AI-native credit intelligence on digital public infrastructure: evidence from India's account aggregator ecosystem

Purpose This paper examines how India's Account Aggregator (AA) framework (269 million consents, 2.12 billion enabled accounts) can support an AI-native consumer credit intelligence platform. It develops testable hypotheses for expanding financial product eligibility to 300–450 million underserved adults across housing, personal lending, MSME credit and insurance. Design/methodology/approach The study combines institutional analysis of India's regulatory architecture with econometric specifications: probit models, difference-in-differences exploiting staggered AA rollout, panel IV estimation, regression discontinuity and heterogeneous treatment effects. The framework extends Rochet and Tirole’s (2003) two-sided market model to DPI-augmented platforms. The paper is conceptual. It sets out these identification strategies, a five-scenario impact assessment and a platform cost model as a research agenda, and does not report estimated causal effects. Findings India's AA data layer, UPI transaction histories and Aadhaar verification enable credit scoring for populations invisible to bureaux. Scenario analysis across five cases projects incremental demand between $180 billion and $480 billion over a decade, with a genuinely incremental impact of $108–216 billion after organic growth adjustment. Originality/value This study develops a formal model of a credit platform on government-provided digital infrastructure, showing DPI-augmented platforms face different pricing dynamics. It proposes India-specific algorithmic fairness criteria for caste, gender and geography, and treats regulatory constraints as binding design parameters.