AI Prompting Workshop — URBANtxt Nonprofit Training

Section 1

Learning Objectives

By the end of this session, participants will be able to do all five of the following. Every segment of the workshop is designed to move participants toward these outcomes.

Objective 1

Use ChatGPT Confidently

Use ChatGPT confidently for real nonprofit research and writing tasks.

Objective 2

Recognize the Difference

Recognize the difference between a weak prompt and a structured one — and see why it matters.

Objective 3

Apply the Framework

Apply the 5-part prompting framework to write prompts that produce polished, usable output.

Objective 4

Research & Write Grants

Use ChatGPT to research grant opportunities and draft a needs statement.

Objective 5

Build Your Action Plan

Identify at least 3 tasks in their own workflow they can hand to ChatGPT this week.

Section 2

The 5-Part Prompting Framework

This is the core skill of the session. Every element matters. The more clearly you define each part, the more useful ChatGPT's output will be. Bookmark this page and use it before every prompt you write.

Element	What It Means	What to Ask Yourself
Context	Give background so ChatGPT understands your organization and goals. Example: "We are a nonprofit providing free tech education to underserved teens in Los Angeles. We serve 200 students annually."	What is the project and who are we?
Task	Be specific about the action you want completed. Example: "Search for grant opportunities from foundations or government programs that fund youth technology education in urban communities."	What exactly needs to be done?
Constraints	Set boundaries — tone, accuracy, limitations, do's and don'ts. Example: "Urban communities focus, relevant funders only — no general scholarship databases or irrelevant sectors."	What must be avoided or followed?
Output Format	Define structure — list, paragraph, bullet points, word count, etc. Example: "List 5 opportunities with the funder name, grant amount if known, and why it's a match for us."	What should the result look like?
Success Criteria	Define what a strong result includes — how you'll know it's good. Example: "Each item should name a real, active funder with enough detail to take action immediately."	How will we know it's good?

Section 3

Session Flow — 7 Segments

The arc: See it first → Understand why it worked → Do it yourself. Every segment builds on the one before. The example comes first — before any explanation — so participants feel the impact before they learn the theory.

Segment 1

Welcome & Hook

Open with a live demo — no slides, no agenda. Type a weak prompt live, then a structured one. Let the room feel the difference before any explanation.

0:00 – 0:08 · 8 min

Segment 2

What Is ChatGPT

Explain what ChatGPT is in one sentence. Cover what it's good at, what it's not, and bust three common myths. Introduce the nonprofit opportunity.

0:08 – 0:18 · 10 min

Segment 3

The Prompting Framework

Walk through all 5 framework elements using the Segment 1 prompts as examples. Live refinement demo. Ask participants what tasks eat their team's time.

0:18 – 0:35 · 17 min

Segment 4

Live Scenario: Grant Research

Walk through the full grant research workflow from scratch — find funders, research one in depth, then draft a needs statement — narrating every decision live.

0:35 – 0:55 · 20 min

Segment 5

Hands-On Practice

Every participant uses ChatGPT on a real task from their own org using one of three prompt starter templates — not a fake exercise.

0:55 – 1:15 · 20 min

Segment 6

Showcase & Debrief

Facilitator reads standout outputs from chat. For each one, ask what prompt they used — then show what a weaker version would look like.

1:15 – 1:25 · 10 min

Segment 7

Responsible Use & Action Plan

Responsible use guidelines, three highest-ROI uses for nonprofits, and one concrete action each participant will take this week.

1:25 – 1:35 · 10 min

Section 4

Live Demo Prompts

These are the exact prompts used during the live demonstrations. Type them exactly as written — don't paraphrase. The side-by-side contrast between the weak and structured prompt is the core teaching moment of the entire session.

Segment 1 — The Hook: Weak vs. Structured

Weak Prompt — Type This First "Find me grants for my nonprofit."

Read the output aloud. Ask the room: "Raise your hand in chat if this is useful." Pause. Let it land. Then say: "Now watch what happens when we change how we ask."

