The café owner writing next week's specials on a sticky note, the artist drafting grant applications alone at midnight, the surf instructor who wants a website but doesn't know where to start, and everyone who's quietly thought — there must be a better way.
Real work done — on day one
Every module is hands-on. You'll complete actual tasks with AI — not watch a demo. By the end of Module 2 you'll have used AI on real work from your job or practice.
A personal toolkit, mapped to your role
There are hundreds of AI tools. You'll leave with a shortlist of the right ones for you — matched to your actual tasks — and the prompting habits to use them well.
Something you built
Modules 5–8 move from browser to build. You'll finish with a real project — a website, a tool, an agent — saved on GitHub and ready to keep working on.
Confidence about what's next
Most AI confusion comes from not knowing what's possible. After this you'll have a clear picture of the landscape, what the tools can and can't do, and exactly where to go deeper.
$ claude "Write a month of Instagram captions for my Raglan surf school"
✨ Creating 30 captions with local flavour...
$ claude "Build a simple booking page for my art studio"
🚀 On it. Generating project files now...
$
The AI Landscape in 2026
You've heard people talking about it at the school gate, in the line-up, at Raglan Roast. Time to actually understand what's happening.
- Tour of models: GPT, Claude, Gemini, DeepSeek, Qwen, Llama
- What makes them different — reasoning, coding, creativity, speed
- Free vs paid: the honest breakdown
- The shift from chat to agents — where this is all going
AI for Your Work, Right Now
No terminal. No code. Just you, a browser, and AI doing a surprising amount of your actual work — starting today.
- Prompting that works — the gap between bad and great is enormous
- Writing tasks: emails, social posts, proposals, grant applications
- Planning: lesson series, café menus, business ideas, event schedules
- Which free tool (Claude, Gemini, ChatGPT) to use for which job
AI for Creative Work
Before we touch the terminal, the fun bit. AI for artists, musicians, photographers, and makers — most of it free, right now, in a browser.
- Image generation: Canva AI, Adobe Firefly, DALL-E, Midjourney
- Music: Suno and Udio — full songs from a text prompt
- Video and motion: Runway, Kling — what's actually good right now
- Grant writing, artist statements, and Google AI Studio
Your Terminal Unlocked
This is the doorway. We'll open it carefully so Module 5 stops feeling like magic and starts feeling manageable.
- Terminal basics for absolute beginners (Mac, Windows, Linux)
- Installing Claude Code, Gemini CLI, and other AI tools
- Your first AI conversation from the command line
- Why this doorway matters for everything that comes after
Build Something Real
Use the class build space to turn one clear brief into a real small project on screen. No coding background required.
- What the build-space panes do: files, preview, AI panel, terminal
- Google AI Studio as a free multimodal build tool
- Paste one ready brief and watch a first draft appear
- Steer one clean revision and rescue one wobble without panic
Make It Real
Your first build exists. Now we make it behave like a real project: tighter prompts, stronger IDE habits, and GitHub as your safety net.
- Choose the next change instead of restarting the project
- Ask for smaller, smarter edits and inspect what actually changed
- Use preview, files, terminal, and AI panel as one build loop
- Put the project on GitHub with a first honest commit
Advanced Vibe Coding
You've built something. It works. Now learn the habits that separate people who vibe-code once from people who keep shipping real things.
- Context loading — teach the AI your project before it touches anything
- Screenshot-to-code: build from any image in 30 seconds
- The 5-step debugging loop that works every time
- Checkpoint commits, model shopping, and multi-file projects
Make Your Own Agent
An agent isn't science fiction. It's a program that can take actions, loop through steps, and report back. In this session you build one.
- What agents actually are — the Think → Act loop demystified
- Build a working research agent with nothing but a prompt
- Connecting agents to files, the web, spreadsheets, and APIs
- Your personal agent roadmap — the one thing you'd automate first
Finished a module and want to dig further? These self-directed deep dives expand on specific topics — no extra cost, complete them in your own time.
