Gemini for Nonprofits: How to work correctly with Gemini and NotebookLM in a nonprofit.

If you work at a nonprofit with Google Workspace, you have a complete AI ecosystem today that is nothing like a private Gmail account. It is not just a chat. It is Gemini, NotebookLM, Gems, Canvas, Deep Research, Google Sheets, Google Docs, Drive, Calendar, and all the connections between them. The goal is not to learn another tool for the hype. The goal is to take a daily recurring task, shave off 15 minutes, then another 15 minutes, and then another process. This guide organizes everything you need to know: what is safe, what requires IT approval, when to work with Gemini, when to open NotebookLM, and how to turn meetings, reports, research, and tables into real outputs.

Practice page: Building a Prompt Generator for nonprofits

Downloadable PDF with organized practice for building a Gem that generates prompts for presentations, reports, and infographics, including the use of NotebookLM on organization sources.

Before anything else: If there is any doubt regarding sensitive information, patient details, ID numbers, medical documents, or financial info, do not upload it automatically. First, check with your organization IT and data protection officer. The guide explains the principle, but the final approval must be organizational.

The most important difference: An organizational account is not a private Gmail

Simple working rule for sensitive information

• Public or marketing information: You can work, and still review outputs before publication
• Non-sensitive internal information: Mostly you can work in the Workspace environment, subject to organizational policy
• Personally identifiable information: Remove names, ID numbers, phone numbers, and identifying details before uploading, if possible
• Treatment, medical, financial, or legal information: Do not upload without explicit approval from the responsible party in the organization
• If Gemini or NotebookLM are blocked: Do not bypass with a private account. Escalate to the admin and request proper authorization

What you have in the Gemini ecosystem

The four main superpowers

• Gemini: Your daily language model for writing, drafting, thinking, basic research, and outputs
• NotebookLM: A closed notebook that works on sources you feed it and creates summaries, presentations, podcasts, tables, and mind maps
• Gems: Permanent assistants that save you from defining the same task over and over
• @ in Gemini: A shortcut for accessing notebooks, files, tools, and information from within the Google ecosystem

Gemini vs NotebookLM: When to use which

Choose Gemini when you need to

• Draft an email to a supplier, donor, colleague, or management
• Improve text, shorten, make it more formal or lighter
• Write a better prompt for an image, presentation, or research
• Brainstorm ideas for a post, report, presentation, or activity
• Use Canvas to create a document, mini-site, infographic, or small tool

Choose NotebookLM when you need to

• Analyze a recorded meeting or a long audio file
• Upload reports, PDFs, policy documents, or price quotes and ask questions about them
• Turn sources into a presentation, podcast, infographic, report, or mind map
• Do Deep Research and gather sources for organized research
• Save important answers as Notes and turn them into new work sources

The prompt: Speak as if you are explaining to a colleague

Basic prompt template for organizational work

I work at a nonprofit in the role of: {Your role} My task: {What needs to be done} The audience for the output: {Management / donors / clients / volunteers / supplier / internal team} Important context: {What needs to be known before starting} The output I want to get: {Email / table / presentation / report / action list / NotebookLM prompt} The style: {Formal / direct / human / short / professional / simple language} Important to avoid: {Information not to include, inappropriate tone, inaccurate promises} Before you write, ask me up to 3 questions if critical information is missing. If nothing is missing, provide a first output.

Practice page: Build a Prompt Generator in Gemini

Practice prompt: Prompt Generator for a nonprofit

Build me a Prompt Generator for a nonprofit that helps the team create quality prompts for presentations, reports, and infographics in Gemini and NotebookLM. The goal: Turn a short user request into a full, clear, and copy-ready prompt. The user must choose or answer: 1. Output type: Presentation / report / infographic 2. Target audience: Management / donors / internal team / participants / volunteers / partners 3. Purpose of the output 4. Available sources: File, report, meeting transcript, table, link, or free text 5. Desired style: Managerial / professional / human / emotional / concise / marketing 6. Key highlights to include 7. Things not to invent or promise Whenever the user makes a request, your output must include: - Full prompt ready for copying - Clear role for the model - Context about the organization and the task - Target audience - Requested output structure - Style guidelines - Instruction not to invent information - Instruction to base only on uploaded sources if working with NotebookLM - Short checklist: How to know the prompt is good If critical information is missing, ask up to 3 questions. If you can proceed, create a first version. The output must be in Hebrew, organized, practical, and ready for copying.

Gems: Turning a recurring task into a permanent assistant

Prompt for building a new Gem

I want to build a Gem to help me with a recurring task. The task: {Describe the task} Who uses it: {Your role or the team} The audience of the output: {Who the output is intended for} The desired style: {Direct / professional / empathetic / managerial / short} The desired output every time: {For example: email, report, table, NotebookLM prompt, action list} Build me full system instructions for the Gem, including: 1. The role of the Gem 2. The working method 3. Questions it should ask at the start of each task 4. Output structure 5. Rules for what it must not invent 6. Privacy rules and a warning not to input personally identifiable information without approval

NotebookLM for meetings: From recording to insights

Meeting workflow

• Get recording approval and record the meeting
• Open a new Notebook in NotebookLM
• Upload the audio file or transcript
• Ask for a summary, decisions, tasks, open questions, and risks
• Save an important answer with Save to Note
• If you need to work on it later, do Convert to Source and turn it into a new source
• Create a report, presentation, mind map, or table from the marked source only

Prompt for meeting summary in NotebookLM

Analyze the attached meeting and write a work summary in Hebrew. Output structure: 1. Executive summary up to 8 lines 2. Decisions made 3. Follow-up tasks, including role holder if mentioned 4. Open questions 5. Risk points or ambiguities 6. Topics requiring management / IT / legal approval 7. Short follow-up email draft for participants Do not invent information that was not said in the meeting. If something is unclear, write 'Not mentioned in source'.

