Automate responses to common supporter emails

The problem

You get the same emails repeatedly: 'How can I volunteer?', 'Can I visit your project?', 'Thank you for your work', 'I'd like to donate goods'. Writing individual responses takes 5-10 minutes each. You've got templates but they still need personalising. Staff spend hours on email admin that could be spent on actual work. The responses are important (supporter relationships matter) but they're repetitive.

The solution

Use Claude or ChatGPT to draft responses to common supporter emails. Create prompt templates for each email type (volunteer enquiry, donation thank you, visit request). Paste the incoming email, the AI drafts a personalised response using your tone and information. You review, adjust if needed, send. The AI handles the repetitive structure, you add the personal touch. 10 minutes becomes 2 minutes.

What you get

Draft email responses ready to review and send: personalised to the sender's specific question, in your organisation's tone, with relevant information included (how to volunteer, visit policies, donation process). Each draft needs quick review (check facts, add personal details AI can't know) but the heavy lifting is done. Same quality, fraction of time.

Before you start

• Common email types you receive (volunteer enquiries, donation thanks, visit requests, etc.)
• Standard information you share (how to volunteer, donation process, visiting policies)
• Your organisation's email tone (formal, friendly, professional)
• A Claude or ChatGPT account
• Willingness to review AI drafts before sending - never send without checking

When to use this

• You get the same types of supporter emails repeatedly
• Staff spend significant time on email responses that follow similar patterns
• You have templates but they still need substantial customisation
• You want consistent quality in supporter communications

When not to use this

• You get very few emails - writing them manually is fine
• Every email is genuinely unique and needs completely custom responses
• You're planning to send AI responses without reviewing them - always review first
• Emails contain sensitive personal information that shouldn't go to AI services (use paid tiers with data protection)

Steps

1. 1

   Identify your common email types

   Look at last month's emails: what questions come up repeatedly? Typical patterns: volunteer enquiries, donation thank yous, visit/tour requests, partnership enquiries, feedback/compliments, general information requests. List 5-10 types that make up 80% of your email admin.

2. 2

   Gather standard information for each type

   For each email type, collect: what information do you typically include? (volunteer application process, visiting hours, donation methods, who to contact). What tone do you use? (warm and personal? Professional and formal?) What shouldn't the AI say? (making commitments, sharing confidential details).

3. 3

   Create prompt templates

   For each email type, write a reusable prompt. Example: 'Draft a friendly response to this volunteer enquiry. Include: application process (link to form), typical roles available (youth mentoring, events support, admin), time commitment (4 hours/month minimum), and next steps. Use warm, encouraging tone. Here's the enquiry: [paste email]'. Save these prompts.

4. 4

   Test with real examples

   Take 3-5 recent emails of each type. Use your prompt template with Claude or ChatGPT. Review the drafts: Is the tone right? Is information accurate? Is it personalised to the sender's specific question? Does it sound like you? Refine your prompts based on what works and what doesn't.

5. 5

   Establish review workflow

   Critical: never send AI-drafted emails without reviewing. Check: facts are correct (AI might reference details that don't exist), tone is appropriate (overly enthusiastic or too formal?), personalisation fits (using the person's name right, addressing their actual question). Add anything AI can't know (personal connection, specific context).

6. 6

   Use for routine emails

   When common emails arrive: open Claude/ChatGPT, use relevant prompt template, paste the incoming email, review the draft, make adjustments, send. You're still in control, AI is just doing the heavy lifting. Your 10-minute email task becomes 2 minutes of review and personalisation.

7. 7

   Refine over time

   Notice what needs editing repeatedly? Update your prompt template. Find a new common email type? Create a new template. AI says something inaccurate or off-tone? Refine the prompt to be more specific. Your prompts improve with use, making drafts better over time.

8. 8

   Track time savings(optional)

   Monitor: how much time does this save? If you're drafting 15 emails/week and saving 5 minutes each, that's 75 minutes/week. Use that time for work that actually needs human expertise: relationship building, strategic planning, complex supporter issues.

Tools

Resources

Part of this pathway