A lot of nonprofit teams are curious about generative AI, but that curiosity often sits alongside a second reaction: where do we even start?
Most people are not looking for a deep technical explanation. They are trying to work out whether AI can help with real tasks like drafting updates, summarising notes, rewriting content, or getting through repetitive admin a bit faster. They want something practical, not another layer of hype, much like the automated assessment and dashboards used in the Diversity Works New Zealand case study.
That is why we created this generative AI cheat sheet for nonprofits. It is a simple starting point for teams that want to use AI in a useful, grounded way without turning it into a big tech project.
What the cheat sheet covers
The cheat sheet gives a practical introduction to generative AI and how nonprofit teams can actually use it in day-to-day work. It covers the basics of what generative AI is, what makes a good prompt, and how to get better results by adding context, asking for the right format, and improving outputs through iteration.
It also includes example prompts for common nonprofit tasks such as drafting board updates, summarising feedback, rewriting content for different audiences, and turning rough notes into something clearer and more usable.
Who this is for
This cheat sheet is designed for nonprofit staff who are interested in using AI, but do not want to spend weeks figuring out tools, terminology, and best practice before trying anything.
It is especially useful if your team is dealing with the kinds of tasks that often get patched together under pressure: internal summaries, first-draft communications, rough planning documents, workshop notes, or content that needs rewriting for a different audience.
In other words, it is for teams who want practical help with the work in front of them.
What generative AI is good for in a nonprofit context
Used well, generative AI can help with parts of the job that are important but time-consuming. It can help you get started faster, organise information more clearly, and reduce some of the drag that builds up around writing, summarising, and repetitive formatting.
That might include drafting a first version of a donor email, summarising survey comments into themes, turning meeting notes into actions, or rewriting a long update in plainer English for a board or funder audience.
The value is not in handing everything over to AI. It is in making common tasks easier to move forward.
What is inside the cheat sheet
The cheat sheet includes:
- a plain-English explanation of generative AI
- guidance on writing clearer prompts
- a simple prompt formula you can reuse
- practical prompt examples for nonprofit teams
- tips for reviewing outputs and using AI with care
It is designed to be simple enough to use straight away, while still helping people avoid some of the most common mistakes.
A quick note on using AI carefully
Generative AI can be genuinely useful, but it still needs human judgement. Outputs can be too generic, miss context, or state things confidently that are not actually right. That is why the best use of AI is usually as a drafting and thinking partner, not a final decision-maker.
It is also important to avoid putting confidential or identifying information into public AI tools. If you are working with sensitive data, internal material, or information about clients, staff, or donors, extra care is needed.
Download the generative AI cheat sheet
If you want a practical starting point, you can download the cheat sheet here:
Download the Generative AI Cheat Sheet for Nonprofits
It is designed to help you get started with real tasks, using plain language and practical examples.
Start small, not big
You do not need an AI strategy deck to begin. A better place to start is with one real task that feels harder, slower, or clunkier than it should be.
That might be a board summary, a project update, a rough draft email, or a page of notes that needs organising. Start there. Use the cheat sheet to write a better prompt, review the output properly, and see whether it helps.
That is usually a much better test than trying to answer the question of “how should we use AI?” in the abstract.
