Switch course ▾
Session 5. Upskilling, Measuring and Briefing the Board
Objective. Plan teacher upskilling in tiers with a paper version, build a scorecard of six indicators of value and risk with baselines and triggers, and present a two-page AI strategy to a board or education department in three minutes, with honest uncertainty.
A strategy is only as good as the people who carry it and the habit of checking whether it works. Training that is one lecture changes little, and a dashboard that counts clicks tells a board nothing. This lesson goes beyond the key ideas. It shows how to build a training plan that teachers will actually attend, how to choose a few measures that can show the plan is failing in time, and how to speak to a board so that your honesty becomes your strength.
Key ideas
Upskilling in tiers
- Not everyone needs the same depth. Tier 1 for all staff: safe use, data rules, checking output (about two 45-minute sessions). Tier 2 for users: practice on real tasks. Tier 3 for champions: coaching colleagues and reporting problems.
- The UNESCO teacher competency framework groups competencies into mindset, ethics, AI foundations, pedagogy and professional learning, in three progression levels. Use it as a checklist of what to cover, not as a rule of how long.
- Pair input with practice and follow-up. EEF guidance on professional development stresses well-designed, well-selected and well-implemented PD: a weekly 20-minute slot with a peer is a practical way to follow input with practice.
A scorecard with six indicators
- Pick six: three for value (teacher time saved against a baseline, a pupil learning measure linked to the problem, teacher confidence from a short paper form) and three for risk (errors found in AI drafts, data incidents, an equity gap such as pupils without home access).
- For each indicator write the source (a log, an exit ticket, a form), the baseline, who records it and how often. If you cannot name the source, drop the indicator.
- Avoid vanity measures such as logins or number of drafts. They show activity, which can rise while value falls. The OECD Outlook 2026 reviews promising uses and key challenges together, which is why a scorecard needs risk indicators beside value.
Triggers and review
- A trigger is a level that forces a decision before anyone has to argue. Example: any pupil data incident means pause the tool and review within a week; no fall in planning time after eight weeks means stop or redesign.
- Review quarterly, on a fixed date, with the same six indicators. Report by school, not as one average, and show the equity indicator first.
- Small numbers are noisy. Eight weeks in three schools is a signal, not proof. Say so.
Brief the board in three minutes
- Use six parts in order: problem, proposal, evidence so far, cost, risks and safeguards, the ask with a decision date. Put the ask on the first line of the page so no one has to guess what you want.
- Say what you do not yet know. A board trusts a leader who names limits and a plan to learn more more than one who promises certainty.
- Prepare for two hard questions: 'What if it fails?' (your stop rule) and 'What does it cost if we do nothing?' (the problem and its cost). Have the one-page strategy on paper.
One scorecard, two kinds of indicator
Value indicators
Is it helping?
- Weekly planning time against the baseline
- A pupil learning measure tied to the problem
- Teacher confidence from a short paper form
Risk indicators
Is it safe and fair?
- Errors found in AI drafts
- Data incidents (any is a trigger)
- Equity gap: pupils without home access
Training that sticks
Start from what teachers must be able to do in the next month, three or four tasks, not a tour of tools. Pair a 45-minute introduction with a weekly 20-minute slot in which two teachers do a real task and check the result against a checklist. EEF guidance on professional development stresses design, selection and implementation, so build in follow-up and peer support from the start.
Use champions: two per school, with a little protected time, even one lesson a week, and a clear job to help peers, collect problems and report them. Prepare a paper handout version so that training runs in a school without a projector. End each week with one question: what did you do differently this week?
Measures that can embarrass the plan
A good measure can show failure. Planning time against the baseline may not fall. The error log may rise. The equity indicator may widen. Choose six measures, name the paper source for each and set triggers beforehand, so that nobody has to argue after the fact. Review on a fixed quarterly date and show the equity indicator first.
Be careful with causes: a change may come from a new textbook or an exam week, so say 'associated with' rather than 'caused by'. Report small numbers as counts, such as 3 of 12, not as percentages. If a trigger is reached, act within a week and record what you did, because that record is what makes your next report believable.
Examples from the classroom
From lecture to weekly slot
Scene. Aziz, a deputy head in Navoi, runs one two-hour lecture on AI for 40 teachers. A month later few have used anything. He changes the plan: a 45-minute introduction, then a 20-minute slot each week for pairs, a handout with a checklist and a champion who visits each pair once a month. After six weeks 9 of 14 teachers in the pilot have a checked task on file.
Insight. Practice with follow-up changes behaviour more than a single lecture.
The error that tripped the trigger
Scene. Malika's pilot in Kashkadarya has a baseline of 3 errors a month in teacher-made materials. In month two the log shows 7. The trigger fires. She pauses new materials for a week, adds a second pair of eyes for every worksheet and tells the department what happened. In month three the log shows 2. The board is told all three numbers.
Insight. A trigger set in advance turns a worrying number into a quick, calm response.
Honest uncertainty wins the vote
Scene. Dilnoza presents to a district board in Tashkent. Her first line is the ask: six more months and 800 units for printing. She says eight weeks in two schools is a signal, not proof. A member asks 'what if it fails?' and she shows the stop trigger. The board approves the extension and sets a review in three months.
Insight. Naming limits and a stop rule makes a request easier to approve, not harder.
Common pitfalls
- Running one lecture and counting it as training.
- Choosing indicators that cannot show failure, such as logins.
- Reporting only the best school and hiding the equity indicator.
- Waiting for the quarterly meeting when a trigger demands action now.
Training Plan, Scorecard and Briefing
- Open the Activity Kit: the SAMPLE cluster plan summary, tier template, scorecard template and briefing frame (5 min).
- Pairs (15 min): fill the three-tier training plan for the sample cluster: who, what, how long, weekly slot, paper version, who coaches.
- Pairs (15 min): complete the scorecard: six indicators with source, baseline, recorder, frequency and one trigger for each. Remove any vanity indicator.
- Groups of four (10 min): one person gives the three-minute briefing, two ask the hard questions, one times and notes what was unclear. Rotate once if time allows.
- Part B (12 min): assemble the two-page AI strategy: page 1 problem, value map, readiness and roadmap; page 2 data and safety, training, scorecard and the ask. Check that every part from sessions 1 to 4 appears.
- Alone (3 min): write your leadership task from the Kit.
Your facilitator will guide you through this activity in the session.
Questions to discuss
- Which indicator did you remove, and what vanity did it hide?
- What was the hardest question from the board group, and what would you change in your answer?
- Looking at your two pages, what would you tell the board you do not know yet?
Reflect
- Who in your school could be a champion, and what time would you give them?
- Which number would you most fear to measure, and why measure it anyway?
- What decision do you want from your board or department, and by when?
Try it tomorrow
- Plan one 20-minute weekly slot for two teachers and write the real task they will do.
- Choose one value and one risk indicator for a tool you use, with source, baseline and trigger.
- Write the first line of a briefing that states your ask and the date you need the decision.
Further reading and links
Train in tiers with a weekly slot, measure value and risk with six indicators and triggers, and ask the board for a clear decision while saying what is still unknown.