Feb 24, 2026
# SOLUTION PROPOSAL: AI-Verified Cash Transfer Infrastructure for Crisis-Responsive Poverty Reduction
## SOLUTION TITLE: "Dual-Use Direct Transfer Platform" — Poverty Reduction Infrastructure That Doubles as Crisis Response Capacity
---
## THE PROBLEM (PRECISELY)
**712 million people live below $2.15/day globally, but the delivery infrastructure to reach them is fragmented, expensive, and collapses under crisis conditions.**
Specifically:
- **GiveDirectly's 6-12% delivery costs** ($60-120 per $1,000 transferred) are best-in-class but still leave ~$75 million annually in operational overhead rather than recipient hands
- **Crisis response rebuilds targeting infrastructure from scratch each time** — CoWIN in India succeeded precisely because JAM Trinity already existed; most countries lack equivalent rails
- **The 300+ million people reached by India's DBT system represent proof of concept, but replication elsewhere has stalled** — only 3-4 countries have comparable digital delivery infrastructure
- **Target population for pilot:** 500,000 extreme-poor households in 2-3 African countries with existing mobile money penetration >40% (Kenya, Rwanda, Ghana) who currently receive fragmented aid through 5+ different programs with duplicative verification costs
---
## THE SOLUTION
**Build a shared verification and delivery layer that multiple cash transfer programs can use, explicitly designed to "surge" during crises (climate disasters, health emergencies, economic shocks).**
The platform combines three proven components into a unified stack: (1) **satellite imagery + mobile data for household targeting** (GiveDirectly's approach), (2) **biometric or SIM-linked identity verification** (JAM Trinity's approach), and (3) **pre-registered "dormant beneficiary" lists** that can be activated within 72 hours during declared emergencies. Unlike current systems where poverty programs and emergency response operate on separate rails, this treats the same infrastructure as dual-use — reducing per-program costs through shared overhead while creating crisis-ready capacity as a byproduct.
**Delivery model:** The platform operates as a **nonprofit utility** (similar to M-PESA's original structure or India's NPCI) that charges participating NGOs, governments, and multilaterals a per-transaction fee significantly below their current standalone costs. Revenue from routine poverty transfers cross-subsidizes the maintenance of surge capacity. The platform does NOT make transfer decisions — it provides verified identity, targeting data, and payment rails that program operators use according to their own criteria.
---
## PROOF OF CONCEPT
1. **India's JAM → CoWIN pivot:** The same infrastructure that delivers $360+ billion in poverty benefits was repurposed to administer 2.2 billion vaccine doses. This proves the dual-use concept works at nation-scale, though it required a decade of prior investment.
2. **GiveDirectly's Kenya operations:** Already demonstrates satellite + mobile money targeting at $60-120 per $1,000 delivered. Their 2020 COVID response showed they could identify and reach 300,000 new recipients in 8 weeks — but only because they'd pre-built the targeting infrastructure for routine transfers.
3. **Togo's Novissi program (2020):** Used mobile money data + machine learning to identify and pay 570,000 informal workers within 10 days of COVID lockdown. Achieved 92% targeting accuracy for the poorest quintile. Proved rapid deployment possible but was built ad-hoc and not sustained post-crisis.
---
## ECONOMICS
**Unit economics at pilot scale (500,000 households):**
- Current state: 5 programs each spending $40-80 per household on verification/targeting = $200-400 per household in duplicative overhead
- Proposed state: Shared platform costs $25-35 per household for initial enrollment, $3-5 per subsequent transaction
- **Net savings: $100-250 per household over 3 years**, split between participating programs
**Cost drivers:**
- Satellite imagery licensing: $0.50-2.00 per km² (Planet Labs, Maxar pricing)
- Mobile money transaction fees: 1-3% of transfer value (negotiable at volume)
- Field verification (10% sample): $15-25 per household
- Platform development and maintenance: $2-4 million annually at pilot scale
- Local staff and partnerships: $1-2 million annually
**Who pays:**
- **Phase 1 (pilot):** Philanthropic capital (GiveDirectly, Open Philanthropy, Gates Foundation) covers platform build; participating NGOs pay per-transaction fees
- **Phase 2 (scale):** Government social protection programs pay subscription fees; crisis surge capacity funded by humanitarian pooled funds (CERF, Start Fund) or pre-negotiated World Bank contingent financing
- **Phase 3 (sustainability):** Transaction fees at volume cover operating costs; crisis capacity becomes a public good funded by development finance institutions
---
## SCALE PATH
**Pilot (Year 1):** 500,000 households across Kenya, Rwanda, Ghana — chosen for mobile money penetration, existing GiveDirectly presence, and government openness to digital delivery.
**Expansion (Years 2-3):** Add 3-5 countries, reaching 2-3 million households. Critical test: successfully surge during at least one declared emergency, demonstrating dual-use value.
**Scale (Years 4-5):** Integrate with government social protection systems in 2+ countries as official delivery infrastructure. Target: 10+ million households, with platform costs below $15 per household annually.
**Critical bottleneck:** Government adoption. NGO-only usage caps scale at ~5 million households; government social protection programs are 10-50x larger. The pivot from "NGO tool" to "public infrastructure" requires demonstrating reliability, data security, and political neutrality that governments will trust.
**Secondary bottleneck:** Interoperability across mobile money providers. Kenya (M-PESA dominant) is easier than Ghana (4+ major providers). Platform must be provider-agnostic from day one.
---
## WHAT NEEDS TO HAPPEN NEXT
1. **Convene a technical working group (April 2026):** GiveDirectly, Togo's Novissi team, India's NPCI, and 2-3 African mobile money operators to define minimum viable platform specifications. Specific output: Technical requirements document and data-sharing MOU template.
