The two reports sit in different places
Under Circular Memorandum 221/2025, the midterm report is submitted on 30 September 2027 and the final report on 30 September 2028, with the funding available up to 31 August 2028. The two reports serve different purposes. The midterm report shows progress, letting the EDB see the direction of the school's work partway through. The final report shows completion, the last account of the five commitments and the use of the funding.
The difference is in what is expected. At the midterm, the three subjects have started, some demonstrations have taken shape, and the first activities have been held, which is enough to pass. By the final report, the EDB is looking at whether the commitments were kept: whether the subjects really ran across the year levels promised, and whether the counts of demonstrations, open lessons, sharing sessions, and student activities hit their targets. The midterm can talk about the plan; the final has to talk about results.
From midterm to final: how the evidence carries on
Once the midterm report is in, evidence gathering should not stop. The same filing system keeps running: one evidence folder for each of the five commitments, plus one for the accounts, with the coordinator reviewing a master progress sheet each month. The midterm report is itself a ready stocktake of the middle stretch, showing which commitments are near done and which are still behind, which makes it the work list for the 2027/28 school year.
The key to the handover is to fill in what the midterm left unfinished while letting the finished parts keep building up results data. Take marking: at the midterm there may be only one term's usage record, while at the final you can offer a full year's before-and-after comparison. Time is on the side of the school that files systematically.
Closing out the five commitments
The final report's frame is still the five commitments: AI-assisted teaching in at least three subjects (two year levels each), at least six teaching demonstrations or resources, at least three open or demonstration lessons, at least three sharing sessions, and at least two student AI literacy activities. By this point, each one has to point to concrete, verifiable evidence, no longer an intention in a proposal.
- Three subjects, two levels: lesson plans and student work for every subject and every level, showing the work went beyond a lesson or two of trial teaching
- Six teaching demonstrations: the examples or resources have taken shape and can be referred to inside and outside the school, ideally with the context in which they were used
- Three open lessons: lesson plans, observation forms, post-lesson discussion notes and photos all present, with dates spread across the programme period
- Three sharing sessions: slides, attendance records and discussion points, with in-school subject panel sharing counted in
- Two student activities: the activity design, the student output and the number who took part, doubling as evidence of AI literacy
Settling the accounts and records
The funding comes from the Quality Education Fund, HK$500,000 per school, available up to 31 August 2028, with any unspent balance to be returned. In its February 2026 clarification, the EDB reminded schools to keep a separate account and to hold records for at least seven years. The financial part of the final report is where that account is settled: which commitment each item of spending goes to, whether the receipts are complete, and whether any unspent balance needs to be returned.
Filing purchase receipts by commitment is the easiest way through this step. If a receipts folder was set up by the midterm, the final only needs the last year's spending added in and the total worked out. Clear accounts are also what gives you confidence in any later audit, and the records have to be kept for seven years, so they do not end when the report is submitted.
After the programme ends, how the results carry on
The funding has an end date, but the teaching practices the school built up need not stop with it. Besides accounting for the commitments met, the final report is the moment to look at which practices are worth keeping: where AI-assisted teaching genuinely eased teachers' workload, which activities students responded well to, and which tools the school can afford to renew on its own.
Criteva, founded by former Hong Kong teacher David Leung, is an independent organisation with no affiliation to the EDB. Our position has always been that the funding is a starting point, not an end. Turn two years of gathered evidence into school-based experience and it can still guide the next stage of planning after the programme is over.