Learn AI the practical way.
Short lessons with exercises you can run on your own work today. No jargon, no hype, and every lesson ends with a check you can mark yourself.
Short answer
The Levy Street learning hub is a free set of practical AI lessons for New Zealand teams. Each lesson takes 10 to 45 minutes, teaches one skill such as writing a useful prompt or checking an answer against sources, and ends with an exercise and a self-check. Start with AI foundations for small business, then follow the pathway.
Start here
Five short lessons that cover what AI is good at, how to ask for what you want, and how to check the answer.
- AI foundations for small businessWhat AI chat assistants are good at, where they fail, and how to test one on your own work in 25 minutes. (Beginner · 25 minutes)
- Choosing your first AI workflowScore your recurring tasks on five criteria and pick the one AI workflow worth starting with. (Beginner · 30 minutes)
- Writing useful business promptsA five-part prompt structure that turns vague requests into drafts you can send with light edits. (Beginner · 20 minutes)
- Checking AI answers against sourcesBreak an AI answer into claims, trace each one to a primary source, and keep only what you could verify. (Beginner · 25 minutes)
- Giving AI useful contextBuild a one-page context pack so AI answers from your own facts, policies and voice instead of guessing. (Beginner · 25 minutes)
Using AI safely as a team
Privacy, policy, subscriptions and the habits that keep a team’s AI use safe and useful.
- Building a reusable prompt libraryTurn prompts that worked once into tested, owned templates the whole team can reuse. (Intermediate · 40 minutes)
- Protecting customer informationDecide in a minute what can go into an AI tool, de-identify what can, and check your tool’s data settings. (Beginner · 30 minutes)
- Creating a team AI use policyDraft a one-page AI use policy with AI’s help, then test it against real scenarios before your team sees it. (Intermediate · 45 minutes)
- Calculating time saved by AITime a task before and after AI, count review and rework, and work out the real net saving in hours and dollars. (Intermediate · 30 minutes)
- Evaluating an AI pilotBuild a 30-case test set, set pass criteria in advance, and make a clear go, adjust or stop decision. (Intermediate · 45 minutes)
- Choosing business AI subscriptionsCompare business AI plans on data terms, fit and real cost in NZD, then test two tools on the same work before buying. (Beginner · 30 minutes)
- Training an internal AI championPick the right internal AI champion, write their role card, and train them over six weeks without burning them out. (Intermediate · 40 minutes)
- Rolling out AI to a small teamRun the first 30 days of AI adoption with your team: an honest kickoff, two shared workflows and weekly practice. (Intermediate · 45 minutes)
- Managing AI mistakes and escalationGrade AI mistakes by severity, escalate them the right way, and rehearse it before the real one happens. (Intermediate · 30 minutes)
Everyday work
Practical lessons for the documents, emails, reports and research that fill a working week.
- Using AI for meeting preparationTurn your notes, emails and numbers into a one-page meeting brief, then rehearse the hard questions. (Beginner · 20 minutes)
- Turning meeting notes into actionsTurn rough notes or a transcript into an action table with owners, dates and source quotes, with gaps flagged. (Beginner · 25 minutes)
- Drafting emails with human reviewA brief, draft, review and send routine for AI-written emails, with a five-point check before anything is sent. (Beginner · 20 minutes)
- Researching a customer accountBuild a sourced one-page brief on a prospect before a sales call, then catch what the AI got wrong. (Beginner · 30 minutes)
- Creating a business proposal with AIDraft a client proposal from your own scope, prices and evidence, with a claims register that catches invented promises. (Beginner · 40 minutes)
- Analysing a spreadsheet with AIGet real answers from a job list or sales export, and catch the date, total-row and sample-size traps. (Beginner · 30 minutes)
- Summarising documents with citationsMake every point in an AI summary carry a page and an exact quote, then check the citations in five minutes. (Beginner · 25 minutes)
- Building a searchable business knowledge baseTurn 10 to 30 policies into an assistant staff can question, and test it before anyone relies on it. (Intermediate · 45 minutes)
- Using AI to prepare management reportsDraft month-end commentary from your P&L and managers' notes, with reasons that come from people, not the AI. (Intermediate · 40 minutes)
- Using AI for customer support draftsDraft customer replies from a one-page rules sheet, with a person checking and sending every message. (Beginner · 30 minutes)
Automation, agents and tools
From writing a workflow specification to connecting AI to your CRM and knowing when not to use an agent.
