Your inbox has three AI training pitches in it right now. A generic "AI fundamentals" course. A Copilot subscription. A one-day workshop on prompt engineering. None will solve your actual problem.
Your actual problem is that AI tools have arrived in your organisation faster than your training, your policies, and your management practices have adapted.
Your staff are using ChatGPT, Claude, Copilot, and Gemini every day. Your managers are improvising. Your privacy framework was written before any of these tools existed. Your editorial standards do not mention AI. And somewhere in your organisation, a piece of AI-drafted content is about to leave the building with a hallucinated statistic in it.
The first instinct when senior leadership says "we need AI training" is to buy a prompt engineering course. This is the wrong instinct. Prompt engineering is a fifteen-minute skill. What your organisation is struggling with is everything around the prompts — the policy questions, the management questions, the legal questions, the editorial questions.
None of these are answered by a generic course. All of them are answered by training built around your specific policy, your specific tools, and your specific work. That is the gap this guide addresses.
There is no single AI training course.
There is a five-area framework.
Your AI policy exists. Your staff have not read it. Acceptable use training translates your policy into behaviour — names the tools your staff actually use, walks through realistic scenarios from their job, and tells them where to go when they are not sure. It is the cheapest, fastest, and highest-impact area to address.
Read the full guide on acceptable useYour managers are the bottleneck in your AI rollout. They are deciding which workflows get automated, which staff get reassigned, and which performance conversations need to happen when output quality changes. Almost none of them have been trained for any of this. This is the layer most organisations skip — and the one that costs the most when it is missing.
Read the full guide on manager trainingThe 2024 amendments to the Privacy Act 1988 changed the rules for any Australian organisation using AI. The Voluntary AI Safety Standard added another layer. Your existing privacy training, almost certainly written before any of this, is now out of step. This is the area that satisfies the regulatory direction.
Read the full guide on AI privacyAI tools produce confident, fluent, plausible content. Sometimes that content is wrong. Your staff are not currently trained to spot the difference, and "verify everything" is not a training plan. Of the five areas, this is the one that pays for itself fastest — it catches the mistakes the others do not prevent.
Read the full guide on critical thinkingYour comms, marketing, and HR teams are drafting with AI. None of your editorial standards account for it, and brand voice fragmentation is already happening. This is the most operationally specific area — it applies everything from the other four to the single most common AI use case in your organisation: writing.
Read the full guide on writing standardsHere is the sequence that works, based on what your budget allows.
Acceptable use.
The foundation. Everything else assumes your staff understand the basic rules.
Acceptable use plus critical thinking.
The first prevents the policy violations; the second catches the mistakes the policy does not cover.
Add the manager training.
The inflection point. The rollout starts working as a system: all-staff baseline, verification layer, leadership layer.
Add privacy or writing standards.
Heavily regulated organisations should do privacy. Organisations with high-volume external communications should do writing standards.
Sequence in the order above.
Acceptable use first, finishing within four to six weeks. Then critical thinking. Then manager training. Then privacy and writing standards in parallel.
Generic AI training has a structural problem: it cannot mention specific tools, specific policies, or specific workflows.
It teaches the abstract concept of "be careful with AI" and leaves every actual decision to the learner. That is not training. That is a quiz on common sense.
Custom AI training can name ChatGPT, Claude, Copilot, Gemini, and whatever else your staff are actually using. It can reference your specific AI policy, your specific data classification scheme, your specific approval process. It can walk learners through scenarios drawn from real situations in your organisation. It can update when the tools change — which they will, every quarter.
We build all five focus areas in three weeks each, with fixed pricing. Most fit within our Tier 1 or Tier 2 pricing. The legally complex ones — privacy especially — sit at Tier 2 or Tier 3 because they need proper legal review baked in.
For a closer look at how we approach AI training as a service, see our AI training category page.
The actual document, not a summary.
This determines what gets named in the course.
AI use cases from your context, anonymised. These become the scenarios.
Person, team, or inbox. The course needs to point somewhere.
So the new course extends rather than duplicates what is there.
That is enough for a single fifteen-minute kickoff call. If you do not have the policy yet, get one before commissioning the training. The Voluntary AI Safety Standard is a reasonable starting point.
For a deeper look at briefing well, see our guide to writing an effective eLearning brief.
SCORM-packaged for any LMS. Source files included. Unlimited revisions during the development period.
What you get: a custom AI training course, built around your policy, your tools, your scenarios.
What you do not get: a six-month engagement, surprise costs, generic content with your logo on it, or a course that will be out of date the day it launches.
One fifteen-minute call. Fixed-price quote within 24 hours. Three weeks from kickoff to launch.