Module 9 Classroom AI Practicum
The module where the workflow meets real students. Each teacher chooses one class, one chapter and 10–30 students, teaches the chapter normally, runs a diagnostic, identifies the gaps, assigns interventions, reassesses and records what happened in a weekly practicum log. The three facilitated hours are a 60-minute launch session and four 30-minute weekly clinics; the classroom work happens between the clinics, in the teacher’s own timetable, over four weeks. Student data stays in the school’s platform, students are told what is happening, and no student names appear in anything shared. What the teacher brings out is evidence: before and after on the same competencies, who received what, and the decision made each week.
By the end, teachers can…
- 1 Design a practicum — one class, one chapter, 10–30 tracked students, a four-week calendar — that fits the school’s timetable and annual plan.
- 2 Run the teach → assess → identify → diagnose → intervene → verify → review loop with real students at least once, end to end.
- 3 Keep a weekly practicum log that records what was taught, diagnosed, assigned (and to whom), reassessed and decided — in about ten minutes a week.
- 4 Handle student data responsibly through the practicum: keep it in the school’s platform, use codes rather than names in anything shared, and explain the practicum to students — and to parents where the school asks.
- 5 Tell evidence from anecdote: before/after results on the same competencies, who received which intervention, and a dated teacher decision.
- 6 Bring a problem to a clinic in the form "what I saw, what I tried, what I decided, what I need" and act on what the group suggests.
Facilitated time — 3 hours
| Length | Activity | Format |
|---|---|---|
| 30 min | Launch: the practicum design One class, one chapter, 10–30 students — small enough to actually do. The four-week calendar: week 1 teach and run the diagnostic; week 2 diagnose and intervene; week 3 intervene again and reassess; week 4 review and write up the evidence. Choosing the chapter (one due next month — Class 8 Maths Linear Equations in One Variable, Class 10 Science Light, Class 7 Social Science, Class 6 English) and the tracked students: a whole section of 30 or fewer, or a named subset of a larger one. The whole section still does the homework; the subset is who you follow in the log. Each teacher drafts their design on page one of the practicum log before leaving the room. | Input |
| 30 min | Launch: consent, ethics and what counts as evidence Student data stays in the school’s platform; it is not exported to personal tools or chat assistants. Student names never appear in clinic slides, chats or the capstone — codes (S01, S02 …) only. Students are told, in two minutes, what the practicum is and why; the school decides what parents are told and how. The AI tutor on student phones is provisioned by the school or pilot for practicum classes; teachers never press a family to buy anything. Then evidence: before/after on the same competencies, who received which intervention, the dated decision — and what is not evidence ("they seemed more engaged"). Teachers set up their tracked-student code list and the evidence page. | Hands-on |
| 30 min | Clinic 1 — reading the diagnostic Between the launch and this clinic each teacher has taught the chapter normally, logged the lessons, and run an 8–12 item diagnostic. Five-minute round; then results on the table, read by question and by student; the group helps each teacher write a diagnosis (which concept, which students, prerequisite or misconception) and an intervention plan for the week. Problems in the card format: what I saw, what I tried, what I decided, what I need. | Group work |
| 30 min | Clinic 2 — interventions in the wild Between clinics the interventions went out: targeted practice, a reteach in class, a note shared to the library. Who was assigned what; who used the tutor at home and who did not — and why (no phone in the evening, a plan not yet provisioned, not interested) — treated as data, with paper practice and class time as the alternative for those students. Early-warning signals for the group are read together. Plans adjusted; the log updated. | Group work |
| 30 min | Clinic 3 — reassessment Between clinics each teacher assigned a second check on the same competencies with different items. Results side by side with the diagnostic, by student code: who moved, who did not, who never attempted. The group helps decide week four: a second round for the students still weak, a whole-class reteach, or both — and which students need a conversation rather than a worksheet. | Group work |
| 30 min | Clinic 4 — evidence and the teacher review The review: three lists (improved / still needs help / reteach in class), one teacher-written parent line per tracked student who needs one, and the Monday class log. The evidence page completed — before/after by competency, by code, who got what, dated decisions. Each teacher says what goes into the capstone and what they would do differently. Coordinators note the clinic on the school’s CPD sheet. | Group work |
| Total 3 h · not one sitting — a 60-minute launch in Week 5, then four 30-minute clinics, one a week, across the practicum weeks. | ||
Key ideas
Teach normally
The practicum does not change how you teach the chapter. It adds two minutes of class log after each lesson, a fifteen-minute diagnostic, targeted practice for the students who need it, a second check, and a decision. If it feels like a different way of teaching, something has gone wrong.
Small enough to actually do
One chapter, 10–30 students, four weeks. A complete loop with a small class teaches more — and produces better evidence — than a plan for the whole school that never gets past week two. Pick the chapter you were going to teach anyway.
Evidence is boring on purpose
Before and after on the same competencies. Who received which intervention. The dated decision and the reason. Not "students were more engaged", not a screenshot of a cheerful tutor conversation. Boring evidence is what survives a sceptical colleague — and a DLD Appreciation Committee.
Students are not data
Their results stay in the school’s platform; their names never leave the classroom; they are told what is happening and why; and the teacher remains accountable for every intervention a child receives. The AI suggested; the teacher decided — and can say so to a parent.