Structured Prompt — Type This Second "We are a nonprofit providing free tech education to underserved teens in Los Angeles. We serve 200 students annually and focus on closing the digital divide. Search for grant opportunities from foundations or government programs that fund youth technology education in urban communities. List 5 opportunities with the funder name, grant amount if known, and why it's a match for us."

Read the output. Let the room react. Don't explain it yet — just say: "We're going to spend the next 90 minutes learning exactly how to do that for your organization."

Segment 3 — Follow-Up Refinement Prompts

After the initial output, show participants how to refine without starting over. ChatGPT remembers the conversation.

Refinement Prompt 1 "Focus only on funders that accept applications year-round."

Refinement Prompt 2 "Which of these funders has the highest likelihood of funding a new applicant?"

Refinement Prompt 3 "Summarize the top 2 opportunities in 3 sentences each so I can share with my board."

Segment 4 — Full Grant Research Workflow (Use URBANTXT as the example org)

Step 1 — Research the Funding Landscape (7 min) "You are a nonprofit grant researcher. URBANTXT is a nonprofit based in Los Angeles that provides free technology education and mentorship to underserved teens. We serve approximately 200 students per year and focus on closing the digital divide in low-income communities. Find 6 grant opportunities from private foundations, corporate giving programs, or government sources that fund youth technology education. For each one include: funder name, estimated grant size, application deadline if known, and one sentence on why it's a strong match for us."

Step 2 — Dig Deeper on One Funder (5 min) "Tell me more about [Funder Name]. What are their stated priorities, what kinds of orgs have they funded before, and what should our application emphasize to be competitive?"

Pick one result from Step 1. Show how this surfaces strategic insight — not just a name and a deadline.

Step 3 — Draft a Needs Statement (8 min) "Now act as an experienced nonprofit grant writer. Using everything you know about URBANTXT and this funder's priorities, write a 350-word needs statement for our grant application. The tone should be compelling and community-focused. Include one or two statistics about the digital divide in underserved urban communities to support our case."

After the output, show one quick refinement: "Make the opening sentence more urgent and lead with the students, not the organization." Show the before and after. Key message: "We just did 3 hours of research and writing in under 10 minutes. That's the shift."

Section 5

Hands-On Practice Tracks

Drop all three templates in the Zoom chat at the start of Segment 5. Participants choose based on their role. They use their own org's real information — not a fake exercise. Give 12–14 minutes of independent work time.

Track A — Grant Research "You are a nonprofit grant researcher. [Your org name] is a nonprofit based in [city] that [what you do]. We serve [who] and our mission is [mission]. Find 5 grant opportunities that match our work. For each include the funder name, grant size, and one reason we're a strong fit."

Track B — Needs Statement "You are an experienced nonprofit grant writer. Write a 300-word needs statement for [your org] applying for a grant focused on [topic]. Our community faces [specific challenge]. Include relevant statistics and write in a tone that is [warm/urgent/formal]."

Track C — Funder Research "Research [specific funder name] and tell me: what are their current funding priorities, what types of organizations have they funded, and what should our application emphasize to be competitive?"

Facilitator check-ins — drop in Zoom chat every 4 minutes:

"Getting started? Paste your org's mission into the Context field first."

"Not loving the output? Ask ChatGPT to try again — add one more specific detail."

"Done already? Ask it to refine: make it shorter, more urgent, or add a statistic."

Section 6

Facilitator Notes & Prep

5 Things Every Facilitator Must Know:

The hook is everything. Do not skip the live demo at the top. If the room isn't engaged in the first 3 minutes, you'll spend the whole session trying to win them back.

Protect Segments 4 and 5. If you run long on context or framework, trim there — not on the demo and practice time.

Tech backup plan. Have screenshots of expected ChatGPT outputs saved as slides. If screen share or internet fails, you can still run the session.

Quiet practice room. If participants go silent during Segment 5, post energy in chat every few minutes. Silence usually means stuck, not done.

