Explain another way
Offer shorter steps, a key idea or a relevant diagram when the learner needs another route into the lesson.
EDUCATION / ADAPTIVE TEACHING
A different answer.
A better next teaching step.
VexSpark™ learner support is being developed across the education family to respond to practice, revisit mistakes and offer clearer explanations while keeping the learner and responsible adults in control.
Adaptive teaching with VexSpark™
A learner’s next step can change with their answers, practice and chosen explanation style.
Practise the part that needs help
Use answer feedback and guided practice to revisit a mistake, then offer a more demanding check after success.
Return at the right point
Keep approved learning progress and explanation preferences, and recommend lessons that need another look.
Feature availability varies by App and version. Accepted™ has recorded offline rule-based adaptation; SafeSpark™ has public adaptation demonstrations. Wider education-App upgrades require their own implementation and checks.
Explore adaptive teaching and App stagesHow the teaching can adapt
A practical sequence: try, receive useful feedback, practise, review and continue.
Respond to the attempt
Use the learner’s actual answer and saved practice history to identify the part to revisit. An unanswered question is not proof of a missing skill.
Adjust the explanation
Offer another presentation or smaller steps. Learners retain their choice of lesson and explanation style.
Support another try
Provide targeted feedback and a guided check, then a more demanding practice task. Practice alone does not automatically complete the lesson.
Recommend revision
Bring lessons needing another check into a review queue and retain bounded progress/preferences where implemented.
Keep adult review visible
Parent and teacher support, approval and handoff are App-specific. Changes to learner presentation do not grant new access or remove approval requirements.
Evaluate the learning
Check whether the support helps real learners. Completion and repeated quiz answers are separate from independently demonstrated mastery.
Current implementation and next upgrades
Each App has its own release stage. This website explains the work; it does not install a tutor into an APK.
Accepted™ 2.4.0 — implemented component
Offline deterministic VexSpark rules use saved attempts to offer shorter teaching, answer feedback, two-choice/four-choice practice and revision recommendations. Explanation preferences persist. Simulated-flow checks are recorded; Android and educational acceptance remain outstanding.
SafeSpark™ — adaptation demonstrations
Public demonstrations show learner adaptation, profile correction and human approval. These are bounded demonstrations with fictional data; they do not establish that every role APK includes the same implementation.
BootCamp Hub — upgrade direction
Extend the course editions with attempt-aware support, alternative explanations, review practice and retained preferences. The reviewed 4.0.0 record establishes structured courses and quizzes; this review did not establish a completed adaptive-tutor upgrade.
WithMe and FindWithMe — upgrade direction
Build on stories, missions, quizzes, progress and existing learner help. A shared adaptation approach must be implemented and checked separately for each language and role App.
Vocational and companion Apps — upgrade direction
Apply task-specific coaching to CAD, culinary, painting/decorating and digital skills. Current adaptive behaviour needs version-specific evidence; subject competence cannot be inferred from a generic quiz.
Betta — welfare-first upgrade direction
Adaptive lesson support should remain within the approved education scope, with trainer review where appropriate. It must not invent diagnoses, treatment advice or unsafe training actions.
Clear limits, useful support
The learner’s history should help the lesson, with transparent and reviewable use.
Rules and models are different
Accepted’s current adaptation is rule-based. It does not contain a generative teaching model or secretly retrain model weights. Generative tutoring would require a separate implementation and evaluation.
Written and spoken work
Accepted’s written transfer task is explicitly ungraded. Read-aloud, transcription and pronunciation assessment are different capabilities; a speech engine alone does not establish teaching or assessment accuracy.
Learner data and safety
Explanation preferences are not diagnoses. Storage, export, deletion and adult controls must follow the specific App’s documented behaviour. Safeguarding and educational review remain part of acceptance.