Worked example
Four weeks with 10-A on Light — one teacher’s log (illustrative)
A Class 10 Science teacher chooses Light: reflection and refraction, with 10-A — 45 students, of whom 28 are provisioned on the school’s pilot cohort; those 28 become S01–S28 in the log. Week 1: she teaches the chapter as she always has, logs each lesson, and runs a 10-item diagnostic. Sixteen of 28 get the sign-convention items wrong; nine of those sixteen also cannot draw a ray diagram for an object beyond C; four others are fine on both but miss the mirror-formula rearrangement — an algebra prerequisite, not optics.
Clinic 1 sharpens the diagnosis: two real problems (sign convention; ray construction) and one prerequisite. Week 2: sign-convention practice for the sixteen; ray-diagram practice for the nine; a one-page "four rules for signs" note shared to the library; ten minutes of algebra in class for the four. At home, the tutors re-explain from the class log; eleven of the sixteen use it, five do not — three have no phone in the evening. Those three get the paper version and five minutes with her at lunch.
Week 3: a second check on the same competencies. Sign convention: 6 of 16 still weak. Ray diagrams: 2 of 9. The four algebra students: all fine. Early-warning flags S11 — engagement dropping; she asks; a cousin’s wedding, back next week. Not a worksheet.
Week 4: she reteaches sign convention to the whole class with a number-line trick, writes four parent lines — in her own words, from the record — and completes the evidence page by code. Her one-line "what I would do differently": run the diagnostic before teaching the second half of the chapter, not after. Illustrative: the shape of a practicum, not a claim about any class.
Assignment — counts towards certification
The completed practicum log
Keep the practicum log (resource: Classroom practicum log) for the four weeks: the design page; four weekly entries — taught, diagnosed, assigned and to whom, reassessed, decided; and the evidence page with before/after on the same competencies for the tracked students, by code, plus a dated decision list. Screenshots are welcome with names hidden.
Deliverable: The completed log as a PDF — design page, four weekly entries, evidence page — with no student names, plus a one-paragraph "what I would do differently".
| Criterion | Meets the standard when… |
|---|---|
| The loop was run with real students | The log shows teach → assess → diagnose → intervene → verify → review completed at least once on the chosen class and chapter, with dates. |
| Evidence, not anecdote | Before/after results on the same competencies for the tracked students, by code; and a record of who received which intervention. |
| Decisions recorded | Each week ends with a teacher decision — reteach, check-in, change of practice, parent line — and the reason for it. |
| Ethics kept | No student names anywhere; data stayed in the school’s platform; students were told; any departure from the plan is noted and explained. |
Where this lands in the PrepGraph workflow
Each certification task corresponds to a behaviour in the teacher portal — so the programme produces a working habit, not a slide deck. The pedagogy comes first; the surface implements it.
| Certification task | Pedagogical move | PrepGraph surface |
|---|---|---|
| Log each lesson of the practicum chapter | Class log so the tutor continues the lesson at home | Class logRecord what you taught today — topic, tagged concepts, homework, next test — so every student's AI tutor carries the lesson home. |
| Run the diagnostic and the reassessment; read results by question and student | Short competency diagnostic; mastery check with different items | Assignments & resultsSet a short diagnostic or homework from the question bank; see who attempted, how each question went, and give feedback. |
| Open a tracked student before the clinic | Student-level reading of practice and weak concepts | Student profilesOpen one student and see their practice, weak concepts and recent work in one place. |
| Assign targeted practice that the tutor carries home | Personalised practice and reteach at the student’s level | Student's AI tutorEach student has a personal tutor — voice in Hindi or English, chat in English today — that knows your class log, re-explains the homework items the student opens, and works at their level. Part of the student AI plan — practicum classes are provisioned with it; free-tier students still get homework, quizzes, doubts and progress. |
| Check the practicum group’s signals each week | Weekly early-warning review — who needs a conversation, not a worksheet | Early-warning signalsNightly signals per group — engagement dropping, scores slipping — so you know who needs a check-in this week. |
Materials
- •Classroom practicum log (printable and digital) — design page, four weekly entries, evidence page
- •Practicum design one-pager: class, chapter, tracked-student codes, four-week calendar against the school timetable
- •Two-minute student explanation script, and the school’s parent-communication note if it uses one
- •Clinic problem card: what I saw / what I tried / what I decided / what I need
- •Evidence page template: before/after by competency, by student code; who received what; dated decisions
- •PrepGraph teacher portal with the teacher’s real class, provisioned by the school or pilot
Facilitator notes
- •The launch must end with every teacher having chosen a class, a chapter and their tracked students, and with page one of the log filled. Otherwise week one is lost and the practicum slips into exam season.
- •Guard the clinics: thirty minutes, a five-minute round, then problems in the card format, then decisions. Not a lecture, and not a product help-desk — route login and setup questions to the coordinator outside the clinic.
- •Some students will not use the tutor at home — no phone in the evening, not provisioned, not interested. Treat it as data, not failure; paper practice and class time are valid interventions. Record them.
- •Never let a student name into a clinic slide or a shared chat. If a screenshot arrives with names, ask for it to be redacted before it goes further.
- •Coordinators record the clinics and the practicum on the school’s CPD sheet; the school decides how the hours are recorded under its CPD guidelines.
Bring the programme to your school
Tell us your board, grades and teacher count. We’ll reply within two working days with a cohort plan and the CPD record templates.