2. **Secure $5-8 million in catalytic funding (Q2 2026):** Approach Open Philanthropy, Co-Impact, and
## SOLUTION TITLE: "Dual-Use Direct Transfer Platform" — Poverty Reduction Infrastructure That Doubles as Crisis Response Capacity
---
## THE PROBLEM (PRECISELY)
**712 million people live below $2.15/day globally, but the delivery infrastructure to reach them is fragmented, expensive, and collapses under crisis conditions.**
Specifically:
- **GiveDirectly's 6-12% delivery costs** ($60-120 per $1,000 transferred) are best-in-class but still leave ~$75 million annually in operational overhead rather than recipient hands
- **Crisis response rebuilds targeting infrastructure from scratch each time** — CoWIN in India succeeded precisely because JAM Trinity already existed; most countries lack equivalent rails
- **The 300+ million people reached by India's DBT system represent proof of concept, but replication elsewhere has stalled** — only 3-4 countries have comparable digital delivery infrastructure
- **Target population for pilot:** 500,000 extreme-poor households in 2-3 African countries with existing mobile money penetration >40% (Kenya, Rwanda, Ghana) who currently receive fragmented aid through 5+ different programs with duplicative verification costs
---
## THE SOLUTION
**Build a shared verification and delivery layer that multiple cash transfer programs can use, explicitly designed to "surge" during crises (climate disasters, health emergencies, economic shocks).**
The platform combines three proven components into a unified stack: (1) **satellite imagery + mobile data for household targeting** (GiveDirectly's approach), (2) **biometric or SIM-linked identity verification** (JAM Trinity's approach), and (3) **pre-registered "dormant beneficiary" lists** that can be activated within 72 hours during declared emergencies. Unlike current systems where poverty programs and emergency response operate on separate rails, this treats the same infrastructure as dual-use — reducing per-program costs through shared overhead while creating crisis-ready capacity as a byproduct.
**Delivery model:** The platform operates as a **nonprofit utility** (similar to M-PESA's original structure or India's NPCI) that charges participating NGOs, governments, and multilaterals a per-transaction fee significantly below their current standalone costs. Revenue from routine poverty transfers cross-subsidizes the maintenance of surge capacity. The platform does NOT make transfer decisions — it provides verified identity, targeting data, and payment rails that program operators use according to their own criteria.
---
## PROOF OF CONCEPT
1. **India's JAM → CoWIN pivot:** The same infrastructure that delivers $360+ billion in poverty benefits was repurposed to administer 2.2 billion vaccine doses. This proves the dual-use concept works at nation-scale, though it required a decade of prior investment.
2. **GiveDirectly's Kenya operations:** Already demonstrates satellite + mobile money targeting at $60-120 per $1,000 delivered. Their 2020 COVID response showed they could identify and reach 300,000 new recipients in 8 weeks — but only because they'd pre-built the targeting infrastructure for routine transfers.
3. **Togo's Novissi program (2020):** Used mobile money data + machine learning to identify and pay 570,000 informal workers within 10 days of COVID lockdown. Achieved 92% targeting accuracy for the poorest quintile. Proved rapid deployment possible but was built ad-hoc and not sustained post-crisis.
---
## ECONOMICS
**Unit economics at pilot scale (500,000 households):**
- Current state: 5 programs each spending $40-80 per household on verification/targeting = $200-400 per household in duplicative overhead
- Proposed state: Shared platform costs $25-35 per household for initial enrollment, $3-5 per subsequent transaction
- **Net savings: $100-250 per household over 3 years**, split between participating programs
**Cost drivers:**
- Satellite imagery licensing: $0.50-2.00 per km² (Planet Labs, Maxar pricing)
- Mobile money transaction fees: 1-3% of transfer value (negotiable at volume)
- Field verification (10% sample): $15-25 per household
- Platform development and maintenance: $2-4 million annually at pilot scale
- Local staff and partnerships: $1-2 million annually
**Who pays:**
- **Phase 1 (pilot):** Philanthropic capital (GiveDirectly, Open Philanthropy, Gates Foundation) covers platform build; participating NGOs pay per-transaction fees
- **Phase 2 (scale):** Government social protection programs pay subscription fees; crisis surge capacity funded by humanitarian pooled funds (CERF, Start Fund) or pre-negotiated World Bank contingent financing
- **Phase 3 (sustainability):** Transaction fees at volume cover operating costs; crisis capacity becomes a public good funded by development finance institutions
---
## SCALE PATH
**Pilot (Year 1):** 500,000 households across Kenya, Rwanda, Ghana — chosen for mobile money penetration, existing GiveDirectly presence, and government openness to digital delivery.
**Expansion (Years 2-3):** Add 3-5 countries, reaching 2-3 million households. Critical test: successfully surge during at least one declared emergency, demonstrating dual-use value.
**Scale (Years 4-5):** Integrate with government social protection systems in 2+ countries as official delivery infrastructure. Target: 10+ million households, with platform costs below $15 per household annually.
**Critical bottleneck:** Government adoption. NGO-only usage caps scale at ~5 million households; government social protection programs are 10-50x larger. The pivot from "NGO tool" to "public infrastructure" requires demonstrating reliability, data security, and political neutrality that governments will trust.
**Secondary bottleneck:** Interoperability across mobile money providers. Kenya (M-PESA dominant) is easier than Ghana (4+ major providers). Platform must be provider-agnostic from day one.
---
## WHAT NEEDS TO HAPPEN NEXT
1. **Convene a technical working group (April 2026):** GiveDirectly, Togo's Novissi team, India's NPCI, and 2-3 African mobile money operators to define minimum viable platform specifications. Specific output: Technical requirements document and data-sharing MOU template.
2. **Secure $5-8 million in catalytic funding (Q2 2026):** Approach Open Philanthropy, Co-Impact, and