Vision
── A Next-Generation Educational Model That Defies Taboos
While global education aid amounting to $15–$19 billion annually has improved school enrollment rates, 53% of 10-year-olds remain trapped in "learning poverty" unable to read a simple short sentence (World Bank). Traditional cash transfer models face a common limit: "they increase attendance, but fail to improve academic performance."
We chose Cambodia as our first field of proof.
The annual inflow of $200–$250 million in educational aid to Cambodia is more than enough to directly distribute $40–$50 per month to all children in the poorest segment (approx. 400,000 to 500,000 students) and fully fund their educational opportunities.
Although school enrollment rates exceeded 90% following three decades of support since the 1990s, the majority of funds vanish into intermediate costs. Consequently, 90% of upper-grade students cannot read basic sentences (90% learning poverty rate), and half of rural children drop out for child labor paying $1–$2 a day.
Building on the proven "immediate motivational impact of cash transfers" demonstrated by GiveDirectly and the UN, we deploy an experimental approach that clears all three walls of existing models:
Leveraging national QR payment infrastructure (such as Bakong) that has deeply penetrated local street vendors and mom-and-pop shops, we completely eliminate physical cash transport and intermediaries, directly disbursing funds into children's digital wallets.
Rejecting empty handouts and formalism, we build a mechanism where rewards are unlocked only when children truly understand and solve problems independently through dialogue with an AI private tutor. By directly valuing learning (immediate reflection of up to $1 per day), we offset the opportunity cost of child labor while firmly establishing foundational academic skills.
Slicing away intermediate overhead, we deliver 85% of the capital directly into the hands of the children, fully visualizing and publicly disclosing all learning and payment logs through real-time data.