AI Ethics & Corporate Accountability
Who builds AI, and for what purpose? Anthropic vs OpenAI on military contracts, the labour displacement question, Bernie Sanders, and what "responsible AI" actually means in practice.
Economics & Your Stack
When do you start paying? Which tools actually work together? Pick your stack, plan your first week, and understand what you're actually buying when you upgrade your subscriptions.
Creative Tools Deep Dive
Proper Midjourney prompting, building songs in Suno from scratch, what's actually good in video generation right now, and an honest guide to AI for grant writing and artist statements.
Google AI Studio Deep Dive
Build with Gemini's free API, understand system instructions and temperature, use multimodal inputs, get a free API key, and turn a working prompt into deployable code in minutes.
Automation & Workflows
Make.com, Zapier, n8n — no-code tools that connect your apps and let AI act on real data. Learn to build automations that triage email, schedule posts, and run while you sleep.
Git & GitHub Deep Dive
Version control from first principles. Commits, branches, pushing to GitHub, and GitHub Pages — free hosting for anything you build. Get your project live with a public URL.
Claude Code
Terminal-based AI that understands your whole project. The most powerful way to build.
Real-Time Voice Agents
Build AI that listens and speaks back in under a second using the Gemini Multimodal Live API.
Security & Adversarial AI
How AI gets manipulated — and how to protect yourself. Based on Nicholas Carlini's research.
MCP & RAG
Give your AI a brain. Model Context Protocol connects AI to live data — files, databases, APIs. RAG grounds responses in documents you own.
Prompt Engineering Deep Dive
System prompts, few-shot examples, chain-of-thought, adversarial critique loops — the techniques that turn average AI outputs into something you'd actually use. Build your own prompt library.
Built for specific roles. Each module gives you job-ready prompts and workflows tailored to your sector — complete them in your own time, no extra cost.
AI for Admin & Operations
Cut admin time in half. Email drafts, meeting summaries, report structures, and document workflows — practical prompts you can use the same day.
AI for HR & Communications
Job ads that attract the right people, onboarding packs that get read, internal comms that land. With responsible use guidance for people work.
AI for Customer Service
Review responses, booking enquiries, FAQ pages, and seasonal campaigns — every customer gets a thoughtful reply without spending hours at your keyboard.
AI for Finance & Reporting
Turn numbers into plain language. Board summaries, funder reports, impact statements, budget justifications, and annual reports — without the dread.
AI for Teachers & Education
Lesson plans, differentiated resources, report comments, parent emails, and assessment rubrics — in the NZ curriculum context. Less marking, more teaching.
AI for Health Practitioners
Heidi Health, Indici, and ACC paperwork — 6 modules built for NZ GPs, practice nurses, and allied health. Includes capture quality, referral letters, and HIPC compliance.
Bring this to your team or venue
Running a half-day AI workshop for your staff, members, or community group is straightforward. Minimum 6 people. Bring your own space or use Old School Arts Centre, Raglan.
Works well for: arts organisations, small business networks, school staff PD, councils, community trusts. Every participant leaves having built something real. BYO laptops and headphones — several modules include audio/visual content.
Book now → Enquire first → See full outcomes & pricing breakdown →Or find the facilitator guide at tekete.co.nz/ai-workshop/facilitator-guide.html if you want to run it yourself.
Save a spot
No payment now — just let me know you're keen. I'll email you session dates, the venue, and a payment link once it's confirmed.
Tobias McRae
Secondary teacher · Developer · AI practitioner · Raglan local
Graduate Diploma in Teaching (Secondary) student at UoA, full-stack developer, and builder of Te Kete Ako — a free educational platform with 12,900+ resources for Aotearoa. Uses AI tooling every day to build real things, and wants to share that knowledge with the community he lives in.
Ready to Level Up? 🌊
Join the Google Classroom for module schedule, materials, and to connect with other Raglan learners.
Claim your 60-day access code
Workshop participants get 60 days of full Kaiako Pro access on Te Kete Ako — 12,929 NZC-aligned resources and Te Wānanga AI lesson drafting.