- Writing a workflow specificationLet an AI interview you, one question at a time, into a one-page spec of how the work really runs. (Intermediate · 40 minutes)
- Automating a simple approval workflowTurn a purchase approval rule into a decision table, test it on paper, then build it with a person still approving. (Intermediate · 45 minutes)
- Connecting AI to your CRMConnect an AI assistant to your CRM read-only first, check its answers, then allow one supervised change. (Intermediate · 45 minutes)
- Understanding AI agentsWhat makes an AI agent different from a chatbot, and a 30-minute exercise mapping what one actually does. (Beginner · 30 minutes)
- When a workflow does not need an agentA decision rule for picking the cheapest option that works: delete, automate, add one AI step, or use an agent. (Intermediate · 25 minutes)
- Evaluating AI generated codeA seven-check routine for AI-written scripts and formulas, practised on an NZ phone number clean-up. (Intermediate · 40 minutes)
- Building your first internal AI toolBuild a saved assistant that does one job well, test it on 10 real cases, and share it without leaking files. (Intermediate · 60 minutes)
- Testing a business AI assistantBuild a 20-case test set with pass rules, run it three times, and know whether your assistant is ready. (Intermediate · 45 minutes)
- Creating an AI budgetA 12-month AI budget with seats, usage, people’s time and contingency, built from dated prices and stated assumptions. (Beginner · 40 minutes)
- Writing AI vendor requirementsTurn rough notes into numbered, testable requirements that make every AI vendor answer the same questions with evidence. (Intermediate · 45 minutes)
Rollout, quality and risk
Rolling AI out across a business, measuring whether it worked, and the risks to manage along the way.
- Planning a safe AI rolloutWrite a one-page rollout plan with pilot gates, red lines and the conditions that make you pause. (Beginner · 30 minutes)
- Measuring adoption after trainingFind out whether AI training changed how people work, with a 30/60/90-day plan and honest ROI arithmetic. (Intermediate · 30 minutes)
- Using AI for market researchSize a New Zealand market from official data and keep the AI’s guesses out of your numbers. (Intermediate · 30 minutes)
- Source-checked content draftsDraft blog posts and newsletters with AI where every factual sentence points to a source you chose. (Beginner · 25 minutes)
- Customer FAQs from support ticketsTurn a month of support tickets into an accurate FAQ, drafted from your policies and checked against NZ consumer law. (Beginner · 30 minutes)
- Working with images in business AIGet reliable text and numbers out of photos and scans, and know when an image should not go into AI at all. (Beginner · 25 minutes)
- Understanding RAGLearn how AI answers from your own documents by doing the retrieval step yourself, then watching what breaks. (Intermediate · 30 minutes)
- API or subscriptionWork out whether a workflow belongs in a per-seat AI subscription or a per-token API build, with the arithmetic shown. (Intermediate · 30 minutes)
- Understanding AI tool permissionsDecide exactly what a connected AI assistant may read, change and send before you click Allow. (Intermediate · 30 minutes)
- Building a no-code AI prototypeBuild a reusable assistant for one task without code, then test it properly before anyone relies on it. (Beginner · 40 minutes)
- Documenting an AI workflowWrite the runbook that lets someone else run your AI workflow when you are away. (Beginner · 30 minutes)
- Managing AI subscriptions and accessFind every AI tool your team uses, move work onto business accounts, and remove access when people leave. (Beginner · 30 minutes)
- Recognising prompt injectionSpot hidden instructions in emails, files and web pages before your AI assistant acts on them. (Intermediate · 25 minutes)
- Evaluating AI translation qualityCheck an AI translation before it reaches customers or staff, and know when only a fluent human will do. (Intermediate · 30 minutes)
- AI accessibility checksRun a first-pass accessibility review with AI, mapped to WCAG 2.2, and know what it cannot check. (Intermediate · 30 minutes)
- Planning human review for AI outputDecide which AI outputs a person must check, who checks them, and how to keep the checks honest. (Intermediate · 30 minutes)
- Learning to code with AIBuild a small internal tool with an AI coding assistant, review it, protect secrets and know when to call an engineer. (Intermediate · 30 minutes)
Questions people ask
What self paced AI courses are suitable for New Zealand small business owners?
Self-paced options include vendor training from the AI providers, university short courses and free practical lessons. Levy Street’s learning hub has more than 50 short, free lessons written for New Zealand businesses, each with an exercise and a self-check, covering prompting, checking AI answers, privacy under the Privacy Act 2020, automation and agents. Start with AI foundations for small business, and see our guide to AI courses in New Zealand to compare options.
Are the Levy Street AI lessons free?
Yes. Every lesson is free to read and use, with no sign-up. If you want the lessons run as hands-on sessions for your team, using your own documents and tools, that is our AI training service.
Which AI tool do I need for the lessons?
Any mainstream assistant works: ChatGPT, Claude, Microsoft Copilot or Gemini. Lessons are written to be tool-agnostic, and where a step differs between tools the lesson says so. Our AI tools guides compare plans and data settings.
Where should a complete beginner start?
Start with AI foundations for small business, then follow the Start here pathway in order. Each lesson takes 10 to 45 minutes and ends with a self-check.
Want this working across your team?
Our team training turns these lessons into habits, with exercises built on your own documents and tools.