Learning poverty
Across Sub-Saharan Africa, 89% of children cannot read a simple text by age ten. Kenya's national results show persistent underperformance in mathematics and English in government schools.
Our work
Learning that survives disruption
Drought, displacement, a lost harvest, an unpaid fee balance. Our programmes are designed around the moments that ordinarily end a child's education — and are built to keep working when everything else stops.
What we are up against
Four forces that compound
Financial exclusion
Roughly 73% of Kenyan adults are shut out of formal credit. Annual school costs of KES 40,000–120,000 sit against median household incomes below KES 15,000 a month.
Climate displacement
Kenya hosts more than 650,000 refugees and 316,000 internally displaced people. Schools in Turkana and Garissa face repeated drought- and flood-driven closure.
Teachers stretched thin
In arid and semi-arid counties, teacher shortages reach 40%. Manual assessment and planning consume the hours that should go to instruction.
How the model works
Three mechanisms, one system
These are not parallel interventions. Each one holds the next in place.
Tutoring that reaches the phone
Professor Hoot delivers personalised, curriculum-aligned tutoring over WhatsApp at around 10 KB a session — no app download, no video, and no high-bandwidth requirement.
Fee continuity
Embedded, collateralised school-fee financing removes the barrier at the exact moment a household is most likely to withdraw a child from class.
Teacher capacity
AI-assisted, CBC-aligned planning reduced preparation from roughly three hours to fifteen minutes across 340 teachers — keeping instruction going when staffing is thin.
Evidence
What we have measured
Results from eighteen months of deployment across 82+ partner schools.
+41%
Numeracy
38% → 79% in mathematics
1M+
Tutoring queries
Answered by Professor Hoot
340
Teachers supported
Planning: ~3 hrs → 15 mins
72
Parent NPS
13% organic monthly growth
And what we have not yet proved.
Our current evidence is self-reported pilot data. We have not yet measured
directly in refugee camp settings, the caregiver-led module for four- to
six-year-olds has been designed but not externally evaluated, and the
government-school onboarding model is untested at scale. All three are named
deliverables in our next funding phase.
The next eighteen months
What grant funding buys
-
Months 1–6
Open the government-school route
County education office agreements in Turkana, Garissa and Nairobi, and a documented onboarding framework built for public procurement rather than private-school sales.
-
Months 3–9
Build the early-childhood module
A caregiver-led, play-based companion for children aged four to six, who cannot yet use an AI tutor unaided — piloted with 500 children in crisis-context schools.
-
Months 4–18
Commission independent evaluation
An external RCT or quasi-experimental study, with baseline, midline and endline collection, replacing self-reported results with evidence a ministry can act on.
-
Months 12–18
Prove it in displacement settings
Direct measurement in refugee and IDP contexts, shared device pools tested in camp environments, and an offline-first fallback for total network loss.
Targets
Where this gets us
- 350+ government schools on the platform
- 75,000+ students learning over WhatsApp
- 30,000 children aged 4–6 in caregiver-led sessions
- 8,500 teachers with AI planning support
- A national framework agreement with the Ministry of Education
Fund the evidence, not just the delivery
The largest single line in our current request is independent evaluation. We would like to be measured properly.