The boredom drain
When work is too easy, students disengage in ways teachers rarely catch — homework gets done, scores look fine, but the brain isn't building anything new. Years of this and a child quietly stops trying.
← Back to home Teaching Philosophy
Every child has an optimal edge. Our job is to find it — every single day.
The work that grows a child's mind is neither the work they can already do, nor the work they cannot do at all. It is the narrow band in between. We obsess over keeping every student inside that band — measured, not guessed.
65–80%
Target accuracy band per session
90 min
Re-calibration cycle
1:1
Difficulty tuned per child
The three zones
Drag the slider. Watch what changes.
Vygotsky proposed three regions of difficulty for any given task. Move the marker below to see what happens inside a child's head in each one.
Difficulty 52/100
🎯
OPTIMAL
Zone of Proximal Development
The child cannot do this alone yet — but can do it with a hint, an example, or a tutor nudge.
This is where new neural pathways form. The struggle is real but solvable, the win is earned, and confidence grows alongside skill. Vygotsky called this the only zone where learning actually happens.
What you'd see: Pauses · asks a focused question · "wait, let me try…" · finishes with effort
“What a child can do with assistance today, she will be able to do by herself tomorrow.” — Lev Vygotsky, 1934
Why this matters
Most tuition aims at the class. Real learning happens at the child.
The frustration spiral
When work is too hard, the child fails repeatedly without traction. The brain learns one lesson very well: “I am bad at this.” Self-efficacy collapses long before grades do.
The ZPD payoff
Inside the ZPD, the child meets a problem just past their edge, struggles productively, and breaks through with a hint. That is the moment new neural connections form — and it is the moment confidence is built honestly.
It scales when measured
A great human tutor reads the edge by feel. That doesn't scale to 30 students. AI can measure the edge continuously, in real time, for every child — freeing the human to do the part that actually requires a human.
How we keep students in the band
The 90-minute calibration loop.
Click each step to see what's happening — for the child, for the tutor, and for the data layer that ties them together.
01
Diagnose
Where is this child, right now?
Each session begins with a 3–5 minute calibration set covering the topic at multiple difficulties. The AI infers the current edge — not from a stale test from last term, but from how the child responded in the past 24 hours.
Live signals
Response accuracyTime-to-solveHint depth usedMisconception fingerprint
New · Focuslab
The signal that closes the loop.
Until now, calibrating the ZPD relied on lagging signals — accuracy, time-to-solve, hint depth. All measured after the child got stuck. Focuslab is our in-house computer-vision model that reads microexpressions twice a second and tells us, in real time, whether the child is in Deep Focus, hitting a Eureka, or starting to Struggle. The edge isn't inferred. It's observed.
Live signal Recording
DEEP FOCUS 97%
Current state
DEEP FOCUS
Focus timeline
Three states, twice a second.
100%75%50%25%0%
0s4s8s14s
Deep Focus
97%
Eureka
1%
Struggle
2%
Observation, not inference.
Wrong-answer streaks tell you a child is stuck after the fact. Microexpressions tell you 5–15 seconds earlier. The tutor walks over while the child is still curious — not after they've given up.
Eureka means “level up”.
A clean Eureka spike followed by sustained focus is the most reliable signal that the child has internalised the concept. The system promotes them to the next difficulty rung — the ZPD has shifted, and we shift with it.
Struggle is data, not failure.
Sustained Struggle > 8 seconds without a Eureka spike means we picked the wrong rung. The next problem auto-targets a known scaffold for that misconception. The child never sits in frustration alone.
Built in-house at BXB
See it for yourself.
Focuslab runs in your browser — no install, no upload. Point your camera, write some code or read a passage, and watch the model classify your own focus state in real time.
Try the live demo ↗Sample student dashboard
What “in the band” looks like.
12 weeks of a real-shape progression. Accuracy stays in the 65–80% target band even as difficulty climbs — the signature of a child working at the edge of their ZPD, not above or below it.
Accuracy Problem difficulty Target ZPD band (65–80%)
Read this chart: The accuracy line stays inside the orange band — the child is winning ~70% of the
time. The difficulty line rises week-on-week. That gap closing without accuracy collapsing is the proof that the edge
actually moved.
What parents see
DAILY
Homework log
What was attempted, where the tutor stepped in, and one specific concept that's firming up.
WEEKLY
Friday note
The week's edge — what moved, what's still stuck, what we're trying differently next week.
MONTHLY
Report card
Mastery map by topic, ZPD band time, accuracy trends, and a written summary from the lead tutor.
Ready to see it in action? Book a centre visit.
Book a centre visit.
Watch a live session.
Bring your child. We'll run a short calibration and show you, on screen, exactly where their current edge sits — and what we'd work on first.