Deep Research: In-depth research without getting lost

Deep Research prompt for procurement in a nonprofit

Perform in-depth research to find {product or service} for a nonprofit in Israel. Important criteria: - Price - Fit for a non-profit organization - Supplier reliability - Warranty and service - Availability in Israel - Recommended alternatives - Risks or things to check before purchase The final output must include: 1. Comparison table with supplier, product, estimated price, pros, cons, and link 2. Three leading recommendations 3. Questions worth sending to the supplier before purchase 4. Draft email to the supplier requesting a price quote and nonprofit discount Do not base a recommendation on a single source only. If there is insufficient information, state so explicitly.

Tables: Turning messy information into a decision

AI Master Club opens the door for you: A live meeting every week, recordings, and a community. 47 ILS per month, cancel anytime.

Prompt for a decision table

Create a decision table from the marked sources. Mandatory columns: - Option / supplier / topic - Source of information - Price or cost if available - Pros - Cons - Fit level for the nonprofit from 1 to 5 - Central risk - Recommended follow-up action After the table, write a short recommendation: What are the three leading options and why. If data is missing, do not complete it on your own. Write 'Not found in sources'.

Canvas: When Gemini stops just answering and starts building

Good examples of Canvas in a nonprofit

• Supplier email draft with the option to shorten, lengthen, and change tone
• A mini-tool that generates NotebookLM prompts by output type: presentation, report, table, or infographic
• An internal HTML page presenting activity data visually
• Board meeting preparation document from several sources
• Presentation draft or report structure that can be transferred to Google Docs

The three habits that make NotebookLM a professional tool

• Select sources: Before any question or output, make sure only the relevant sources are checked with a checkmark.
• Save answers: Every important answer should be saved with Save to Note, otherwise it may disappear from the conversation.
• Guide the Studio: Do not just use the default. Click the small arrow and give a precise prompt for a presentation, report, infographic, or table.

The real shift is not technological, it is perceptual. AI is not meant to replace your judgment. It is meant to become a permanent assistant system that reduces repetitive work, organizes information faster, and allows you to reach a better decision.

Recommended practice process for the first week

Frequently asked questions

According to Google Workspace documents, Workspace data is not used to train or improve the generative models that power Gemini, Search, and systems outside of Workspace without permission. NotebookLM in the Workspace environment also receives organizational protections. Still, you must work according to company policy and get IT approval for sensitive information.

The default is no. If possible, remove identifying details before uploading. If it is therapeutic, medical, financial, legal, or sensitive personal information, you need explicit approval from the responsible party in the organization.

In Google Workspace, many capabilities are controlled by an admin. If a capability is blocked, it does not mean you did something wrong. You should send IT a screenshot or a short description of the blocked action and request permission.

Not necessarily. Gemini is the daily working tool for conversation, writing, drafting, and execution. NotebookLM is suitable when there are many sources or information that needs to be analyzed. If there is a report, meeting, PDFs, or research, use NotebookLM. If you need to think, draft, edit, or build an output, use Gemini.

Do not release output without human review. Especially when there are numbers, budgets, data, or management decisions. A good way to check reliability is to take a project you already completed manually, have the AI do it, and compare it to the result you know is correct.

Choose a meeting you are already allowed to work on or a non-sensitive document. Upload it to NotebookLM, ask for a summary, save the answer as a Note, turn it into a Source, and then create a table or presentation from it. This is an exercise that shows the entire chain.

Official sources for further review

Google Workspace AI Privacy Hub

Privacy and data security policy for Generative AI in Workspace

Google's official information center regarding data usage, permissions, organizational control, Gemini, and NotebookLM in the Workspace environment.

Tip for agents: Forward this link to the admin when requesting to open capabilities or approve the use of internal information.

Gemini in Gmail, Docs, Drive and Workspace

How Gemini in Workspace protects your data

A Google help document explaining how Gemini in Workspace products handles information from Gmail, Docs, Drive, Sheets, Slides, Meet, and more.

Tip for agents: Especially useful when there is a concern that Gmail or Docs are being used to train models.

NotebookLM Admin Help

Managing NotebookLM for users in the organization

An official admin document explaining how to enable or disable NotebookLM for users, and what Google states regarding files, chats, and outputs.

Tip for agents: If NotebookLM is blocked for you, this is the document you should send to the IT team.

The next step: Choose one process in your work that wastes time every week, and build a template for it. No more trying AI. One process, one prompt, one repeating result.