Free Library · CC BY 4.0
AI MARKETING
PROMPT LIBRARY.
Battle-tested AI prompts used across $102M+ in tracked client revenue. Ad copy, email sequences, voice agent scripts, lead scoring, attribution frameworks. Copy, paste, use. Free under Creative Commons.
Ad Copy (2) →Email Sequences (2) →Landing Pages (2) →AI Voice Agent (2) →CRM Automation (2) →Content + SEO (2) →Lead Qualification (1) →Personalization (1) →Analytics (1) →
Replace [BRACKETED FIELDS] with your specific context · Tested on GPT-4o, Claude 3.5 Sonnet, Gemini 2.0
Ad Copy
Meta Ad Headline Variations
Generate 20 high-converting Meta ad headlines for any local service business
You are a direct-response copywriter who's written ads that generated $50M+ in revenue for local service businesses.
Write 20 Meta ad headline variations for a [INDUSTRY] business in [CITY/REGION] targeting [SPECIFIC AUDIENCE].
Requirements:
- Each headline 5-8 words
- Mix curiosity, urgency, specificity, and social proof
- Include 4 questions, 4 hard numbers, 4 specific outcomes, 4 contrarian takes, 4 transformation statements
- No clichés ("game changer", "revolutionary", "cutting-edge")
- Each one must work as a standalone Meta ad headline
Format as numbered list.Google Search Ad Generator
Create RSA variants for Google Search campaigns
Generate Google Responsive Search Ad variants for [BUSINESS NAME], a [INDUSTRY] business serving [LOCATION]. Service offered: [SPECIFIC SERVICE] Unique value: [DIFFERENTIATOR] Avg ticket: [$AMOUNT] Produce: - 15 headlines (max 30 chars each, mix of brand, benefit, urgency, social proof) - 4 descriptions (max 90 chars each, each addressing a different objection) - 4 sitelink extensions - 2 callout extensions Make sure headlines work in any combination (Google rotates them randomly).
Email Sequences
Lead Nurture Sequence (Cold to Booked)
5-email sequence that converts cold leads into booked calls
Write a 5-email cold lead nurture sequence for [INDUSTRY] business [BUSINESS NAME]. Context: Leads come in via paid ads but don't book on first contact. Need a sequence that re-engages them over 14 days. Constraints: - Each email under 150 words - Subject lines under 50 chars, no clickbait - Days 1, 3, 7, 11, 14 - Email 1: Friendly check-in, surface objections - Email 2: Specific case study with hard numbers - Email 3: Address #1 objection directly - Email 4: Offer a no-call alternative (free audit, free guide) - Email 5: Last touch — empathetic, "if not now, when" Voice: confident operator who knows what works. Not salesy.
Post-Quote Follow-Up Sequence
Re-engage prospects who got a quote but didn't sign
Write a 4-email follow-up sequence for prospects who received a quote from [BUSINESS] but haven't signed within 7 days. Specifics: - Industry: [INDUSTRY] - Avg project size: [$AMOUNT] - Sales cycle: 14-30 days typical Sequence: - Day 8: "Quick check-in" with specific question that surfaces objection - Day 14: Case study of similar customer who also took time to decide — what happened - Day 21: Address pricing/financing objection directly - Day 30: Honest "is this still on your radar?" with clear yes/no path Voice: not pushy, but direct. Sales is service.
Landing Pages
Hero Section Generator
Generate landing page hero section copy that converts
Write a landing page hero section for [BUSINESS], a [INDUSTRY] business. Target: [SPECIFIC CUSTOMER] in [LOCATION] Offer: [SPECIFIC OFFER] Differentiator: [WHAT MAKES YOU DIFFERENT] Output: 1. Pre-headline (5-8 words, sets context) 2. Main headline (8-12 words, big benefit + specificity) 3. Subheadline (15-25 words, expand on the benefit, add proof) 4. 3-bullet value props (each 8-12 words, concrete benefits) 5. Primary CTA button text (3-5 words, action-oriented) 6. Social proof line below CTA (1 line, specific number or quote) No vague phrases. Every word earns its place.
