📨 AI for Social Impact Deep Dive: Prompt Engineering

✍🏼 A Note From the Editor

Wow! What an amazing response to the first edition—a 76% open rate! I am so grateful for your engagement and your feedback. Based on your responses, I am going to supplement the monthly newsletter with a deep dive — more in-depth info to help you activate AI in your workflows. Let’s get started!

🤖 Prompt Engineering Deep Dive

• What is prompt engineering?

  • Prompt engineering is the art and science of crafting instructions that get AI tools to produce what you need. Think of it as learning to speak AI's language instead of putting in vague requests, getting frustrated with the output, and giving up.

• Why does it matter?

  • Good prompting gets usable first drafts immediately.

  • Strategic prompting embeds your values, voice, and impact into every output.

🦀 Tiny Crabs Ride Enormous Iguanas

Do you remember Tina Huang’s mnemonic device? (ICYMI, check out the first issue here). Let’s take a deep dive into each element.

Task: What exactly do you want the AI to do?

Context: What background information does the AI need?

Resources: What materials, data, or examples can you provide?

Evaluate: How will you assess the quality of the output?

Iterate: How will you refine the prompt for better results?

🎬 Putting Prompting into Action - Example 1

🛑 Don’t say: “Write a thank you letter to a donor” because you will get a generic letter.

Try this instead:

Task: “You are a fundraising expert and your task is to write a personalized thank you letter for a first-time donor.”

Context: “Our organization provides after-school tutoring for middle school students in Chicago. This donor gave $250 for the first time after attending our annual gala. Our voice is warm, personal, and focuses on student success stories to demonstrate our outcomes. The letter should feel personal, mention specific impact, and include a subtle invitation for continued engagement. It should be 200-300 words and avoid flowery fundraising language.”

Resources: “Here's our mission statement: [paste mission]. Here's a brief story about one student we helped: [paste story]. Here are three specific outcomes from last year: [paste data].” Pro tip: Include an example of a donor acknowledgement that a human at your organization wrote and ask it to mimic its tone. 😉 

Evaluate: Did it mirror your organizational tone? Was it specific enough? Did it include any superfluous information?

Iterate: Based on your evaluation, you could ask AI to “rewrite this in a more conversational tone” or to “include an aspirational closing line about the ripple effects of our work.”

🎬 Putting Prompting into Action - Example 2

🛑 Don’t say: "Research nonprofit organizations working on digital equity."

Try this instead:

Task: “You are a subject matter expert in digital inclusion in under invested communities. Conduct a comprehensive landscape analysis of digital equity organizations to inform our foundation's potential $2M annual investment strategy in closing the digital divide.”

Context: “Our foundation is considering entering digital equity as a new program area, particularly focusing on rural and low-income communities. We typically fund evidence-based interventions with clear outcomes measurement, prefer organizations with annual budgets of $500K-$5M, and focus on scalable solutions. Our board values innovation, measurable impact, and avoiding duplication of existing major funders. We need to understand the current funding ecosystem, identify gaps, and assess whether our involvement would add unique value. Analysis should identify: 10-15 key organizations by size/impact, major funding sources and gaps, evidence-based intervention models, and geographic distribution. Include funding recommendations and present findings as executive summary and detailed appendix suitable for board review. Please cite all of your sources.”

Resources: “Foundation's current portfolio: education, economic mobility, workforce development. Geographic focus: U.S. with preference for Midwest/South regions. Previous research: FCC data showing 39% of rural Americans lack broadband access, $65B federal infrastructure investment underway. Board's interests: connectivity, digital literacy training, and small business support. Excluded approaches: direct infrastructure development, telecommunications lobbying, and hardware manufacturing.”

Evaluate: Did the nonprofits that it surfaced meet your criteria? Was there information missing? Did you check the sources to ensure accuracy?

Iterate: If missing foundation fit, ask it to strengthen its analysis. Or ask it to include concrete next steps, for example, which organizations to meet and what additional research is needed to make an informed decision.

*Try Perplexity for research queries. And always fact-check the citations!

🏆 Pro Tips for Prompt Engineering

1. Build Your Prompt Library: Create templates for common tasks. Save your best-performing prompts and customize them for different situations.

2. Include Your Values: Always specify your organization's voice, values, and approach. AI should sound like you.

3. Provide Examples: The more examples you give, the better the output. Include your best past work as a reference point. Remember to omit private or sensitive information!

4. Be Specific About Constraints: Word counts, tone, audience, and format matter. Specify exactly what you need.

5. Test and Refine: Your first prompt won't be perfect. Build refinement into your workflow.

6. Consider Your Audience: Tailor prompts based on whether you're writing for donors, board members, partners, or the general public.

About AI for Social Impact

I’m Joanna, and I am on a mission to help folks in the social impact sector understand, experiment with, and responsibly adopt AI. We don’t have time to waste, but we also can’t get left behind.

Let’s move the sector forward together. 💫

♥️ Spread the Love

Spread the love and forward this newsletter to anyone who might benefit from a dose of AI inspo!