Use permitted course context
Teaching starts from approved curriculum context and materials, with unverified coverage stated plainly.
Inspect the boundary before switching AI on
Tutoring between lessons, with staff in control.
Science tutor
Approved courseEvidence boundary
Text + voiceYou ask
Show why this next step needs review.
Tutor responds
The system can suggest a follow-up; staff approve it.
Ask, answer or explain it back...
Tutor shows it
Approved source→Learner attempt
Human review
You try
Who approves the follow-up?
A member of staff
Tutor remembers
Review the suggested follow-up
Useful AI, with visible ownership
See what the tutor does and who approves it.
Inspect the evidence
Staff can see the learner attempt and teaching response behind a suggested follow-up.
Keep people accountable
Assignments, family communication and consequential changes remain reviewable human decisions.
A controlled route from one cohort to wider use
Start small. Review the evidence. Expand carefully.
Define the boundary before the first learner enters
Agree permitted courses, materials, users and escalation routes before running a controlled cohort.
What happens hereHide details
- Scope approved content and capabilities
- Make human ownership explicit
- Keep unverified coverage outside the promise
Managed workspace boundary
Tutor
Explain. Check. Suggest.
Staff
Review. Decide. Own.
Source → attempt → review
Give every learner one managed tutor workspace
Within that boundary, learners move between explanation, voice, visual work and checked practice without stitching together separate AI tools.
What happens hereHide details
- Course context stays visible
- Learning history stays attached
- The product surface mirrors the permitted workflow
Course workspace
Subjects
Maths
Biology
English
AQA GCSE · Higher
Your course, ready to teach from
Your course
Your materials
Current focus
Factorising expressions
Inspect why the system acted
A learning signal should be traceable from approved source to learner attempt, tutor response and proposed follow-up.
What happens hereHide details
- Review evidence behind next actions
- Distinguish draft from approved action
- Keep accountable people in the chain
Educator handoff
Learner attempt
Increasing light always makes photosynthesis faster
Tutor response
Switched to a limiting-factor graph question
Draft next action
Compare three limiting factors.
ApproveEdit first
See who owns what
Inspect the decision boundary
Teaching
AI may
Explain and check.
A person owns
Set the boundary.
Set the pilot boundary before the first learner joins
Agree the cohort, permitted data, staff owner, review points, support route and exit plan. If our current evidence does not answer a procurement question, ask us directly.