FAQ Section That Closes Objections
Generate FAQ content that handles real buyer objections
Generate 10 FAQs for a [INDUSTRY] business landing page that proactively close common objections. Context: [BUSINESS] sells [SERVICE] for [TYPICAL CUSTOMER]. Avg ticket: [$AMOUNT]. For each FAQ: - Question phrased the way a real prospect would ask it (not corporate-speak) - Answer that ACKNOWLEDGES the underlying concern + provides specific reassurance - 2-3 sentences max per answer - Include hard numbers where relevant Cover: - Pricing objection - "Is this right for me" fit objection - Trust/quality objection - Timeline objection - Refund/guarantee objection - "What if it doesn't work" objection - Onboarding/setup complexity - Specific industry-relevant concern - Comparison to alternatives - Next step clarity Voice: honest, direct, not defensive.
AI Voice Agent
Inbound Qualification Pathway
AI voice agent script for inbound lead qualification (Bland.ai, Vapi)
Design a Bland.ai pathway for an inbound AI voice agent handling [INDUSTRY] service calls for [BUSINESS NAME]. Context: - Avg ticket size: [$AMOUNT] - Service area: [LOCATION] - Operating hours: 24/7 (AI replaces voicemail) - Qualification criteria: [SPECIFIC CRITERIA] Pathway structure: 1. Greeting (under 8 seconds, friendly, brand name + name) 2. Intent identification (qualifier vs emergency vs question) 3. Qualifying questions (3-5 max, conversational not interrogative) 4. Booking flow (if qualified) — calendar integration 5. Escalation rules (when to transfer to human) 6. Closing (confirmation + next step) Voice tone: friendly professional. Sound like a great CSR, not a chatbot. Include exact words for each node + transition logic.
Outbound Cold Call Script (TCPA-Compliant)
AI voice agent script for compliant cold prospecting
Design a TCPA-compliant cold outbound AI voice agent script for [BUSINESS] prospecting [TARGET BUYER] in [INDUSTRY]. Offer: [SPECIFIC OFFER] Goal: book a 15-min discovery call Required: - Opening discloses recording (state-by-state aware: CA/FL/IL/MA need explicit consent) - Identifies as AI in first 15 seconds (best practice + legal in some states) - Handles 5 most common objections (no budget, no time, not interested, send me info, talk to spouse) - Books appointment if interested, sends info if "send me info", gracefully exits if no - Maximum 90 seconds if no engagement - DNC respect (immediate removal on request) Output as pathway with exact words.
CRM Automation
Lead Scoring Logic
Build a lead scoring system for GoHighLevel or HubSpot
Design a lead scoring framework for [INDUSTRY] business [BUSINESS]. Context: Lead volume is [VOLUME]/month from mixed sources (paid ads, SEO, referrals, cold outbound). Output: 1. Demographic scoring (company size, geo, industry fit) — 0-30 points 2. Behavioral scoring (pages visited, content downloaded, email opens, form fills) — 0-40 points 3. Intent scoring (specific actions like pricing page visit, booking page bounce, contact form abandon) — 0-30 points 4. Negative scoring (red flags like wrong geo, wrong industry, abusive contact info) — subtract points Define thresholds: - 80+ score: Route to closer immediately - 60-79: Route to AI qualifier first - 40-59: Multi-touch nurture sequence - Below 40: Long-term nurture or remove Provide specific point values for each criterion based on actual buying behavior.
Automation Workflow: New Lead → Booked Call
Map every step from first form fill to booked appointment
Map the complete automation workflow from "new lead enters system" to "booked appointment" for [INDUSTRY] business using GoHighLevel. Show every step with: - Trigger event - Action taken - Wait/delay if any - Branching logic (yes/no paths) - CRM tag changes - Integration touchpoints (Bland.ai, calendar, SMS provider) Cover both: 1. Lead converts immediately on first contact (booked within 5 min) 2. Lead doesn't convert immediately (multi-touch sequence over 30 days) Format as decision tree with exact GHL action names where applicable.
