Practical AI skills for nonprofits and small businesses — built for the people who know your organization best.
The instinct to train newcomers first is understandable — but it is rarely the highest-return investment. Your most experienced team members know the workflows, the history, and the edge cases that make your organization work. Give them AI skills and they will do something newcomers cannot: redesign processes from the inside.
An experienced staff member who learns to automate a report does not just save their own time — they build the template that saves time for everyone who follows. They know which shortcuts create problems downstream, and which are genuinely safe. That judgment cannot be taught in a training session; it has to already be there.
Trained experienced staff become internal champions who improve processes, create reusable tools, and raise the AI fluency of the whole team over time. They save money, reduce errors, and enrich the experience your organization delivers to the constituents who depend on it.
The foundation everything else builds on. Every tool in this track is free. Staff are routinely surprised that professional-grade development tools cost nothing — and that learning them opens every other door on this page.
Version control for everyone — your work is never lost, every change is traceable, and collaboration becomes effortless. Free, runs in any browser, and is the first skill every other module builds on.
The free code editor used by professional developers worldwide. Set it up once and it becomes the hub for writing, running, testing, and version-controlling everything your team builds.
Extensions are free add-ons that make VS Code smarter for specific tasks. Learn which ones matter for nonprofit and small business work — and how to evaluate any new one before installing it.
Python is the language most AI tools are built in. Learn to set up a clean, isolated environment so you can install packages, run scripts, and avoid the version conflicts that trip up most beginners.
Demystifying code for staff who have never written a line. Three short exercises produce something real: a personalized greeting for a donor list, a word counter, and a file batch-renamer.
Two mistakes that are easy to avoid and expensive if missed. First: keep secrets out of your repository — folders like .claude store conversation history that may contain API keys and passwords; add them to .gitignore before your first commit. Second: control when your workflows run — a documentation edit should never trigger a full website rebuild. An LLM can help you configure both correctly, but you have to remember to ask.
The skills that make AI genuinely useful at work — not as a novelty, but as a daily productivity tool. Prompting well is the single highest-leverage skill most staff can develop right now.
Poor prompts produce poor output. Learn the anatomy of a prompt that consistently delivers useful results — and the common mistakes that produce generic, unhelpful responses. The person who prompts well becomes the person every organization most needs right now.
AI models range from nearly free to expensive. Learn to match model to task — using frontier models only where they earn their cost — and how to test two models side by side to make an informed choice.
Move beyond the chat interface. Learn to set up API access to major AI providers, store credentials safely, and connect AI capabilities directly to your own tools and workflows.
Running LLM APIs without limits is how organizations get surprise bills. Learn to set spending caps, monitor usage by model and feature, cache results, and calculate the cost of a workflow before running it at scale.
Claude Code works within a finite context window — the amount of conversation, open files, and tool output it can hold at once. Managing that window keeps sessions fast, focused, and cost-effective. While our examples use Claude Code, the same principles apply to any AI coding assistant — Copilot, Cursor, Gemini Code Assist, and others face the same constraints.
Ai Fun 4 Us
.claudeignore file tells Claude which folders to skip — the single biggest context gainCLAUDE.md lean; use Plan mode (Shift+Tab) before executing/compact between tasks, start a fresh session for unrelated workFrom free text-to-speech tools to fully conversational AI avatars, this track covers audio and video capabilities that were reserved for large organizations — until now.
Turn any written content into natural-sounding audio using a free, offline-capable Python library. Produce accessible audio versions of newsletters, announcements, and program updates in dozens of languages.
Transcribe recorded meetings with free local models, generate live captions, and build simple voice-command tools. Practical exercise: transcribe a board meeting and produce an LLM-cleaned summary.
Create a photorealistic digital presenter from a single photo or video recording — no production team required. Ideal for donor updates, training content, event recaps, and accessible program communications.
Combine a talking avatar with your own documents so it answers questions grounded in your actual content — not general internet knowledge. Built with Azure AI Foundry, Azure AI Search, and RAG.