Use your own org. Referencing URBANTXT as the example org throughout makes the session feel cohesive and authentic — lean into it.

Before the Session — Prep Checklist:

Day-Before Prep

Slide deck prepared and tested

Prompt Framework Toolkit PDF ready to share in Zoom chat

Sample nonprofit scenario prepared (org name, mission, community served, a real-ish funding need)

ChatGPT logged in and open on facilitator screen — free account is fine

Weak prompt and structured prompt written out in advance — don't wing these

Zoom recording enabled

Backup screenshots of expected ChatGPT outputs in case of tech issues

Day-Of Prep

Start Zoom 10 minutes early for early arrivals

Post welcome message and ChatGPT link (chat.openai.com) in chat at open

Confirm co-facilitator roles if applicable

Post-Session Follow-Up:

After the Session

Send recording link within 24 hours

Email Prompt Framework Toolkit PDF

Include 3-question feedback form: What did you learn? What will you try first? What could be better?

Log common questions and struggles to refine the next session

Section 7

Test Your Knowledge ✏️

10 Questions · Workshop Knowledge Quiz 🎆

Get it right → fireworks + "You Are Awesome!" 🎆 | Wrong once → encouraging message + try again | Wrong twice → a helpful tip appears. You've got this!

0 / 10

Question 1 of 10

What does the "Context" element of the prompting framework provide?

A The specific action you want ChatGPT to complete

B Background so ChatGPT understands your organization and goals

C The structure and format of the response

D The boundaries of what to avoid

Question 2 of 10

Which of these is something ChatGPT IS good at?

A Replacing human judgment on important decisions

B Acting as a reliable fact-checking oracle

C Drafting, summarizing, researching, rewriting, and brainstorming

D Storing and protecting sensitive donor data securely

Question 3 of 10

What should you always do with ChatGPT output before using it?

A Share it immediately for maximum impact

B Review it — ChatGPT output is always a first draft

C Delete it and start over with a better prompt

D Run it through a second AI tool to verify accuracy

Question 4 of 10

What does the "Task" element tell ChatGPT?

A Who you are and what your organization does

B What format you want the response in

C Exactly what needs to be done — the specific action to complete

D How to measure whether the result is good

Question 5 of 10

What type of data should you NEVER paste into ChatGPT?

A Your organization's mission statement

B Publicly available grant opportunity descriptions

C Sensitive donor, client, or financial data

D Statistics about the communities you serve

Question 6 of 10

What does "Output Format" define in the prompting framework?

A The tone and voice ChatGPT should use

B The constraints on what ChatGPT must avoid

C The structure of the result — list, bullets, paragraphs, word count, etc.

D The background context about your organization

Question 7 of 10

Why does the session show the weak prompt BEFORE teaching the framework?

A To give participants a humorous warm-up before the real content

B So participants feel the impact of the difference before they learn the theory

C To demonstrate that simple prompts can sometimes work well enough

D To explain why ChatGPT is unreliable without professional training

Question 8 of 10

What are the three highest-ROI uses of ChatGPT for nonprofits identified in the session?

A Social media posts, blog content, and website copy

B Grant research and needs statement drafting, summarizing long documents, and email drafts

C Financial modeling, data analysis, and board reporting

D Hiring, onboarding, and performance reviews

Question 9 of 10

What does "Success Criteria" define in the prompting framework?

A The background context about your organization

B The specific action you want ChatGPT to complete

C How you'll know the output is good — what a strong result includes

D The word count and format of the response

Question 10 of 10

When ChatGPT includes statistics in its output, what should you always do?

A Trust them — AI only cites verified, accurate statistics

B Only keep statistics that come from government sources

C Remove all statistics from the output before using it

D Fact-check them — ChatGPT-provided statistics can be wrong

0/10

Questions Correct

AI Prompting Workshopfor Nonprofits

10 Questions · Workshop Knowledge Quiz 🎆

AI Prompting Workshop
for Nonprofits