India, June 11 -- In her book Recoding America, Jennifer Pahlka recounts her experience co-leading the Employment Development Department (EDD) task force in California during Covid-19. Covid relief was delayed by weeks due to a backlog stemming from outdated technology and rigid rules. For example, minor name mismatches could trigger a manual review, significantly delaying relief payments. We are often surprised when we first read about such systemic failures in a technologically advanced nation. But this is a global problem, a combination of technological choices, government contracting modalities, and rigid administrative processes. At a very different point in our technological journey within government, we face similar challenges in India, as a recent piece about India's first Aadhaar recipient highlights; Ranjana Sonwane is yet to receive her monthly entitlements because of an error linking her bank account to her Aadhaar. We are fortunately at an earlier starting point regarding technology in the Indian State; many of these systems for enhancing frontline welfare delivery are yet to realise their full effectiveness. At the same time, full-stack solutions are already emerging and have even matured across many states in India. The struggles of even a single citizen should motivate us to envision a more robust social protection system, especially given the maturity of supporting digital public infrastructure, such as the linkages between Jan Dhan bank accounts through Aadhaar and UPI protocols on mobile devices (JAM trinity). It is helpful to start with the fundamentals. Why do eligible citizens sometimes get excluded from receiving their rightful entitlements? As Pahlka shows us, technology is only part of the problem. Instead, a core challenge lies in how technology is designed and deployed to identify eligibility. If we begin with the assumptions that most citizens are likely to cheat to receive entitlements they do not deserve, we end up building extremely rigid processes and benchmarks that raise the administrative burden on citizens in terms of documentation, proof, and visits to government offices, that eventually only impacts the poor, educationally and technologically unprepared citizen. As a policy implementation organisation, through our work at Indus Action, we have found that it takes nearly 10 high-touch transactions, including more than three visits to government offices, to secure access to entitlements like scholarships, pensions and maternity benefits. Solving this problem requires combining new technological solutions and reframing the core problem. Let's reframe first, and adapt our strategy to one where we are comfortable living with a small degree of inclusion errors. Some undeserving citizens might get access to entitlements, but lowering the bar for inclusion will help lakhs of marginalised citizens. Technology allows us to check these inclusion errors if we think innovatively. From over a decade of work with governments across India, we have learnt that there are three key challenges in welfare delivery that new systems and technology can solve. The first is discovery: How can governments discover the eligible citizen instead of the other way around? Second, documentation: How can the government design rules and leverage existing data to validate eligibility for low exclusion/inclusion errors? And third, delivery: How can governments fast-track applications upon eligibility check and redress grievances to ensure on-time and quality delivery of entitlements? Luckily, we have enough bright spots from within India for these three significant challenges. Regarding discovery, states now have better quality information to discover vulnerable citizens. States can estimate spatial and household-level vulnerability through family ID linkage across department databases and auto-validate eligible citizens into relevant schemes. For example, states such as Punjab, Rajasthan, Karnataka and Jammu and Kashmir are providing validation tokens to workers who have finished 50 plus days of MGNREGA building and other construction jobs, making them automatically eligible for construction workers' entitlements administered by the labour department. On the burden of documentation, apart from leveraging validations of other departments (for example, birth/death registries; school-going age/grade), the rules governing eligibility can be reviewed to move towards exclusion-based targeting or targeting vulnerable households instead of individuals. For instance, all informal workers can be broadly divided into three archetypes: Farmers, construction workers and other e-Shram registered informal workers like domestic workers and gig workers. Once the family ID is linked to the key occupational information (like KAALIA ID for farmers in Odisha or shramik ID with labour departments), eligibility matching can provide a report of schemes that could be compounded to the family during crucial life-cycle moments. Finally, on the burden of delivery, we need look no further than passport seva and other government to citizen (G2C) services that have been transformed with commitments to operational excellence metrics like TATs (turnaround times) and SLAs (service level agreements). The RTPS framework and Bihar model provide the ambition for delivering grievance redress services for major G2C entitlements in all states. With the latest developments in Gen AI, it is possible to reach India's digitally unprepared citizens through voice in their Indic language, eliminating text barriers. Whether we hear about one citizen not receiving access, or view this as a systemic failure, we have a critical opportunity for learning to iterate better delivery systems and move towards universal access. We have experienced Covid-19 and find ourselves today increasingly unsettled by climate-related shocks. The next time a public health or climate emergency occurs anywhere across India, we can, through our DPI infrastructure for states, immediately release social protection grants to an auto-validated and eligible set of vulnerable citizens. This future is almost here if we act on every error to iterate on our delivery systems....