |
| [ASSETS DATA AND APPROVAL RULES] | Describe available content assets, reporting data, approval owners, response times, and escalation paths. | Product screenshots, approved customer quotes, Meta and LinkedIn exports; product lead approves feature claims; support owns billing or account issues. |
How to use
- Replace the five inputs with your approved strategy, brand rules, platforms, and escalation owners.
- Select prompts that map to a real workflow, then add the exact asset, campaign details, or metrics before using them.
- Review drafts against current platform rules, approved claims, and the source material before scheduling anything.
- Follow up with: “Adapt the community-triage prompt for our response-time policy and add an escalation path for refund requests.”
Variations
Platform post draft
Use this when an approved asset needs copy for one channel.
Draft [NUMBER] posts for [PLATFORM] promoting [ASSET OR TOPIC] to [AUDIENCE]. The objective is [OBJECTIVE], approved facts are [APPROVED CLAIMS], and the CTA is [CTA]. Follow this voice guide: [VOICE RULES]. Give each version a hook, main copy, visual or on-screen-text cue, CTA, and any suitable accessibility text. Respect the current platform limits I provide: [LIMITS]. Do not invent product capabilities, customer outcomes, statistics, or urgency. Flag any claim that needs approval. Before answering, check that each version has a distinct angle and that the CTA matches the objective. Ask up to 3 questions only if a required input is missing.
Monthly performance report
Use this after exporting verified platform metrics.
Analyze these verified social metrics for [PERIOD]: [METRICS]. Our objectives were [OBJECTIVES], comparison period is [COMPARISON DATA], and notable campaign changes were [CHANGES]. Produce an executive summary, a platform table, calculation notes with denominators, the strongest and weakest content patterns, and 3 testable recommendations for next month. Separate facts from hypotheses; do not attribute results to a post feature without evidence. Explain comparability limits such as changed posting frequency, paid spend, incomplete data, or different date ranges. Do not create benchmarks or calculate a rate when its numerator or denominator is absent. Check every value against the supplied export before answering.
Comment triage
Use this to organize incoming comments and messages before responding.
Triage these incoming social comments and messages: [MESSAGES]. Apply these brand and response rules: [POLICY], with escalation owners [OWNERS]. Return a table with message summary, category, sentiment, urgency, suggested response goal, draft response if safe, and escalation owner. Categories must include praise, question, support, billing, privacy, safety, legal, harassment, press, and spam where applicable. Do not promise refunds, provide account-specific help, disclose personal information, debate harassment, or send anything automatically. Preserve the user’s wording only as needed for context. Check that every high-risk item is escalated and every draft uses only approved claims. Ask up to 3 questions only if policy is missing.
Tips
- Maintain an approved-claims sheet separate from brand voice guidance; a friendly sentence can still make an unsupported product, pricing, or results claim.
- Use platform-native source assets, such as the final video, carousel frames, or approved screenshot, when drafting copy so captions do not promise what viewers cannot see.
- In reports, record numerator and denominator for every rate, especially engagement and click-through measures, before comparing channels.
- Route sensitive messages by topic and owner, not by sentiment alone; a polite billing or privacy question can require escalation.
FAQ
Can AI manage my social media comments?
It can classify messages and prepare approved drafts, but a person should handle publishing and escalation. Keep account, billing, privacy, legal, safety, and harassment decisions in established workflows.
Can AI analyze social performance exports?
Yes, if you provide the raw metric definitions, date ranges, and comparison context. Ask it to show denominators and distinguish observed patterns from explanations.
Should I use one prompt for every platform?
Use a shared brand foundation, then add platform-specific objective, format, audience behavior, and constraints. The same idea often needs different framing and assets across channels.