Across Wellington, Dunedin and Christchurch, more than seventy science teachers attended the workshops in person. Of those, thirty-six completed the post-workshop survey — and thirty-four of thirty-six reported measurable growth. Every city, on every dimension, moved positively. This is what SCENZ funding produced.
Every respondent self-rated their confidence on three dimensions before and after the workshop, on a 1–5 scale. Reception of the format was captured separately. The numbers below describe what teachers walked out with.
of teachers walked out more capable than they walked in — a measurable lift on at least one confidence dimension.
average jump in everyday practical AI confidence — the dimension that translates most directly into Monday-morning classroom use.
said the balance of theory and practice was right for them. Zero teachers asked for less practice.
Three dimensions measured on a 1–5 self-rating scale, paired before and after the three-hour workshop. The largest movement is on the most actionable dimension — everyday practical use of AI tools.
The two non-growers had distinctive starting points: one entered at the ceiling (5/5/5 across all measures) and self-reported as the most experienced AI user in the cohort; the other registered no movement on any axis. The remaining thirty-four teachers reported gains.
Beneath the numbers, three benefits surfaced again and again in the open-text feedback. Each pillar below pairs a benefit with the verbatim words of a teacher who received it.
Many teachers arrived uncertain whether AI belonged in the classroom at all. They left with the language, the framing, and the permission to bring it in — thoughtfully, deliberately, on their own terms.
Thank you for today and affirming that AI has a valuable place in the direction of education in NZ — we just need to be brave and embrace it.
Not abstract concepts — working tools. Several teachers walked out with prototypes built during the session: custom GPTs, prompt templates, and assessment scaffolds they could deploy on Monday morning.
Great session. Benny helped me make a tool and was very knowledgeable and helpful. Being an actual teacher is all the difference. So worthwhile — will improve my teaching.
Some teachers were already steering. What they needed was confirmation that their direction was sound — and a sharper map for the next step. The workshop affirmed departmental approaches and accelerated existing work.
Time felt like it flew by. So good to have many questions about AI answered and affirmed the approach we are taking in our Department.
I had deep kōrero in the hui with other participants and have been able to bring those conversations back to kura immediately.
Thank you for an inspirational presentation, with many opportunities to share ideas.
Awesome — knew little and now have some knowledge.
Thank you, I really enjoyed the session and learned a lot about AI.
Great workshop. Thank you.
Very good personable tutor — thank you.
Really good to consider AI as a tool for learning and not a replacement.
Benny and Danny were awesome.
Beyond confidence and validation, every participant walked out with a tangible toolkit: ready-to-use AI assets built for the New Zealand context, frameworks for assessing student work, and the workshop's full reference materials.
The first item on this list — the NZ AI Policy Gem — was custom-built for this cohort. It generates AI policy drafts for individual schools, adapting to local community values, kura kaupapa, and the particular shape of each school's leadership. The remaining four items support teachers in turning that policy into daily practice.
A custom Gemini Gem, purpose-built by Benny for the New Zealand education context. Drafts AI policy tailored to a school's community, values, and decile setting — saving leadership teams the months of from-scratch policy work that would otherwise be required.
Slash-command patterns that let teachers save common prompts — lesson openers, marking rubrics, differentiation requests — and invoke them with a single keystroke. Demonstrated live in the workshop and included as a take-home worksheet.
A curated bundle of guidance from leading AI-security practitioners on data handling, prompt hygiene, student-data protection and the practical questions schools face when integrating AI tools into classroom practice.
Frameworks and ready-to-adapt templates for assessing students who are working alongside AI — rubrics that distinguish AI-supported learning from AI-dependent submission, plus guidance for designing AI-resilient tasks.
The complete deck, including the new productivity-workflow slides on custom commands, scheduled AI tasks, and the iterative prompt-refinement skill. Provided so teachers can revisit any segment at their own pace, or redeliver to colleagues.
The road trip travelled south from Te Whanganui-a-Tara to Otepoti, then back up to Otautahi. Each city brought a distinct cohort, but the directional finding holds: confidence rose across all three rooms, on all three dimensions.
Two questions, two clear signals. The first asked which topics from the workshop were most valuable for immediate teaching practice. The second asked what teachers want explored in a deep-dive follow-up. The answers point to a single, actionable next step.
Teachers selected up to two from the day's content. Productivity tooling and prompt craft drew the most votes — closely followed by AI ethics, and the foundational "how AI works" segment.
Teachers chose one option for a future deep-dive session. Customising AI for science curricula — specifically NCEA levels and formative feedback — was the runaway winner, claiming 39% of votes.
I would be interested in more AI courses.
Cashmere High · CHC respondent · verbatim
That voice was not alone. Every respondent — 36 of 36 — named a topic for a future deep-dive session when given the option. The largest single block, fourteen teachers (39%), chose “Customising AI for specific science curricula”. Three more wrote unprompted requests for follow-up work. The cohort handed us the next chapter, in their own words.
The data named the destination. The blueprint reaches it.
Indicative trajectory · modeled from Phase-01 dose-response on the same 5-pt self-rating scale, "Practical AI usage" dimension. NCEA-ready threshold defined as conventional ≥ 4.0 confidence-to-integrate band, plus headroom for subject-specific application.