Custom AI Prompts
Master the art of prompt engineering to get the best AI-generated content for your email campaigns. Learn proven prompting strategies and real-world examples.
Anatomy of a Good Prompt
Effective prompts have these components:
- Context: What information should the AI consider?
- Task: What exactly should be generated?
- Format: How should the output be structured?
- Tone: What style or voice should be used?
- Constraints: Any length or content restrictions?
Prompt Templates by Use Case
1. Personalized Opening Lines
Prompt:
"Based on the contact's company, recent news, and industry, write a 1-2 sentence personalized opening that demonstrates genuine research and establishes common ground. Be specific and reference concrete details. Avoid generic statements."
Example Output:
"I noticed Shopify recently launched Shop Pay Installments - congrats on making checkout more accessible for merchants. As someone focused on e-commerce infrastructure, I thought you'd be interested in how we're helping platforms reduce cart abandonment."
2. Pain Point Identification
Prompt:
"Identify the most likely pain point this contact faces based on their job title, company stage, and industry. Write it as a relatable statement (1 sentence) that shows you understand their challenges. Be specific to their role, not generic."
Example Output:
"As a VP of Sales at a high-growth SaaS company, you're likely struggling to scale personalized outreach while maintaining quality and avoiding spam filters."
3. Value Proposition
Prompt:
"Create a 1-2 sentence value proposition tailored to this contact's role and company. Explain how [YOUR PRODUCT] solves their specific challenges with concrete benefits. Include a relevant metric or outcome. Tone: confident but not pushy."
Example Output:
"WarmOpener helps sales teams like yours send AI-personalized emails at scale while maintaining 95%+ deliverability. Our customers see 3-4x higher reply rates compared to generic outreach."
4. Call-to-Action
Prompt:
"Write a soft, low-pressure call-to-action (1 sentence) that invites this contact to take a small next step. Tone: friendly and casual, not salesy. Focus on learning/exploring, not committing."
Example Output:
"Would you be open to a quick 15-minute call next Tuesday or Wednesday to explore if this could help your team?"
5. Social Proof
Prompt:
"Based on the contact's industry and company size, mention a relevant case study or similar customer (1-2 sentences). Be specific about the outcome they achieved. Format: 'We helped [similar company] achieve [specific result].'."
Example Output:
"We recently helped a Series B SaaS company similar to yours increase their outbound reply rate from 2% to 8% by personalizing at scale without hiring more SDRs."
Advanced Prompting Techniques
Include Examples (Few-Shot Prompting)
Give the AI examples of what you want:
Prompt:
"Write a personalized opening line based on the contact's company and recent news. Follow this style:
Example 1: 'Congrats on the Series B! I noticed you're expanding into Europe - we've helped similar companies navigate GDPR compliance during international expansion.'
Example 2: 'I saw you just launched the mobile app. As someone focused on user acquisition, you might find our approach to in-app onboarding interesting.'
Now create a similar opening for this contact."
Specify Tone with Adjectives
- Professional & Consultative: "thoughtful, insightful, data-driven"
- Friendly & Casual: "conversational, warm, approachable"
- Urgent & Direct: "concise, action-oriented, clear"
Add Negative Instructions
Tell the AI what NOT to do:
"Write a personalized opening. Do NOT: use generic phrases like 'I hope this email finds you well,' mention LinkedIn stalking, be overly formal, or exceed 2 sentences."
Common Mistakes to Avoid
❌ Too Vague
"Write something personalized for this contact"
Problem: AI has no direction, outputs generic content
❌ Too Prescriptive
"Write exactly: 'Hi {first_name}, I noticed {company} is in {industry}...'"
Problem: Defeats the purpose of AI, output too templated
❌ Missing Length Constraint
"Describe the contact's pain points"
Problem: AI writes 5 paragraphs instead of 1 sentence
❌ Conflicting Instructions
"Be friendly and warm, but also professional and formal"
Problem: AI can't satisfy both, output is inconsistent
Iterating on Prompts
Prompts rarely work perfectly on the first try. Here's how to refine:
- Test on 5-10 contacts - Generate AI content with initial prompt
- Review outputs - Check for quality, tone, length, relevance
- Identify patterns - What's working? What's not?
- Adjust prompt - Add specificity, examples, or constraints
- Re-test - Generate for same contacts with new prompt
- Scale when satisfied - Apply to full list once quality is good
⚠️ AI Limitations
Remember that AI can:
- Hallucinate facts (make up information not in the data)
- Produce inconsistent results (same prompt, different outputs)
- Misinterpret ambiguous instructions
- Generate content that sounds good but is factually wrong
Always review AI-generated content before launching campaigns, especially for high-value prospects.
💡 Pro Tip
Save your best-performing prompts in a document for reuse. Build a prompt library over time with templates for: openings, pain points, value props, CTAs, follow-ups, breakup emails. This saves time and ensures consistency across campaigns.