A case study from a New Zealand secondary school that piloted AI-assisted lesson planning and marking support. The results — and the unexpected challenges — are worth reading.
When Wellington College introduced an AI pilot programme for its English department in mid-2024, the goal was modest: reduce the time teachers spent on routine tasks without affecting the quality of student experience. Eight months later, the results exceeded expectations — but not without some surprises along the way.
The pilot involved twelve English and ESOL teachers using a combination of Claude and a custom-built school prompt library to support lesson planning, differentiation, and written feedback. Teachers kept time logs for the first three months, tracking how long specific tasks took before and after AI assistance was introduced.
The headline figure — a 30% reduction in total preparation time — came primarily from two sources: differentiated materials (where teachers previously prepared separate resources for different ability groups) and written feedback on student work (where AI produced a first draft that teachers then reviewed and personalised, typically in a fraction of the original time).
But the pilot also revealed challenges that the team hadn't fully anticipated. Some teachers reported an initial increase in workload as they learned to write effective prompts and develop a reliable workflow. "There was a learning curve," said the department head. "For the first month, some teachers felt like it was taking more time, not less. It was only once they'd found their rhythm that the time savings kicked in."
There were also unexpected pedagogical shifts. Several teachers reported becoming more ambitious with differentiation — offering more ability levels, more task variety — because AI made it feasible. "I started doing things I'd always wanted to do but never had time for," said one teacher. The implication is significant: AI may not just save time, but change what teachers consider possible.
