Abstract
This whitepaper sets out a blueprint for a sustainable trust layer within digital public infrastructure (DPI): the systems with which citizens seek redress, and through which lost funds are traced, frozen, and recovered. Bilateral and philanthropic donors have funded DPI technology that has expanded access to financial services for over 1.1 billion citizens across low- and middle-income countries (LMICs), but that access has been systematically exploited by transnational criminal networks – global scam losses reached an estimated US$579.4 billion in 2025. Drawing on live AI-based redress deployments across Africa and Asia, on national anti-scam utility design, and on country case studies from Pakistan and Namibia, this whitepaper sets out three integrated functions – redress, recovery, and instant payment system integration – together with a phased stakeholder strategy and a liability waterfall model for long-term sustainability. This paper is intended as a guide for regulators and government agencies, development partners, and philanthropic organisations seeking sustainable models to ensure trust in digital systems.
“If it weren’t for BOB [Proto’s AI agent in the Philippines], I probably wouldn’t have gotten my refund. It’s a very good way for consumers like us to have a voice… you can talk to it like a regular person. You can use English, Tagalog, or Taglish.”
Kitty Elicay
Financial Consumer in the Philippines
This paper draws on live deployments with regulators in Africa and Southeast Asia, including AI redress agents operating in local languages such as Tagalog, Kinyarwanda, and Oshiwambo, and proposes a parallel model for development and philanthropy partners built around three functions:
- Redress: zero-cost grievance intake across messaging, SMS, and voice channels, with AI-mediated resolution and auto-escalation for scam reports
- Recovery: real-time cross-institution tracing and time-limited fund freezing, targeting recovery within the 24-hour window in which it remains feasible
- IPS integration: federating fraud signals into instant payment systems under a hybrid data sovereignty model, so data stays with financial institutions
The core idea is simple:
Trust, interoperability, and linguistic inclusion can be designed into digital systems – but only if citizen redress and fund recovery are treated as shared infrastructure, sustained by a liability model that does not fall on low-income consumers, not as one-off, donor-dependent projects.
Key takeaways
These key takeaways are grounded in evidence from live national deployments and focus on practical design choices that determine whether digital systems build long-term trust or expose new systemic risks.
- Citizen redress and fund recovery are no longer optional add-ons – they are core infrastructure required to sustain trust in scaled digital public systems, alongside identity and payments.
- Global scam losses reached an estimated US$579.4 billion in 2025, and in several LMICs studied, none of the scam reports processed by regulatory AI agents resulted in successful fund recovery.
- A trust layer of three integrated functions – redress, recovery, and instant payment system integration – can resolve grievances and trace, freeze, and recover consumer funds within the 24-hour window in which recovery remains feasible.
- A phased stakeholder strategy – establish, scale, sustain – sequences engagement from the smallest workable coalition through to permanent governance and funding, so that no single participant can structurally veto implementation.
- A liability waterfall model, informed by reimbursement frameworks in the UK, Singapore, and Australia, assigns first, second, and third-pay responsibility across case types, with a pre-funded protection pool so low-income consumers never bear the cost.
- Country case studies from Pakistan and Namibia show how a donor-funded pilot can transition to a self-sustaining model – for example, a prospective success fee on recovered funds that would fund thousands of social-protection grievance cases each year.
- Development funding has the greatest impact when it acts as a catalyst – establishing durable public capacity that transitions to shared-service and domestic financing models.