Content + SEO
Blog Post Outline Generator
Generate SEO-optimized blog post outlines for any topic
Generate an SEO-optimized blog post outline for the topic: [TOPIC] targeting the keyword: [PRIMARY KEYWORD]. Target word count: 2,500-3,500 E-E-A-T requirements: first-hand experience, expertise, case studies, opinions Output: 1. SEO title (under 60 chars, includes keyword, compelling for CTR) 2. Meta description (under 160 chars, includes keyword + benefit) 3. URL slug suggestion 4. Article structure: - Hook intro (2-3 paragraphs) - 8-12 H2 sections - 2-3 H3 subsections under key H2s - 5-8 FAQ entries at end (target People Also Ask) 5. 5 internal link suggestions (to other relevant pages) 6. 3 LSI keywords to weave in naturally 7. 1-2 specific stats or studies to research and include 8. CTA placement strategy (top, middle, bottom) The article must read like it was written by someone who actually does the work, not by AI.
FAQ for Featured Snippet Targeting
Generate Q&A content optimized for Google Featured Snippets and AI Overviews
Generate 15 question-answer pairs for [TOPIC] optimized for Google Featured Snippets and AI Overviews.
For each:
- Question phrased EXACTLY as someone would search it (use Google's People Also Ask format)
- Answer: 40-60 words, first sentence is a direct answer (this is what Google picks for snippets)
- Include 1 specific stat or data point per answer
- Avoid hedging language ("might", "could", "possibly")
Mix:
- 5 definitional ("what is..." questions)
- 4 comparison ("X vs Y" questions)
- 3 how-to ("how do I..." questions)
- 3 list ("best X for Y" questions)Lead Qualification
ICP (Ideal Customer Profile) Builder
Build a specific ICP from existing customer data
Build a detailed Ideal Customer Profile (ICP) for [BUSINESS NAME], a [INDUSTRY] business. Based on these patterns from existing best customers (highest LTV, highest satisfaction): [PASTE: 3-5 best customer descriptions including company size, revenue, industry, pain points, deal size] Output: 1. Firmographic profile (size, revenue, industry, geography, growth stage) 2. Demographic profile (decision-maker role, age, experience level, motivation) 3. Behavioral signals (what they Google, what tools they currently use, what triggers their buying decision) 4. Pain points specifically (rank top 5 from most painful to least) 5. Disqualifying criteria (who should NEVER be sold to — too small, wrong industry, wrong fit) 6. Where to find them (specific channels, communities, ad platforms, partnerships) Be specific enough that a sales team could screen prospects in under 60 seconds using this ICP.
Personalization
Personalized Cold Email Opener (Per Prospect)
Generate genuinely personalized cold email openers at scale
Write 5 personalized cold email opener variations for this prospect: Name: [PROSPECT NAME] Company: [COMPANY] Role: [TITLE] Recent news/post: [RECENT LINKEDIN POST OR COMPANY NEWS] Industry: [INDUSTRY] Goal: book a 15-min call about [OFFER] Constraints: - 1-2 sentences each - First sentence references something genuinely specific to them (not "I see you're at [COMPANY]") - Second sentence transitions to relevance for their business - No "I noticed", "I came across", "saw you on LinkedIn" — actually reference what matters - Voice: confident operator who's done their homework, not desperate Each variation should hit different angle: (1) their recent post, (2) their industry challenge, (3) competitor comparison, (4) hard stat in their industry, (5) pattern interrupt.
Analytics
Attribution Model Decision Framework
Pick the right attribution model for your business
Help me choose the right multi-touch attribution model for [BUSINESS] in [INDUSTRY]. Context: - Sales cycle length: [DAYS] - Avg touches before conversion: [NUMBER] - Primary channels: [LIST] - Budget allocation question: [SPECIFIC QUESTION] Compare these models for my specific situation: 1. First-touch 2. Last-touch 3. Linear 4. Time-decay 5. Position-based (40-20-40) 6. Data-driven (Google's algorithmic) For each: - When it's best - When it misleads - Recommended for my situation: yes/no with reasoning Final recommendation with specific implementation guidance for [PLATFORM: GHL/HubSpot/Salesforce/etc].
WANT THIS BUILT
FOR YOUR BUSINESS?
We deploy these prompts (and the full AI infrastructure they live inside) for local service businesses. Average client outcome: 9.2× peak ROAS.
BOOK A STRATEGY CALL →