Generate consistent, photorealistic images of a person across different visual styles — professional, illustrated, branded — using the Azure OpenAI image edits API. Maintain a coherent visual identity across all communications.
A fully conversational AI presenter that listens, thinks, and speaks in real time. Visitors can interrupt and ask follow-up questions, just like talking to a person. Combines Azure Voice Live with a live avatar in a single API call.
| Goal | Recommended Module |
|---|---|
| Pre-recorded explainer video with a human presenter | 3.3 — Batch video avatar |
| Animated FAQ on a website | 3.3 — Real-time photo avatar |
| Interactive Q&A grounded in your documents | 3.4 — RAG avatar agent |
| Consistent branded imagery for communications | 3.5 — GPT-image-1.5 |
| Fully conversational spoken interaction | 3.6 — Voice Live + avatar |
Understanding how your digital infrastructure works — and how to maintain it without a full-time IT department. Practical skills that save money and prevent the panicked calls when something breaks.
Examples in this track reference Microsoft Azure, the cloud platform our team knows best. Every concept and service has a direct equivalent on AWS, Google Cloud, and other providers — the skills and thinking transfer directly.
Navigate your cloud subscription with confidence — read cost dashboards, restart stuck services, set budget alerts, and check logs when something stops working. No developer required for routine tasks.
Understand how the internet finds your website and routes your email. Learn to read, add, and verify DNS records — and which records must never be accidentally deleted.
Code changes deploy to your website automatically — no FTP, no manual uploads. Learn to read a workflow file, understand when it triggers, and fix the most common failures by reading the log.
Serverless code that runs only when triggered — no servers to manage, no idle cost. Trace a contact form submission from button click through the function log to the email received.
Use AI tools to maintain living architecture documents and human-readable enhancement logs alongside your work. Good documentation makes future changes faster, decisions traceable, and new staff onboarding smoother.
Your website will not help anyone who cannot find it. Learn what search engines and social platforms look for beyond content: meta titles and descriptions, Open Graph tags for rich sharing previews, canonical URLs, JSON-LD structured data so Google understands your organization type, and how to manage your sitemap and robots.txt. Practical exercise: audit a live site, implement missing elements, and submit to Google Search Console.
Simple interactive experiences that increase engagement on a website, at a fundraising event, or in a training session — no game-development background required.
Build a text-based quiz in Python, add a web-based version using HTML and JavaScript, and use an LLM to generate question banks around your organization's mission or program area. Finish with a working quiz you can embed on your website or present at a fundraising event. Delivers as a standalone half-day session or as part of a longer program.
Capstone sessions that combine skills from earlier tracks into something your organization can actually use — and a structured working session to find where AI fits your specific workflows.
Read a spreadsheet of donors, use an LLM API to write a personalized thank-you for each, convert the messages to audio with edge-tts, and commit the whole project to GitHub with a README. Combines Tracks 1, 2, and 3. Estimated 3–4 hours with an instructor.
A structured 2-hour working session — not a lecture. Map your five most time-consuming tasks, score them for AI potential, and leave with a prioritized list of experiments to run in the next 30 days. No prerequisites required.
Modules can be delivered standalone (1–3 hours), combined into a half-day workshop, or sequenced into a multi-week program. Every format is available as a flat-fee cohort — easier to budget and to report on.
| Audience | Recommended Modules | Suggested Format |
|---|---|---|
| Executive directors, board members | 2.1, 2.2, 6.2 | Half-day workshop |
| Program staff | 2.1, 2.4, 3.1, 3.3, 5.1 | Three 2-hour sessions |
| Administrative / operations staff | 2.1, 2.3, 4.1, 6.1 | Three 2-hour sessions |
| Staff assigned to manage the website | 1.1, 1.2, 1.3, 2.5, 4.2, 4.3 | Full-day workshop |
| Staff who want to learn to code | 1.1 – 1.5, 2.3, 2.5, 4.4 | 6-week evening series |
Tell us about your team and we will design a training plan around your workflows, your tools, and your goals.
Get in touch