AI Workshops and Incursions for Schools: A Buyer's Guide

Quick answer

A genuinely useful AI workshop or incursion - the term for when an outside provider runs a session at the school itself, rather than students travelling out - looks nothing like an assembly-hall demo. The good ones are hands-on: students build or direct something themselves rather than watching slides. They upskill teachers alongside students, so the learning has somewhere to live once the visitors leave. And they leave a follow-up pathway - a next step, a resource, a way to keep going - rather than ending cleanly with the final "any questions?" If a provider cannot describe what students will actually do, what teachers will take away, and what happens afterward, that is the answer you need before you book anything.

Why this decision matters

Schools get pitched a lot of AI content right now, and it is not always easy to tell a substantive program from a well-produced one. Both can look impressive in a proposal. The difference only shows up in the room - and by then the budget is spent and the timetable slot is gone.

The stakes are not trivial. A wasted session costs a school more than money: it costs a timetable slot that could have gone to something that worked, and it costs student and teacher goodwill for the next AI session someone tries to run. Getting the selection right the first time matters more than it might seem from the proposal stage.

The assembly-hall demo problem

The easiest AI session to deliver is also the least useful: one presenter, one screen, a room full of students watching AI do something impressive. It photographs well. It generates almost no lasting capability, because nobody in the room had to direct anything, check anything, or make a decision themselves.

The tell is usually in the verbs. A demo is described in terms of what the presenter does - "shows", "demonstrates", "explains". A genuine workshop is described in terms of what students do - "build", "prompt", "test", "present". If a proposal is heavy on the first list and light on the second, you are looking at a demo wearing a workshop's name.

What separates a good incursion from a demo

Three features consistently separate the sessions that leave something behind from the ones that do not.

• Hands-on, not watch-on. Students should be directing AI themselves - writing prompts, building something small, checking and correcting output - not observing someone else do it. The learning lives in the doing, not the watching.
• Teacher upskilling built in. A session that only reaches students leaves teachers unable to follow up, reinforce, or extend anything covered. The stronger providers run a genuine component for staff, not a five-minute aside while students pack up.
• A follow-up pathway. What happens the week after matters as much as the day itself. A good provider leaves a resource, a next step, or an ongoing relationship - not a one-off event that ends when the van leaves the car park.

Questions worth asking before you book

A short, specific list of questions will tell you almost everything a glossy proposal will not.

1. What will students actually do during the session - not watch, do?
2. What do teachers take away that lets them extend the learning afterward?
3. What happens in the following weeks? Is there a resource, a follow-up session, or a clear next step?
4. Who is delivering it, and do they have real classroom or student-facing experience, not just AI expertise?
5. Can it flex to our year level and existing curriculum, or is it a fixed script regardless of audience?

If a provider answers all five with specifics, that is a strong signal. Vague answers to more than one or two are worth pausing on.

Making the case internally

If you are the one proposing an AI workshop to a principal or budget committee, frame it around what it builds, not what it shows. A session where students build something small and defensible - even a simple one - gives you something concrete to point to afterward, and it aligns naturally with the kind of AI literacy the Australian Framework for Generative AI in Schools is pushing schools toward: transparent, considered, taught rather than merely permitted.

It also helps to be honest about the follow-up commitment. A single incursion, however good, is a spark, not a curriculum. The schools that get the most from these sessions treat them as the start of something - a taster that feeds into ongoing AI literacy work, referenced in our broader look at what a good AI curriculum for secondary students should include - not a box ticked once a year. For the wider case on why this capability matters for this age group in the first place, see AI education for teenagers in Australia.

Common mistakes schools make

• Booking on production values alone. A slick proposal deck says nothing about what happens in the room.
• Leaving teachers out of the session entirely. Without staff upskilling, nothing continues after the provider leaves.
• Treating one incursion as "AI covered". A single session builds awareness, not capability.
• Not asking what students will actually do. If the answer is vague, assume it means "watch."

The recommendation: choose hands-on over impressive, teacher-inclusive over student-only, and a real follow-up pathway over a single memorable afternoon. Ask the five questions above before you book, and judge the session afterward by what students and teachers can do, not by how entertained the room looked at the time.